Hubbry Logo
logo
Projective geometry avatar
Projective geometry
Community hub

Projective geometry

logo
0 subscribers
Read side by side
Projective geometry
View on Wikipedia
from Wikipedia

Geometry
Stereographic projection from the top of a sphere onto a plane beneath it
Projecting a sphere to a plane
Branches
  • Concepts
  • Features
Dimension
Zero-dimensional
One-dimensional
Two-dimensional
Triangle
Quadrilateral
Circle
Three-dimensional
Polyhedron
Solid of revolution
Four-/other-dimensional
Geometers
by name
by period
BCE
1–1400s
1400s–1700s
1700s–1900s
Present day

In mathematics, projective geometry is the study of geometric properties that are invariant with respect to projective transformations. This means that, compared to elementary Euclidean geometry, projective geometry has a different setting (projective space) and a selective set of basic geometric concepts. The basic intuitions are that projective space has more points than Euclidean space, for a given dimension, and that geometric transformations are permitted that transform the extra points (called "points at infinity") to Euclidean points, and vice versa.

Properties meaningful for projective geometry are respected by this new idea of transformation, which is more radical in its effects than can be expressed by a transformation matrix and translations (the affine transformations). The first issue for geometers is what kind of geometry is adequate for a novel situation. Unlike in Euclidean geometry, the concept of an angle does not apply in projective geometry, because no measure of angles is invariant with respect to projective transformations, as is seen in perspective drawing from a changing perspective. One source for projective geometry was indeed the theory of perspective. Another difference from elementary geometry is the way in which parallel lines can be said to meet in a point at infinity, once the concept is translated into projective geometry's terms. Again this notion has an intuitive basis, such as railway tracks meeting at the horizon in a perspective drawing. See Projective plane for the basics of projective geometry in two dimensions.

While the ideas were available earlier, projective geometry was mainly a development of the 19th century. This included the theory of complex projective space, the coordinates used (homogeneous coordinates) being complex numbers. Several major types of more abstract mathematics (including invariant theory, the Italian school of algebraic geometry, and Felix Klein's Erlangen programme resulting in the study of the classical groups) were motivated by projective geometry. It was also a subject with many practitioners for its own sake, as synthetic geometry. Another topic that developed from axiomatic studies of projective geometry is finite geometry.

The topic of projective geometry is itself now divided into many research subtopics, two examples of which are projective algebraic geometry (the study of projective varieties) and projective differential geometry (the study of differential invariants of the projective transformations).

Overview

[edit]

Projective geometry is an elementary non-metrical form of geometry, meaning that it does not support any concept of distance. In two dimensions it begins with the study of configurations of points and lines. That there is indeed some geometric interest in this sparse setting was first established by Desargues and others in their exploration of the principles of perspective art.[1] In higher dimensional spaces there are considered hyperplanes (that always meet), and other linear subspaces, which exhibit the principle of duality. The simplest illustration of duality is in the projective plane, where the statements "two distinct points determine a unique line" (i.e. the line through them) and "two distinct lines determine a unique point" (i.e. their point of intersection) show the same structure as propositions. Projective geometry can also be seen as a geometry of constructions with a straight-edge alone, excluding compass constructions, common in straightedge and compass constructions.[2] As such, there are no circles, no angles, no measurements, no parallels, and no concept of intermediacy (or "betweenness").[3] It was realised that the theorems that do apply to projective geometry are simpler statements. For example, the different conic sections are all equivalent in (complex) projective geometry, and some theorems about circles can be considered as special cases of these general theorems.

During the early 19th century the work of Jean-Victor Poncelet, Lazare Carnot and others established projective geometry as an independent field of mathematics.[3] Its rigorous foundations were addressed by Karl von Staudt and perfected by Italians Giuseppe Peano, Mario Pieri, Alessandro Padoa and Gino Fano during the late 19th century.[4] Projective geometry, like affine and Euclidean geometry, can also be developed from the Erlangen program of Felix Klein; projective geometry is characterized by invariants under transformations of the projective group.

After much work on the very large number of theorems in the subject, therefore, the basics of projective geometry became understood. The incidence structure and the cross-ratio are fundamental invariants under projective transformations. Projective geometry can be modeled by the affine plane (or affine space) plus a line (hyperplane) "at infinity" and then treating that line (or hyperplane) as "ordinary".[5] An algebraic model for doing projective geometry in the style of analytic geometry is given by homogeneous coordinates.[6][7] On the other hand, axiomatic studies revealed the existence of non-Desarguesian planes, examples to show that the axioms of incidence can be modelled (in two dimensions only) by structures not accessible to reasoning through homogeneous coordinate systems.

Growth measure and the polar vortices. Based on the work of Lawrence Edwards

In a foundational sense, projective geometry and ordered geometry are elementary since they each involve a minimal set of axioms and either can be used as the foundation for affine and Euclidean geometry.[8][9] Projective geometry is not "ordered"[3] and so it is a distinct foundation for geometry.

Description

[edit]

Projective geometry is less restrictive than either Euclidean geometry or affine geometry. It is an intrinsically non-metrical geometry, meaning that facts are independent of any metric structure. Under the projective transformations, the incidence structure and the relation of projective harmonic conjugates are preserved. A projective range is the one-dimensional foundation. Projective geometry formalizes one of the central principles of perspective art: that parallel lines meet at infinity, and therefore are drawn that way. In essence, a projective geometry may be thought of as an extension of Euclidean geometry in which the "direction" of each line is subsumed within the line as an extra "point", and in which a "horizon" of directions corresponding to coplanar lines is regarded as a "line". Thus, two parallel lines meet on a horizon line by virtue of their incorporating the same direction.

Idealized directions are referred to as points at infinity, while idealized horizons are referred to as lines at infinity. In turn, all these lines lie in the plane at infinity. However, infinity is a metric concept, so a purely projective geometry does not single out any points, lines or planes in this regard—those at infinity are treated just like any others.

Because Euclidean geometry is contained within a projective geometry—with projective geometry having a simpler foundation—general results in Euclidean geometry may be derived in a more transparent manner, where separate but similar theorems of Euclidean geometry may be handled collectively within the framework of projective geometry. For example, parallel and nonparallel lines need not be treated as separate cases; rather an arbitrary projective plane is singled out as the ideal plane and located "at infinity" using homogeneous coordinates.

Additional properties of fundamental importance include Desargues' Theorem and the Pappus's hexagon theorem. In projective spaces of dimension 3 or greater there is a construction that allows one to prove Desargues' Theorem. But for dimension 2, it must be separately postulated.

Using Desargues' Theorem, combined with the other axioms, it is possible to define the basic operations of arithmetic, geometrically. The resulting operations satisfy the axioms of a field – except that the commutativity of multiplication requires Pappus's hexagon theorem. As a result, the points of each line are in one-to-one correspondence with a given field, F, supplemented by an additional element, ∞, such that r ⋅ ∞ = ∞, −∞ = ∞, r + ∞ = ∞, r / 0 = ∞, r / ∞ = 0, ∞ − r = r − ∞ = ∞, except that 0 / 0, ∞ / ∞, ∞ + ∞, ∞ − ∞, 0 ⋅ ∞ and ∞ ⋅ 0 remain undefined.

Projective geometry also includes a full theory of conic sections, a subject also extensively developed in Euclidean geometry. There are advantages to being able to think of a hyperbola and an ellipse as distinguished only by the way the hyperbola lies across the line at infinity; and that a parabola is distinguished only by being tangent to the same line. The whole family of circles can be considered as conics passing through two given points on the line at infinity — at the cost of requiring complex coordinates. Since coordinates are not "synthetic", one replaces them by fixing a line and two points on it, and considering the linear system of all conics passing through those points as the basic object of study. This method proved very attractive to talented geometers, and the topic was studied thoroughly. An example of this method is the multi-volume treatise by H. F. Baker.

History

[edit]
Further information: Mathematics and art

The first geometrical properties of a projective nature were discovered during the 3rd century by Pappus of Alexandria.[3] Filippo Brunelleschi (1404–1472) started investigating the geometry of perspective during 1425[10] (see Perspective (graphical) § History for a more thorough discussion of the work in the fine arts that motivated much of the development of projective geometry). Johannes Kepler (1571–1630) and Girard Desargues (1591–1661) independently developed the concept of the "point at infinity".[11] Desargues developed an alternative way of constructing perspective drawings by generalizing the use of vanishing points to include the case when these are infinitely far away. He made Euclidean geometry, where parallel lines are truly parallel, into a special case of an all-encompassing geometric system. Desargues's study on conic sections drew the attention of 16-year-old Blaise Pascal and helped him formulate Pascal's theorem. The works of Gaspard Monge at the end of 18th and beginning of 19th century were important for the subsequent development of projective geometry. The work of Desargues was ignored until Michel Chasles chanced upon a handwritten copy during 1845. Meanwhile, Jean-Victor Poncelet had published the foundational treatise on projective geometry during 1822. Poncelet examined the projective properties of objects (those invariant under central projection) and, by basing his theory on the concrete pole and polar relation with respect to a circle, established a relationship between metric and projective properties. The non-Euclidean geometries discovered soon thereafter were eventually demonstrated to have models, such as the Klein model of hyperbolic space, relating to projective geometry.

In 1855 A. F. Möbius wrote an article about permutations, now called Möbius transformations, of generalised circles in the complex plane. These transformations represent projectivities of the complex projective line. In the study of lines in space, Julius Plücker used homogeneous coordinates in his description, and the set of lines was viewed on the Klein quadric, one of the early contributions of projective geometry to a new field called algebraic geometry, an offshoot of analytic geometry with projective ideas.

Projective geometry was instrumental in the validation of speculations of Lobachevski and Bolyai concerning hyperbolic geometry by providing models for the hyperbolic plane:[12] for example, the Poincaré disc model where generalised circles perpendicular to the unit circle correspond to "hyperbolic lines" (geodesics), and the "translations" of this model are described by Möbius transformations that map the unit disc to itself. The distance between points is given by a Cayley–Klein metric, known to be invariant under the translations since it depends on cross-ratio, a key projective invariant. The translations are described variously as isometries in metric space theory, as linear fractional transformations formally, and as projective linear transformations of the projective linear group, in this case SU(1, 1).

The work of Poncelet, Jakob Steiner and others was not intended to extend analytic geometry. Techniques were supposed to be synthetic: in effect projective space as now understood was to be introduced axiomatically. As a result, reformulating early work in projective geometry so that it satisfies current standards of rigor can be somewhat difficult. Even in the case of the projective plane alone, the axiomatic approach can result in models not describable via linear algebra.

This period in geometry was overtaken by research on the general algebraic curve by Clebsch, Riemann, Max Noether and others, which stretched existing techniques, and then by invariant theory. Towards the end of the century, the Italian school of algebraic geometry (Enriques, Segre, Severi) broke out of the traditional subject matter into an area demanding deeper techniques.

During the later part of the 19th century, the detailed study of projective geometry became less fashionable, although the literature is voluminous. Some important work was done in enumerative geometry in particular, by Schubert, that is now considered as anticipating the theory of Chern classes, taken as representing the algebraic topology of Grassmannians.

Projective geometry later proved key to Paul Dirac's invention of quantum mechanics. At a foundational level, the discovery that quantum measurements could fail to commute had disturbed and dissuaded Heisenberg, but past study of projective planes over noncommutative rings had likely desensitized Dirac. In more advanced work, Dirac used extensive drawings in projective geometry to understand the intuitive meaning of his equations, before writing up his work in an exclusively algebraic formalism.[13]

Classification

[edit]

There are many projective geometries, which may be divided into discrete and continuous: a discrete geometry comprises a set of points, which may or may not be finite in number, while a continuous geometry has infinitely many points with no gaps in between.

The only projective geometry of dimension 0 is a single point. A projective geometry of dimension 1 consists of a single line containing at least 3 points. The geometric construction of arithmetic operations cannot be performed in either of these cases. For dimension 2, there is a rich structure in virtue of the absence of Desargues' Theorem.

The Fano plane is the projective plane with the fewest points and lines.

The smallest 2-dimensional projective geometry (that with the fewest points) is the Fano plane, which has 3 points on every line, with 7 points and 7 lines in all, having the following collinearities:

  • [ABC]
  • [ADE]
  • [AFG]
  • [BDG]
  • [BEF]
  • [CDF]
  • [CEG]

with homogeneous coordinates A = (0,0,1), B = (0,1,1), C = (0,1,0), D = (1,0,1), E = (1,0,0), F = (1,1,1), G = (1,1,0), or, in affine coordinates, A = (0,0), B = (0,1), C = (∞), D = (1,0), E = (0), F = (1,1)and G = (1). The affine coordinates in a Desarguesian plane for the points designated to be the points at infinity (in this example: C, E and G) can be defined in several other ways.

In standard notation, a finite projective geometry is written PG(a, b) where:

a is the projective (or geometric) dimension, and
b is one less than the number of points on a line (called the order of the geometry).

Thus, the example having only 7 points is written PG(2, 2).

The term "projective geometry" is used sometimes to indicate the generalised underlying abstract geometry, and sometimes to indicate a particular geometry of wide interest, such as the metric geometry of flat space which we analyse through the use of homogeneous coordinates, and in which Euclidean geometry may be embedded (hence its name, Extended Euclidean plane).

The fundamental property that singles out all projective geometries is the elliptic incidence property that any two distinct lines L and M in the projective plane intersect at exactly one point P. The special case in analytic geometry of parallel lines is subsumed in the smoother form of a line at infinity on which P lies. The line at infinity is thus a line like any other in the theory: it is in no way special or distinguished. (In the later spirit of the Erlangen programme one could point to the way the group of transformations can move any line to the line at infinity).

The parallel properties of elliptic, Euclidean and hyperbolic geometries contrast as follows:

Given a line l and a point P not on the line,
Elliptic
there exists no line through P that does not meet l
Euclidean
there exists exactly one line through P that does not meet l
Hyperbolic
there exists more than one line through P that does not meet l

The parallel property of elliptic geometry is the key idea that leads to the principle of projective duality, possibly the most important property that all projective geometries have in common.

Duality

[edit]
Further information: Duality (projective geometry)

In 1825, Joseph Gergonne noted the principle of duality characterizing projective plane geometry: given any theorem or definition of that geometry, substituting point for line, lie on for pass through, collinear for concurrent, intersection for join, or vice versa, results in another theorem or valid definition, the "dual" of the first. Similarly in 3 dimensions, the duality relation holds between points and planes, allowing any theorem to be transformed by swapping point and plane, is contained by and contains. More generally, for projective spaces of dimension N, there is a duality between the subspaces of dimension R and dimension N − R − 1. For N = 2, this specializes to the most commonly known form of duality—that between points and lines. The duality principle was also discovered independently by Jean-Victor Poncelet.

To establish duality only requires establishing theorems which are the dual versions of the axioms for the dimension in question. Thus, for 3-dimensional spaces, one needs to show that (1*) every point lies in 3 distinct planes, (2*) every two planes intersect in a unique line and a dual version of (3*) to the effect: if the intersection of plane P and Q is coplanar with the intersection of plane R and S, then so are the respective intersections of planes P and R, Q and S (assuming planes P and S are distinct from Q and R).

In practice, the principle of duality allows us to set up a dual correspondence between two geometric constructions. The most famous of these is the polarity or reciprocity of two figures in a conic curve (in 2 dimensions) or a quadric surface (in 3 dimensions). A commonplace example is found in the reciprocation of a symmetrical polyhedron in a concentric sphere to obtain the dual polyhedron.

Another example is Brianchon's theorem, the dual of the already mentioned Pascal's theorem, and one of whose proofs simply consists of applying the principle of duality to Pascal's. Here are comparative statements of these two theorems (in both cases within the framework of the projective plane):

  • Pascal: If all six vertices of a hexagon lie on a conic, then the intersections of its opposite sides (regarded as full lines, since in the projective plane there is no such thing as a "line segment") are three collinear points. The line joining them is then called the Pascal line of the hexagon.
  • Brianchon: If all six sides of a hexagon are tangent to a conic, then its diagonals (i.e. the lines joining opposite vertices) are three concurrent lines. Their point of intersection is then called the Brianchon point of the hexagon.
(If the conic degenerates into two straight lines, Pascal's becomes Pappus's theorem, which has no interesting dual, since the Brianchon point trivially becomes the two lines' intersection point.)

Axioms of projective geometry

[edit]

Any given geometry may be deduced from an appropriate set of axioms. Projective geometries are characterised by the "elliptic parallel" axiom, that any two planes always meet in just one line, or in the plane, any two lines always meet in just one point. In other words, there are no such things as parallel lines or planes in projective geometry.

Many alternative sets of axioms for projective geometry have been proposed (see for example Coxeter 2003, Hilbert & Cohn-Vossen 1999, Greenberg 1980).

Whitehead's axioms

[edit]

These axioms are based on Whitehead, "The Axioms of Projective Geometry". There are two types, points and lines, and one "incidence" relation between points and lines. The three axioms are:

  • G1: Every line contains at least 3 points
  • G2: Every two distinct points, A and B, lie on a unique line, AB.
  • G3: If lines AB and CD intersect, then so do lines AC and BD (where it is assumed that A and D are distinct from B and C).

The reason each line is assumed to contain at least 3 points is to eliminate some degenerate cases. The spaces satisfying these three axioms either have at most one line, or are projective spaces of some dimension over a division ring, or are non-Desarguesian planes.

Additional axioms

[edit]

One can add further axioms restricting the dimension or the coordinate ring. For example, Coxeter's Projective Geometry,[14] references Veblen[15] in the three axioms above, together with a further 5 axioms that make the dimension 3 and the coordinate ring a commutative field of characteristic not 2.

Axioms using a ternary relation

[edit]

One can pursue axiomatization by postulating a ternary relation, [ABC] to denote when three points (not all necessarily distinct) are collinear. An axiomatization may be written down in terms of this relation as well:

  • C0: [ABA]
  • C1: If A and B are distinct points such that [ABC] and [ABD] then [BDC]
  • C2: If A and B are distinct points then there exists a third distinct point C such that [ABC]
  • C3: If A and C are distinct points, and B and D are distinct points, with [BCE] and [ADE] but not [ABE], then there is a point F such that [ACF] and [BDF].

For two distinct points, A and B, the line AB is defined as consisting of all points C for which [ABC]. The axioms C0 and C1 then provide a formalization of G2; C2 for G1 and C3 for G3.

The concept of line generalizes to planes and higher-dimensional subspaces. A subspace, AB...XY may thus be recursively defined in terms of the subspace AB...X as that containing all the points of all lines YZ, as Z ranges over AB...X. Collinearity then generalizes to the relation of "independence". A set {A, B, ..., Z} of points is independent, [AB...Z] if {A, B, ..., Z} is a minimal generating subset for the subspace AB...Z.

The projective axioms may be supplemented by further axioms postulating limits on the dimension of the space. The minimum dimension is determined by the existence of an independent set of the required size. For the lowest dimensions, the relevant conditions may be stated in equivalent form as follows. A projective space is of:

  • (L1) at least dimension 0 if it has at least 1 point,
  • (L2) at least dimension 1 if it has at least 2 distinct points (and therefore a line),
  • (L3) at least dimension 2 if it has at least 3 non-collinear points (or two lines, or a line and a point not on the line),
  • (L4) at least dimension 3 if it has at least 4 non-coplanar points.

The maximum dimension may also be determined in a similar fashion. For the lowest dimensions, they take on the following forms. A projective space is of:

  • (M1) at most dimension 0 if it has no more than 1 point,
  • (M2) at most dimension 1 if it has no more than 1 line,
  • (M3) at most dimension 2 if it has no more than 1 plane,

and so on. It is a general theorem (a consequence of axiom (3)) that all coplanar lines intersect—the very principle that projective geometry was originally intended to embody. Therefore, property (M3) may be equivalently stated that all lines intersect one another.

It is generally assumed that projective spaces are of at least dimension 2. In some cases, if the focus is on projective planes, a variant of M3 may be postulated. The axioms of (Eves 1997: 111), for instance, include (1), (2), (L3) and (M3). Axiom (3) becomes vacuously true under (M3) and is therefore not needed in this context.

Axioms for projective planes

[edit]
Main article: Projective plane

In incidence geometry, most authors[16] give a treatment that embraces the Fano plane PG(2, 2) as the smallest finite projective plane. An axiom system that achieves this is as follows:

  • (P1) Any two distinct points lie on a line that is unique.
  • (P2) Any two distinct lines meet at a point that is unique.
  • (P3) There exist at least four points of which no three are collinear.

Coxeter's Introduction to Geometry[17] gives a list of five axioms for a more restrictive concept of a projective plane that is attributed to Bachmann, adding Pappus's theorem to the list of axioms above (which eliminates non-Desarguesian planes) and excluding projective planes over fields of characteristic 2 (those that do not satisfy Fano's axiom). The restricted planes given in this manner more closely resemble the real projective plane.

Perspectivity and projectivity

[edit]

Given three non-collinear points, there are three lines connecting them, but with four points, no three collinear, there are six connecting lines and three additional "diagonal points" determined by their intersections. The science of projective geometry captures this surplus determined by four points through a quaternary relation and the projectivities which preserve the complete quadrangle configuration.

An harmonic quadruple of points on a line occurs when there is a complete quadrangle two of whose diagonal points are in the first and third position of the quadruple, and the other two positions are points on the lines joining two quadrangle points through the third diagonal point.[18]

A spatial perspectivity of a projective configuration in one plane yields such a configuration in another, and this applies to the configuration of the complete quadrangle. Thus harmonic quadruples are preserved by perspectivity. If one perspectivity follows another the configurations follow along. The composition of two perspectivities is no longer a perspectivity, but a projectivity.

While corresponding points of a perspectivity all converge at a point, this convergence is not true for a projectivity that is not a perspectivity. In projective geometry the intersection of lines formed by corresponding points of a projectivity in a plane are of particular interest. The set of such intersections is called a projective conic, and in acknowledgement of the work of Jakob Steiner, it is referred to as a Steiner conic.

Suppose a projectivity is formed by two perspectivities centered on points A and B, relating x to X by an intermediary p:

The projectivity is then Then given the projectivity the induced conic is

Given a conic C and a point P not on it, two distinct secant lines through P intersect C in four points. These four points determine a quadrangle of which P is a diagonal point. The line through the other two diagonal points is called the polar of P and P is the pole of this line.[19] Alternatively, the polar line of P is the set of projective harmonic conjugates of P on a variable secant line passing through P and C.

See also

[edit]

Notes

[edit]

References

[edit]
[edit]
Revisions and contributorsEdit on WikipediaRead on Wikipedia

Projective geometry

View on Grokipedia
from Grokipedia
Projective geometry is a fundamental branch of mathematics that extends Euclidean geometry by incorporating points at infinity, where parallel lines are considered to intersect, thereby unifying the treatment of lines and eliminating the distinction between parallel and intersecting lines.[1] It focuses on properties invariant under projective transformations, which are bijective mappings that preserve collinearity and the cross-ratio of points on a line, and is typically defined axiomatically through incidence relations: any two distinct points determine a unique line, and any two distinct lines intersect at a unique point.[2] This framework allows for the study of geometric configurations in a projective space, often represented using homogeneous coordinates, where points are equivalence classes of vectors in a vector space.[1] The origins of projective geometry trace back to the Renaissance era, when artists like Filippo Brunelleschi in the early 15th century developed techniques for linear perspective to realistically depict three-dimensional scenes on two-dimensional canvases, addressing how parallel lines appear to converge at a vanishing point.[3] This artistic innovation was mathematically formalized in the 17th century by Girard Desargues, who introduced key theorems on perspective and conic sections, and Blaise Pascal, who extended properties of conics through projective methods.[3] The field gained rigorous axiomatic foundations in the 19th century through the work of Jean-Victor Poncelet, who in 1822 published Traité des propriétés projectives des figures, emphasizing synthetic methods and duality between points and lines, followed by contributions from August Ferdinand Möbius, Julius Plücker, and others who explored transformations and invariants.[4] Central to projective geometry are concepts like duality, which interchanges points and lines while preserving incidence, enabling theorems such as Desargues' theorem (stating that two triangles are perspective from a point if and only if they are perspective from a line)[5] and Pappus's theorem on collinear points generated by intersecting lines from two sets.[6] The cross-ratio, a projective invariant measuring the division of a line segment by four collinear points, plays a crucial role in classifying configurations and transformations.[7] Projective spaces can be constructed over any field, leading to finite projective planes like the Fano plane, which has seven points and seven lines, illustrating the geometry's abstract nature beyond the real numbers.[8] In modern applications, projective geometry underpins computer vision by modeling perspective projections from 3D scenes onto 2D images, facilitating tasks like camera calibration and 3D reconstruction through homographies and fundamental matrices.[7] It also informs computer graphics for rendering realistic scenes and has historical ties to algebraic geometry, where projective varieties extend the study of curves and surfaces.[1] Overall, projective geometry provides a metric-free foundation that reveals deep symmetries in geometric structures, influencing fields from art to theoretical physics.[5]

Introduction

Overview

Projective geometry is the branch of mathematics that studies properties of geometric figures invariant under projective transformations, which include central projections from one plane to another.[1] The term "projective" derives from these central projections, where rays emanate from a fixed point (the center of projection) to map points from one surface onto another, preserving certain relational structures like incidence and cross-ratios.[9] A fundamental distinction from Euclidean geometry lies in the treatment of parallelism: in projective geometry, there are no parallel lines, as all lines intersect at some point, including those that would be parallel in the Euclidean plane, which meet at points on a line at infinity.[1] This addition of an ideal line at infinity unifies the treatment of lines, eliminating special cases in theorems and enabling a more symmetric study of configurations.[10] An illustrative example is the central projection of a sphere onto a plane, where circles on the sphere map to conic sections—such as ellipses, parabolas, or hyperbolas—on the plane, demonstrating how projective transformations preserve the projective class of curves without regard to metric properties like size or shape.[11] This projection highlights the focus on qualitative incidences rather than quantitative measures, often formalized using homogeneous coordinates to incorporate points at infinity.[1]

Motivations and Basic Properties

Projective geometry emerged as a response to certain limitations in Euclidean geometry, particularly the inconsistent treatment of parallel lines, which are assumed not to intersect within the finite plane. This creates special cases in theorems and proofs, complicating the study of geometric configurations. By introducing points at infinity—ideal points where parallel lines are considered to meet—projective geometry resolves these issues, ensuring a uniform framework where all lines intersect, thus simplifying and generalizing Euclidean results.[12][1] A key motivation lies in the unification of conic sections, which appear as distinct entities (ellipses, parabolas, hyperbolas) in Euclidean geometry but are revealed as equivalent under projective transformations. Specifically, these curves are all projective images of a circle, allowing a single set of properties to describe them without the Euclidean distinctions based on eccentricity or intersection with the line at infinity. This perspective, rooted in the invariance of incidence relations under projection, provides a more elegant classification and facilitates applications in areas like optics and computer vision.[12][13] Among its basic properties, projective geometry adheres to the principle that any two distinct points determine a unique line, mirroring the Euclidean axiom but extended to the full space including infinity. Dually, any two distinct lines intersect at a unique point, eliminating the parallel case and embodying the symmetry between points and lines. These properties form the foundation of incidence geometry in projective spaces.[1][5] The projective plane, as a fundamental model, extends the Euclidean plane by adjoining a line at infinity, which captures all directions and points where parallels converge. This construction compactifies the affine plane into a closed surface topologically equivalent to a sphere with antipodal points identified, preserving essential geometric relations while incorporating infinite elements seamlessly.[12][1]

Historical Development

Origins in Perspective Art and Optics

The origins of projective geometry can be traced to the Renaissance interest in accurately representing three-dimensional space on two-dimensional surfaces, particularly through the development of linear perspective in art. In the early 1420s, the Italian architect Filippo Brunelleschi conducted pioneering experiments in Florence to demonstrate the principles of linear perspective, using a peephole device and a mirror to view the Baptistery of San Giovanni, allowing viewers to see a painted scene align perfectly with the real architecture beyond.[14] These demonstrations emphasized the convergence of parallel lines to a vanishing point, a core idea that would later underpin projective transformations. Building on this, Leon Battista Alberti formalized these concepts in his 1435 treatise Della Pittura (On Painting), the first systematic theoretical work on artistic perspective in Europe, where he described the visual pyramid formed by rays from the eye to objects, enabling painters to construct scenes mathematically using intersecting lines and proportions.[15] Alberti's approach treated the canvas as a cross-section of this pyramid, introducing methods to scale figures based on their distance from the viewer, which influenced generations of artists and laid intuitive groundwork for handling projections.[16] Advancing into the realm of optics, Johannes Kepler provided a scientific foundation for understanding visual projections in his 1604 work Astronomiae Pars Optica. Kepler modeled the eye as a camera obscura, explaining how light rays from external objects form an inverted image on the retina through refraction in the eye's lenses, thus describing the projective nature of vision as a central projection onto a curved surface.[17] This optical theory shifted perspectives from geometric intuition in art to physiological and mathematical mechanisms, highlighting how projections preserve certain incidences and collinearities despite distortions, a principle central to later projective geometry.[18] Kepler's insights connected artistic representation to the physics of light, bridging empirical observation with projective mappings. The transition to explicit mathematical treatments occurred in the 17th century with contributions from French mathematicians inspired by these artistic and optical ideas. In 1639, Gérard Desargues published Brouillon Project d'une Atteinte aux Événements des Rencontres d'un Cône avec un Plan, a seminal but initially overlooked manuscript that introduced projective methods for studying conic sections through perspective and involutions, treating points at infinity uniformly without Euclidean metrics.[19] Desargues' work, circulated privately and not widely published until 1866, emphasized configurations invariant under projection, such as the alignment of intersection points in perspective drawings of triangles.[20] Complementing this, Blaise Pascal, at age 16, developed his theorem on conics in 1640 as part of his Essai pour les Coniques, proving that for any hexagon inscribed in a conic section, the intersections of opposite sides are collinear—a purely projective property independent of the conic's metric form.[21] Pascal's result, derived from Desargues' perspective techniques, demonstrated early recognition of projective invariants in algebraic curves, setting the stage for more rigorous 19th-century developments.[22]

19th-Century Foundations

The foundations of projective geometry as a rigorous mathematical discipline were laid in the early 19th century, building on earlier ideas from perspective but shifting toward abstract, metric-free properties. Jean-Victor Poncelet played a pivotal role with his 1822 publication, Traité des propriétés projectives des figures, which systematically explored projective properties of geometric figures through central projections and established the field on a synthetic basis.[23] In this work, Poncelet introduced the principle of continuity, positing that properties holding for degenerate cases (such as intersecting conics) extend continuously to general cases, enabling the treatment of imaginary elements as real in projective contexts and unifying disparate geometric configurations.[24] This principle allowed Poncelet to derive theorems on conic sections and polygons without relying on metric measurements, emphasizing invariance under projection.[23] Concurrent developments in the 1820s advanced duality concepts central to projective geometry. Joseph Gergonne, in his 1813 paper, articulated the pole-polar relation for conics, where a point (pole) corresponds to a line (polar) such that harmonic properties are preserved under projection, providing a foundational tool for studying reciprocal figures.[25] Poncelet independently developed similar ideas around the same period, integrating them into his projective framework to demonstrate how dual transformations maintain incidence relations between points and lines, thus revealing the symmetry inherent in projective spaces.[25] These relations, though sparking debates over priority, enriched the theory by enabling proofs of cross-ratio invariance and harmonic divisions without coordinates.[25] These ideas were further advanced by August Ferdinand Möbius, who in 1827 introduced barycentric coordinates in Der barycentrische Calcül, providing a coordinate system well-suited for projective geometry by treating points as weighted combinations without metrics, and by Julius Plücker, who in the 1830s and 1840s developed analytic methods for projective duality and introduced line coordinates to study higher-dimensional configurations.[26][27] By the mid-1840s, algebraic approaches began supporting synthetic methods. Hermann Grassmann's Die lineale Ausdehnungslehre (1844) introduced a calculus of extension that formalized multilinear operations on vectors, laying the groundwork for modern linear algebra as applied to projective geometry.[28] Grassmann's framework treated geometric objects through their extensive properties, such as outer products, which naturally encode projective transformations and incidence without metrics, influencing later vector-based treatments of projective spaces.[28] This work provided an abstract toolset for handling higher-dimensional extensions, bridging combinatorial aspects of geometry with algebraic structure.[28] Karl Georg Christian von Staudt further solidified the synthetic foundation in Geometrie der Lage (1847), constructing projective geometry entirely from incidence axioms without reference to distance or angles.[29] Von Staudt defined projective harmonic conjugates purely through complete quadrilaterals and extended this to coordinates via "throwing ratios," demonstrating that cross-ratios could be introduced synthetically to quantify projective invariants.[29] His approach proved the independence of projective geometry from Euclidean metrics, establishing it as a self-contained discipline capable of deriving theorems like Desargues' from basic axioms alone.[29] This text marked a culmination of early 19th-century efforts, emphasizing pure positional relations over analytic methods.[29]

20th-Century Advances

In the early 20th century, Oswald Veblen and John Wesley Young published their seminal two-volume textbook Projective Geometry (1910–1918), which established a rigorous axiomatic foundation for the subject, emphasizing postulates for points, lines, and planes while integrating synthetic methods with emerging algebraic insights. This work, developed during their collaboration at Princeton, became a cornerstone for advanced studies, influencing generations of geometers by providing a systematic treatment that avoided metric assumptions and focused on incidence and order properties.[30][31] Harold Scott MacDonald Coxeter advanced the synthetic tradition in his Projective Geometry (first edition 1964, building on earlier synthetic ideas from the 1930s onward), offering an intuitive yet formal approach centered on perspectivities and projectivities, with constructions limited to straightedge alone. Coxeter's text highlighted the elegance of projective configurations and polarity, making complex theorems accessible through visual and axiomatic clarity, and it reinforced the subject's independence from Euclidean metrics.[32] Emil Artin's Geometric Algebra (1957) bridged projective geometry with linear algebra by reformulating geometric concepts through vector spaces and bilinear forms, demonstrating how projective transformations arise naturally from linear mappings on homogeneous coordinates. This integration facilitated deeper connections to modern algebra, enabling projective geometry to inform developments in representation theory and quadratic forms.[33] Post-World War II, the rise of coding theory spurred significant advancements in finite projective geometries, as structures like projective planes over finite fields provided models for error-correcting codes, such as the Hamming code derived from the Fano plane (the projective plane of order 2). This interdisciplinary influence, beginning with Claude Shannon's foundational work in 1948 and Richard Hamming's codes in 1950, revitalized interest in non-Desarguesian planes and higher-dimensional finite spaces, leading to applications in combinatorial design and information theory.[34]

Fundamental Concepts

Points, Lines, and Incidence

In projective geometry, points and lines serve as the primitive elements of the theory, with no further definition provided beyond their mutual relations. The incidence relation denotes when a point lies on a line, forming the foundational structure known as an incidence geometry, where lines are distinguished subsets of points connected through this relation.[8] This synthetic approach treats points and lines as undefined terms, emphasizing their interdependencies rather than embedding them in a metric or coordinate system.[35] A projective plane is defined as an incidence structure consisting of a set of points, a set of lines, and the incidence relation satisfying three key axioms. First, any two distinct points are incident with exactly one line, ensuring that lines are uniquely determined by pairs of points. Second, any two distinct lines are incident with exactly one point, guaranteeing that lines always intersect. Third, there exist at least four points such that no three are incident with the same line, preventing degeneracy and allowing for non-trivial configurations.[35] These axioms imply the fundamental theorem that in a projective plane, any two distinct points determine a unique line; dually, any two distinct lines determine a unique point of intersection.[8] To extend the incidence structure with an ordering on lines, betweenness is introduced for ordered triples of collinear points, where one point is considered between the other two. Pasch's axiom provides the consistency for this ordering: given a triangle formed by three non-collinear points and a line that intersects one side of the triangle but passes through none of the vertices, the line must intersect exactly one of the other two sides. This axiom ensures that the ordering behaves coherently across the plane, distinguishing projective geometries with affine-like order from purely incidence-based ones.[36] The concept of the line at infinity arises in the construction of the projective plane as the closure of an affine plane. In this embedding, parallel lines from the affine plane, which do not intersect within the finite points, are made to meet at points on an added line at infinity, completing the structure to satisfy the projective axioms uniformly. This line at infinity comprises all ideal points corresponding to directions in the affine plane, with each such point representing the intersection of a pencil of parallel lines.[37]

Homogeneous Coordinates

Homogeneous coordinates provide the primary analytic framework for studying projective geometry, enabling the representation of points, lines, and transformations in a unified manner that incorporates points at infinity. In this system, points in the projective space RPn\mathbb{RP}^n are defined as equivalence classes of (n+1)(n+1)-tuples (x0,x1,…,xn)∈Rn+1∖{0}(x_0, x_1, \dots, x_n) \in \mathbb{R}^{n+1} \setminus \{\mathbf{0}\}, denoted by [x0:x1:⋯:xn][x_0 : x_1 : \dots : x_n], where two tuples are identified if one is a non-zero scalar multiple of the other, i.e., [x]=[λx][x] = [\lambda x] for λ∈R∖{0}\lambda \in \mathbb{R} \setminus \{0\}.[1] This construction arises from the quotient of the vector space Rn+1\mathbb{R}^{n+1} minus the origin by the action of scalar multiplication, effectively identifying each point with a one-dimensional subspace (a line through the origin).[7] The equivalence relation ensures scale invariance, which is essential for projective properties, as geometric incidences and transformations remain unchanged under rescaling. For instance, in RP2\mathbb{RP}^2, a point at infinity can be represented by [1:0:0][1 : 0 : 0], corresponding to the direction of the x-axis, without distinguishing between parallel lines in the affine plane.[1] To connect projective geometry with affine geometry, finite points in the affine space An\mathbb{A}^n are embedded into RPn\mathbb{RP}^n via homogenization: an affine point (a1,…,an)(a_1, \dots, a_n) maps to the projective point [a1:⋯:an:1][a_1 : \dots : a_n : 1].[7] This adds a hyperplane at infinity, defined by the equation xn+1=0x_{n+1} = 0, where ideal points reside, allowing parallel lines to intersect at infinity. The reverse process, dehomogenization, recovers affine coordinates from a projective point by normalizing the last coordinate to 1 (assuming xn+1≠0x_{n+1} \neq 0): (a1,…,an)=(x1/xn+1,…,xn/xn+1)(a_1, \dots, a_n) = (x_1 / x_{n+1}, \dots, x_n / x_{n+1}).[1] Normalization in general involves scaling the tuple so that a specific non-zero component equals 1, facilitating computations while preserving the equivalence class.[7] Lines in projective space are similarly represented using homogeneous coordinates. The line joining two distinct points [x][x] and [y][y] in RPn\mathbb{RP}^n (with x,y∈Rn+1x, y \in \mathbb{R}^{n+1} linearly independent) consists of all points [ax+by][a x + b y] for scalars a,b∈Ra, b \in \mathbb{R} not both zero, forming the one-dimensional projective subspace spanned by xx and yy.[1] A point [z][z] lies on this line if and only if zz is a linear combination of xx and yy, i.e., there exist scalars λ,μ\lambda, \mu not both zero such that z=λx+μyz = \lambda x + \mu y.[7] In the case of RP2\mathbb{RP}^2, this join operation corresponds to the cross product: the line through points p=[X1,Y1,W1]⊤p = [X_1, Y_1, W_1]^\top and q=[X2,Y2,W2]⊤q = [X_2, Y_2, W_2]^\top has homogeneous coordinates l=p×q=[Y1W2−Y2W1,W1X2−W2X1,X1Y2−X2Y1]⊤l = p \times q = [Y_1 W_2 - Y_2 W_1, W_1 X_2 - W_2 X_1, X_1 Y_2 - X_2 Y_1]^\top, up to scale.[7] This coordinate system allows projective transformations to be expressed as linear maps on Rn+1\mathbb{R}^{n+1}, preserving the equivalence classes and thus incidences between points and lines.[1]

Projective Spaces

In projective geometry, the real projective space of dimension nn, denoted RPn\mathbb{RP}^n, is defined as the set of all one-dimensional linear subspaces (lines through the origin) of the vector space Rn+1\mathbb{R}^{n+1}.[1] Each point in RPn\mathbb{RP}^n thus represents an equivalence class of nonzero vectors in Rn+1\mathbb{R}^{n+1} under scalar multiplication by nonzero reals, capturing all possible directions from the origin.[38] This construction generalizes the projective plane RP2\mathbb{RP}^2 (where n=2n=2) to arbitrary dimensions, providing a framework where parallel lines in lower-dimensional affine spaces meet at points at infinity.[39] Projective subspaces, also called flats, are the natural substructures within RPn\mathbb{RP}^n. A kk-dimensional projective subspace (or kk-flat) is the projectivization of a (k+1)(k+1)-dimensional linear subspace of Rn+1\mathbb{R}^{n+1}, consisting of all lines through the origin lying within that vector subspace.[1] For instance, a 0-flat is a single point, a 1-flat is a projective line (generalizing the lines of the projective plane), and a 2-flat is a projective plane embedded in RPn\mathbb{RP}^n. These flats inherit incidence relations from the underlying vector space, where two flats intersect in a flat of dimension at most the minimum of their dimensions, and their join (the smallest flat containing both) spans a higher-dimensional flat. Homogeneous coordinates [x0:⋯:xn][x_0 : \dots : x_n] from RPn\mathbb{RP}^n provide a coordinate representation for points and flats, as discussed previously.[40] A key property governing the structure of these subspaces is Grassmann's dimension formula, which relates the dimensions of subspaces, their intersection (meet), and their join. For any two projective subspaces S1S_1 and S2S_2 in RPn\mathbb{RP}^n, the dimension of their join L(S1,S2)L(S_1, S_2) (the smallest flat containing both) satisfies
dim⁡L(S1,S2)=dim⁡S1+dim⁡S2−dim⁡(S1∩S2), \dim L(S_1, S_2) = \dim S_1 + \dim S_2 - \dim(S_1 \cap S_2),
where the empty intersection has dimension −1-1. Equivalently, the dimension of the intersection (meet) is dim⁡(S1∩S2)=dim⁡S1+dim⁡S2−dim⁡L(S1,S2)\dim(S_1 \cap S_2) = \dim S_1 + \dim S_2 - \dim L(S_1, S_2). This formula ensures that the geometry is consistent across dimensions: if S1S_1 and S2S_2 are disjoint (dimension −1-1), their join has dimension dim⁡S1+dim⁡S2+1\dim S_1 + \dim S_2 + 1; if one contains the other, the join equals the larger subspace. It directly follows from the corresponding Grassmann relation in the ambient vector space Rn+1\mathbb{R}^{n+1}.[41] An illustrative example is the three-dimensional projective space RP3\mathbb{RP}^3, which acts as the projective closure of the affine 3-space R3\mathbb{R}^3. This closure is formed by embedding R3\mathbb{R}^3 as an affine hyperplane in RP3\mathbb{RP}^3 and adjoining a plane at infinity, a 2-flat comprising all projective points corresponding to parallel classes of lines in R3\mathbb{R}^3.[42] In this setup, any two skew lines in R3\mathbb{R}^3 (non-intersecting and non-parallel) meet at a unique point on the plane at infinity, while parallel lines share that same infinite point, unifying affine parallelism into projective incidence. Planes in RP3\mathbb{RP}^3 thus consist of an affine plane in R3\mathbb{R}^3 union a line at infinity, demonstrating how RP3\mathbb{RP}^3 resolves directional behaviors at infinity in three-dimensional geometry.[1]

Axiomatic Systems

Incidence and Parallelism Axioms

The incidence axioms of projective geometry establish the fundamental relationships between points and lines, diverging from Euclidean geometry by eliminating the concept of parallelism. In Euclidean geometry, Playfair's axiom asserts that given a line and a point not on it, there exists exactly one line through the point parallel to the given line. Projective geometry negates this by positing that every pair of distinct lines intersects in exactly one point, ensuring no parallels exist and unifying affine and "infinite" behaviors.[43] This axiom, often termed the elliptic parallel property, forms the cornerstone of projective incidence, allowing all lines to meet, either in the finite plane or at infinity.[30] A comprehensive axiomatic framework for these incidence relations was developed by Oswald Veblen and John Wesley Young in their seminal two-volume work on projective geometry. Their system treats points and lines as primitive elements, with axioms guaranteeing the existence and uniqueness of joins (lines connecting points) and meets (intersection points of lines). The core incidence axioms are:
  • Any two distinct points determine a unique line, known as the join of the points.
  • Any two distinct lines determine a unique point of intersection, known as the meet of the lines.
These ensure a symmetric duality between points and lines while preventing degenerate configurations. To support non-trivial structures, additional axioms specify existence: there exist at least three non-collinear points, ensuring the geometry is not merely linear.[30] The Veblen-Young axioms thus provide a minimal set for constructing projective spaces of arbitrary finite dimension greater than or equal to 2.[8] For geometries in dimensions higher than 2, the axiom of plane existence extends the incidence structure: any three non-collinear points determine a unique plane containing them. This axiom embeds lower-dimensional subspaces within the higher-dimensional space, allowing planes to be generated as the joins of lines or spans of points while maintaining uniqueness to avoid over-dimensioning. It is essential for defining projective 3-space and beyond, where planes serve as the basic hyper-surfaces. A key non-degeneracy condition in the Veblen-Young system is the eponymous axiom, which addresses the intersection behavior in quadrilateral configurations. It states: if points AA, BB, CC, DD are such that the line ABAB intersects the line CDCD at a point, then the line ACAC intersects the line BDBD at a point, and the line ADAD intersects the line BCBC at a point. This axiom, analogous to Pasch's axiom in ordered geometries but adapted for projective incidence, ensures that the space is connected and free of "gaps" in line intersections, facilitating the proof of higher theorems without invoking coordinates.[8][30] These incidence and parallelism-negating axioms enable synthetic proofs of central theorems, notably Desargues' theorem, which relates two triangles in perspective from a point. In a projective 3-space defined by the above axioms, the theorem for a plane follows synthetically: consider two triangles in distinct planes sharing a perspective axis; their vertices' joining lines concur at a point OO by the two-lines-one-point axiom, and the plane determined by three non-collinear vertices intersects the configuration such that the intersections of corresponding sides are collinear, using repeated applications of unique intersections and the plane axiom to trace the alignments without metric assumptions. This proof demonstrates the power of pure incidence in deriving plane properties from spatial structure, independent of order axioms.

Order and Continuity Axioms

In axiomatic systems for projective geometry, order axioms introduce a betweenness relation on the points of each line, enabling the distinction of ordered triples of collinear points without invoking metric concepts. This relation, denoted B(A, B, C) to indicate that B lies between A and C on a line, satisfies properties such as: for distinct points A, B, C on a line, exactly one of B(A, B, C), B(B, A, C), or B(C, A, B) holds; if B(A, B, C), then B(A, C, B) is false; and transitivity holds, so if B(A, B, D) and B(B, C, D), then B(A, C, D). These axioms adapt Hilbert's order group by focusing on projective incidence structures, omitting parallelism to maintain uniformity across lines.[44] A key plane axiom of order in projective geometry is Pasch's axiom, which serves as the projective counterpart to the affine or Euclidean version. It states that if a line intersects two sides of a triangle but does not pass through any vertex, then it must intersect the third side. This ensures consistent separation properties in the plane, preventing pathological configurations and supporting the development of ordered projective spaces. The axiom, introduced in foundational treatments of geometry, guarantees that the betweenness relation extends coherently from lines to planar figures.[44] Continuity axioms further refine the structure by imposing completeness on the ordered lines, particularly in the real projective case. For the real projective line RP1\mathbb{RP}^1, which is topologically a circle but coordinatized by the real projective field, continuity is achieved through the Dedekind completeness of the underlying real numbers: every non-empty subset of points bounded above has a least upper bound. Alternatively, the Archimedean property ensures that for any two positive elements, there exists a natural number multiple exceeding the other, preventing infinitesimal gaps. These properties, adapted from Hilbert's continuity group (removing metric dependencies), ensure that the real projective geometry is "complete" without discrete interruptions.[44] By incorporating these order and continuity axioms atop incidence structures, real projective geometry embeds Euclidean geometry as a substructure: selecting a line as the "line at infinity" yields an affine plane isomorphic to the Euclidean plane, where betweenness aligns with the standard order on R2\mathbb{R}^2. This embedding preserves all projective properties while recovering metric interpretations via additional congruence axioms if desired.

Axioms for Projective Planes

A projective plane is defined by a set of points and lines satisfying the basic incidence axioms: any two distinct points determine a unique line, any two distinct lines intersect in a unique point, and there exist four points no three of which are collinear.[45] These axioms, originally formalized in the context of Hilbert's foundational work on geometry, provide the minimal structure for a projective plane without additional ordering or metric assumptions.[46] To achieve coordinatization, Hilbert's incidence axioms are supplemented with Desargues' theorem, which states that if two triangles are perspective from a point (corresponding vertices joined by lines concurrent at that point), then they are perspective from a line (intersections of corresponding sides are collinear).[47] A projective plane satisfying these axioms (a Desarguesian plane) can be coordinatized by a division ring, where points are represented as equivalence classes of triples from the division ring under scalar multiplication, and lines as linear equations in homogeneous coordinates.[47] Further imposing Pappus' theorem—that the intersections of opposite sides of a hexagon inscribed in two lines are collinear—ensures the coordinating structure is a commutative field, yielding a Pappian plane isomorphic to the projective plane over that field.[48] For more general projective planes not assuming Desargues' theorem, coordinatization proceeds via a planar ternary ring, an algebraic structure consisting of a set RR with a ternary operation t:R×R×R→Rt: R \times R \times R \to R satisfying specific axioms: there exist distinct elements 0 and 1 in RR such that t(0,a,b)=t(a,0,b)=bt(0, a, b) = t(a, 0, b) = b and t(1,a,0)=t(a,1,0)=at(1, a, 0) = t(a, 1, 0) = a; for a≠ca \neq c, the map x↦(t(x,a,b),t(x,c,d))x \mapsto (t(x, a, b), t(x, c, d)) is bijective; for fixed a,ba, b, the map x↦t(a,b,x)x \mapsto t(a, b, x) is bijective; and for a≠ca \neq c, given b,db, d, there is a unique pair (x,y)(x, y) such that t(a,x,y)=bt(a, x, y) = b and t(c,x,y)=dt(c, x, y) = d.[49] This structure, developed in the tradition of von Staudt's early coordinatization efforts, allows construction of the plane's points as ordered pairs from RR plus points at infinity, with lines defined using the ternary operation, satisfying the incidence axioms without requiring a division ring.[50] The existence of non-Desarguesian planes demonstrates that Desargues' theorem is independent of the incidence axioms. A seminal example is the Moulton plane, constructed in 1902 by modifying the Euclidean plane: points are pairs of real numbers, vertical lines remain as in the Euclidean plane, but non-vertical lines have slope m/2m/2 when x≤0x \leq 0 (left of the y-axis) and slope mm when x>0x > 0, for any slope m≠0m \neq 0. This alteration preserves incidence but violates Desargues' theorem, as certain perspective triangles fail to have collinear side intersections. The theorem characterizing Desarguesian planes states that a projective plane is Desarguesian if and only if it is coordinatizable by a division ring. The forward direction follows from embedding the plane into a higher-dimensional space or direct algebraic verification using homogeneous coordinates; the converse constructs the division ring from the plane's ternary operations, verifying field-like properties via Desargues' configurations.[47]

Transformations and Invariants

Perspectivities

A perspectivity is a bijective mapping between two distinct lines (or planes) in a projective space, defined as a central projection from a fixed center point not lying on either line (or plane). Specifically, for two lines $ l $ and $ l' $ in the projective plane, a perspectivity with center $ O $ maps a point $ A $ on $ l $ to the point $ A' $ on $ l' $ such that the line $ AA' $ passes through $ O $.[1][51] This construction extends naturally to higher dimensions, where it projects between hyperplanes via lines through the center.[1] Perspectivities preserve key incidence relations, including collinearity of points and concurrence of lines. If three points are collinear on the source line, their images remain collinear on the target line, as the projection rays from the center maintain the relative alignments. Similarly, concurrent lines map to concurrent lines, ensuring that the intersection structure is invariant under the mapping. These properties follow from the linear nature of the projection and underpin theorems like Desargues's theorem, where two triangles related by a perspectivity have collinear side intersections.[51][1] A chain of perspectivities refers to the composition of multiple such mappings along a sequence of lines or planes, where each subsequent projection shares an appropriate center and axis with the previous one. Such compositions are bijective and form the foundational building blocks for more general projective transformations, with any projectivity between two distinct lines expressible as a chain of at most two perspectivities.[51][1] In the real projective plane $ \mathbb{RP}^2 $, a classic example of a perspectivity arises in perspective drawing: consider two parallel lines $ l $ and $ l' $ (which intersect at a point at infinity), and a center $ O $ not on either. The mapping sends points on $ l $ to $ l' $ along rays through $ O $, effectively simulating the convergence of parallel lines toward a vanishing point, thus unifying affine and projective views.[1]

Projectivities

In projective geometry, a projectivity is defined as a bijective mapping between two projective spaces (or between pencils or ranges within them) that preserves the incidence relation between points and lines, meaning collinear points map to collinear points and vice versa.[51] This transformation is equivalently characterized as a finite composition of perspectivities, which are special projective mappings with a fixed center of projection.[1] Projectivities form the general collineation group of the projective space, acting transitively on ordered sets of points in general position, such as complete quadrangles. In homogeneous coordinates, a projectivity on the real projective space RPn\mathbb{RP}^n is represented by an invertible linear transformation on the underlying vector space Rn+1\mathbb{R}^{n+1}. Specifically, it acts on a point represented by the equivalence class [x]=[x0:x1:⋯:xn][x] = [x_0 : x_1 : \dots : x_n], where x∈Rn+1∖{0}x \in \mathbb{R}^{n+1} \setminus \{0\} and scalar multiples are identified, via [x]↦[Ax][x] \mapsto [Ax], with AA an (n+1)×(n+1)(n+1) \times (n+1) invertible matrix from GL(n+1,R)\mathrm{GL}(n+1, \mathbb{R}).[1] The group of all such projectivities is the projective linear group PGL(n+1,R)\mathrm{PGL}(n+1, \mathbb{R}), obtained as the quotient GL(n+1,R)/R×I\mathrm{GL}(n+1, \mathbb{R}) / \mathbb{R}^\times I, where scalar matrices act trivially.[51] A key result is the fundamental theorem of projective geometry, which states that for a projective space of dimension at least 2 over a field KK, any bijective collineation (incidence-preserving map) is a projectivity induced by a semilinear transformation on the coordinate space.[52] In the real case, this simplifies to projectivities being precisely the elements of PGL(n+1,R)\mathrm{PGL}(n+1, \mathbb{R}). Moreover, every projectivity can be expressed as a composition of perspectivities, and there exists a unique projectivity mapping any complete quadrangle (four points, no three collinear) to any other.[51] Projectivities are classified in part by their fixed points, which correspond to eigenvectors of the representing matrix AA: a point [x][x] is fixed if Ax=λxAx = \lambda x for some scalar λ≠0\lambda \neq 0. If a projectivity fixes three distinct points on a line, it must be the identity on that line.[51] A notable subclass is the harmonic homology, an involutory projectivity (H2=idH^2 = \mathrm{id}) with a center OO (a fixed point) and an axis ℓ\ell (a fixed line not containing OO). It fixes all points on ℓ\ell pointwise and maps any other point XX to its harmonic conjugate YY with respect to OO and the intersection HX=[OX]∩ℓH_X = [OX] \cap \ell, such that the cross-ratio (O,HX;X,Y)=−1(O, H_X; X, Y) = -1.[53] This transformation, with eigenvalue −1-1 on the axis complement, exemplifies projectivities preserving harmonic divisions.[1]

Cross-Ratio

The cross-ratio provides the fundamental invariant for the configuration of four collinear points in projective geometry, determining their relative positions up to projective transformations. For four distinct points A,B,C,DA, B, C, D on a line with affine coordinates a,b,c,da, b, c, d, the cross-ratio is defined as
(A,B;C,D)=(c−a)/(c−b)(d−a)/(d−b)=(c−a)(d−b)(c−b)(d−a). (A, B; C, D) = \frac{(c - a)/(c - b)}{(d - a)/(d - b)} = \frac{(c - a)(d - b)}{(c - b)(d - a)}.
This expression uses directed distances and remains well-defined in the projective setting, where the line is RP1\mathbb{RP}^1, by incorporating the point at infinity; if any denominator vanishes, the value is ∞\infty. In homogeneous coordinates [xi:yi][x_i : y_i] for points PiP_i, the cross-ratio extends to
(A,B;C,D)=det⁡(xAxByAyB)det⁡(xCxDyCyD)det⁡(xBxCyByC)det⁡(xDxAyDyA). (A, B; C, D) = \frac{\det\begin{pmatrix} x_A & x_B \\ y_A & y_B \end{pmatrix} \det\begin{pmatrix} x_C & x_D \\ y_C & y_D \end{pmatrix}}{\det\begin{pmatrix} x_B & x_C \\ y_B & y_C \end{pmatrix} \det\begin{pmatrix} x_D & x_A \\ y_D & y_A \end{pmatrix}}.
This determinant form ensures invariance and handles projective completions directly.[54][55] A defining property of the cross-ratio is its preservation under projectivities, the collineations of the projective line isomorphic to PGL(2,R)\mathrm{PGL}(2, \mathbb{R}), making it the unique (up to the action of this group) complete invariant for four points. The value depends on the ordering of the points: among the 24 possible permutations, only six distinct cross-ratios arise, related by the transformations q↦1−qq \mapsto 1-q, q↦1/qq \mapsto 1/q, and q↦q/(q−1)q \mapsto q/(q-1), where qq is the original value; the cases q=0,1,∞q = 0, 1, \infty correspond to degenerate configurations where points coincide projectively.[54][55] When the cross-ratio equals −1-1, the four points form a harmonic division (or harmonic set), a configuration central to projective constructions. In this case, the points CC and DD are harmonic conjugates with respect to AA and BB, meaning that the complete quadrilateral formed by lines joining them has diagonals intersecting at a point that "balances" the division projectively; an example is the points at 0,∞,1,−10, \infty, 1, -1 in affine coordinates. This property is constructible using ruler and compass in the plane and characterizes many self-dual figures.[54][55] The cross-ratio generalizes to higher-dimensional projective spaces, where it applies to configurations involving simplices through coordinate systems and hyperplane pencils. In FPn\mathbb{FP}^n, a coordinate simplex with vertices B0,…,BnB_0, \dots, B_n defines homogeneous coordinates for points via barycentric combinations, and the cross-ratio extends to four hyperplanes (each a facet-like simplex) containing a common (n−2)(n-2)-plane by projecting to a transversal line and using the 1D definition; the invariance holds analogously, with the value determining whether a fourth hyperplane passes through the projected point. This framework supports volume-based interpretations for simplices, where signed determinants of vertex coordinates yield ratios mirroring the 1D case.[56]

Duality and Polarity

Principle of Duality

The principle of duality in projective geometry asserts that points and lines (or points and planes in higher dimensions) play symmetric roles with respect to incidence relations, allowing any theorem formulated in terms of these elements to be dualized by interchanging "point" with "line" (or "plane"), resulting in an equally valid dual theorem. This symmetry stems from the axiomatic foundations of projective spaces, where the undefined primitives—points and lines—are treated equivalently in the incidence axioms, without privileging one over the other. The principle was first stated in full generality by Joseph Gergonne in 1825–1826, following foundational contributions by Jean-Victor Poncelet in 1817–1818 that emphasized the role of reciprocation in projective transformations.[57][58] A fundamental example is the dualization of the incidence statement "two distinct points lie on a unique line" to "two distinct lines intersect in a unique point," both of which hold true in any projective plane. Another illustration is the theorem "three non-collinear points determine a triangle," whose dual is "three lines in general position (no two parallel, no three concurrent) determine a triangle," preserving the geometric configuration through the interchange. These dual pairs demonstrate how the principle generates corresponding propositions without altering their validity, relying solely on the self-dual nature of the incidence structure.[57][59] The duality principle is inherently preserved under projectivities, the fundamental transformations of projective geometry that map lines to lines and points to points while maintaining collinearity and concurrence. Specifically, a projectivity on the space of points induces a corresponding dual projectivity on the space of lines, ensuring that incidence relations are mirrored exactly in the dual setting. This compatibility underscores the principle's deep integration with the group's action on the geometry, allowing dual maps to exist naturally without additional assumptions.[59][57] While powerful in standard projective spaces, the principle holds primarily in self-dual axiomatic systems focused on unoriented incidence and may fail in non-self-dual settings, such as oriented projective geometries, where the interchange of points and lines disrupts orientation-dependent properties like signed angles or directed incidences. In these extensions, duality requires modifications to account for the added structure, preventing direct application of the classical principle.[60]

Polar Reciprocity

In projective geometry, a polarity is defined as a bijective correlation between the points and hyperplanes of a projective space that reverses incidence and is an involution, meaning applying the map twice yields the identity. This structure provides a concrete realization of duality, mapping points to hyperplanes and vice versa while preserving the relational properties of the geometry.[61][62] Polarities are commonly induced by non-degenerate quadrics, which are hypersurfaces defined by quadratic forms. In a projective space Pn\mathbb{P}^n arising from a vector space VV over a field FF (with char(F)≠2\mathrm{char}(F) \neq 2), a symmetric bilinear form B:V×V→FB: V \times V \to F defines the polarity as follows: for a point [x]∈Pn[x] \in \mathbb{P}^n (the projective point corresponding to the line spanned by x∈Vx \in V), its polar hyperplane is the set of points [y][y] such that B(x,y)=0B(x, y) = 0. The associated quadratic form is q(v)=B(v,v)q(v) = B(v, v), and the quadric is the set {[v]∣q(v)=0}\{ [v] \mid q(v) = 0 \}. This construction ensures the map is an involution because BB is symmetric and non-degenerate, so the orthogonal complement operation satisfies (U⊥)⊥=U(U^\perp)^\perp = U for subspaces U⊆VU \subseteq V.[61] Key properties of such polarities include the existence of absolute points, which are self-polar points satisfying B(x,x)=0B(x, x) = 0 and thus lying on the quadric. These points play a central role in classifying the polarity (e.g., hyperbolic if there are isotropic subspaces of maximal dimension). Additionally, the polar of a subspace UU is its orthogonal complement U⊥={v∈V∣B(v,u)=0 ∀u∈U}U^\perp = \{ v \in V \mid B(v, u) = 0 \ \forall u \in U \}, which inherits the lattice structure of the original space under the duality. Non-degeneracy ensures the radicals (kernels where BB vanishes entirely) are trivial, preserving the bijection between points and hyperplanes.[61] A representative example occurs in the real projective plane RP2\mathbb{RP}^2, where the polarity with respect to the unit circle—defined by the quadric x2+y2−z2=0x^2 + y^2 - z^2 = 0 in homogeneous coordinates [x:y:z][x : y : z]—arises from the bilinear form B((x1,y1,z1),(x2,y2,z2))=x1x2+y1y2−z1z2B((x_1, y_1, z_1), (x_2, y_2, z_2)) = x_1 x_2 + y_1 y_2 - z_1 z_2. The polar line of a point [x0:y0:z0][x_0 : y_0 : z_0] is given by the equation
(x0y0−z0)(xyz)=0, \begin{pmatrix} x_0 & y_0 & -z_0 \end{pmatrix} \begin{pmatrix} x \\ y \\ z \end{pmatrix} = 0,
or x0x+y0y−z0z=0x_0 x + y_0 y - z_0 z = 0. For a point outside the circle (in the affine view where z=1z=1), this polar is the line joining the points of tangency from that point to the circle; points on the circle are self-polar, with their polars being the tangents at those points. This conic-based polarity exemplifies how quadrics (here, a circle as a special quadric) generate the reciprocal relations central to projective duality.[62]

Self-Dual Configurations

In projective geometry, a self-dual configuration is an incidence structure of points and lines that remains unchanged under the principle of duality, featuring a one-to-one correspondence between points and lines that preserves incidence relations. This isomorphism to its dual implies an equal number of points and lines, with symmetric incidence properties such that the number of lines through each point matches the number of points on each corresponding line.[63][64] A fundamental example involves the complete quadrilateral and its dual, the complete quadrangle. The complete quadrilateral comprises 4 lines in general position (no three concurrent) and their 6 intersection points, with each line containing 3 points and each point incident to 2 lines, denoted as a (6_2 4_3) configuration. Its dual, the complete quadrangle, consists of 4 points in general position (no three collinear) and the 6 lines joining them pairwise, forming a (4_3 6_2) configuration. While these are not individually self-dual due to differing counts of points and lines, their mutual duality exemplifies the interchangeability central to self-dual structures.[65] The Desargues configuration provides a prominent self-dual example, denoted as 10_3, with 10 points and 10 lines where each line passes through 3 points and each point lies on 3 lines. This configuration emerges from Desargues' theorem, which equates central and axial perspectivities between triangles, and its self-duality follows directly from the theorem's symmetric statement under duality.[65] A key theorem states that in a self-dual configuration admitting a polarity under which the configuration is self-polar (invariant under the polarity map), the incidence matrix—whose rows and columns represent points and lines with entries indicating incidence—can be labeled such that it is symmetric. This symmetry arises because the polarity induces a bijection that equates point-line incidences with line-point incidences, reflecting the configuration's invariance.[64]

Classifications and Models

Real Projective Geometry

The real projective space RPn\mathbb{RP}^n serves as the canonical model for projective geometry over the field of real numbers R\mathbb{R}. It consists of all one-dimensional subspaces (lines through the origin) of the vector space Rn+1\mathbb{R}^{n+1}, where each point in RPn\mathbb{RP}^n corresponds to such a line. This construction endows RPn\mathbb{RP}^n with a natural smooth manifold structure of dimension nn.[1] Topologically, RPn\mathbb{RP}^n is obtained as the quotient space Sn/∼S^n / \sim, where SnS^n is the nn-dimensional sphere in Rn+1\mathbb{R}^{n+1} and ∼\sim identifies each point xx with its antipode −x-x. This identification yields a compact, connected Hausdorff manifold that is second-countable. The space RPn\mathbb{RP}^n is orientable if and only if nn is odd; for even nn, it is non-orientable, as evidenced by its double cover Sn→RPnS^n \to \mathbb{RP}^n being orientation-reversing along the non-trivial deck transformation.[66][66] The affine Euclidean space Rn\mathbb{R}^n embeds naturally into RPn\mathbb{RP}^n via the standard affine chart, where points [x0:⋯:xn][x_0 : \dots : x_n] with xn≠0x_n \neq 0 are identified with (x0/xn,…,xn−1/xn)∈Rn(x_0/x_n, \dots, x_{n-1}/x_n) \in \mathbb{R}^n. The complement RPn∖RP∞n−1≅Rn\mathbb{RP}^n \setminus \mathbb{RP}^{n-1}_\infty \cong \mathbb{R}^n, where RP∞n−1\mathbb{RP}^{n-1}_\infty is the hyperplane at infinity consisting of points with xn=0x_n = 0. This embedding realizes projective geometry as a compactification of affine geometry, adding the hyperplane at infinity to handle parallel lines and points at infinity uniformly.[1] In the real projective plane RP2\mathbb{RP}^2, ovals are closed curves homeomorphic to the circle S1S^1 that intersect every projective line in at most two points. A key theorem establishes that every oval in RP2\mathbb{RP}^2 is the locus of real points of a non-degenerate conic, defined by a quadratic form ax2+by2+cz2+dxy+exz+fyz=0ax^2 + by^2 + cz^2 + dxy + exz + fyz = 0 with non-vanishing discriminant. Such conics, when possessing real points, form a single connected oval, as their real loci are topologically circles in the projective closure; examples include projectivizations of ellipses and hyperbolas, which become equivalent under projective transformations. Hyperovals, which extend ovals by adding two nuclei in finite even-order planes, do not arise in the real case due to the continuous nature of R\mathbb{R}, though conics provide the maximal analogues, intersecting each line in exactly two points (counting multiplicity).[67][1][68]

Finite Projective Geometries

Finite projective geometries, denoted as PG(n, q), are incidence structures defined over the finite field GF(q), where q = p^k for a prime p and positive integer k. These geometries arise from the (n+1)-dimensional vector space V(n+1, q) over GF(q), with points corresponding to the 1-dimensional subspaces and lines to the 2-dimensional subspaces of V. The notation PG(n, q) was introduced by Oswald Veblen and William H. Bussey in their foundational 1906 paper, which systematically developed these finite analogs of classical projective spaces.[69][70] A fundamental property of PG(n, q) is that the total number of points is given by the formula qn+1−1q−1\frac{q^{n+1} - 1}{q - 1}, which counts the distinct directions (or rays) in the vector space. Every pair of distinct points in PG(n, q) lies on a unique line, and every pair of distinct lines intersects in a unique point, ensuring a highly symmetric structure without parallel lines. These geometries are inherently Desarguesian, meaning they satisfy Desargues' theorem—perspective triangles from a point have their intersection points collinear—due to their construction from a field.[70] The smallest non-trivial example is the Fano plane, PG(2, 2), constructed over the field with two elements. It features 7 points and 7 lines, with each line containing 3 points and each point incident to 3 lines, illustrating the basic incidence relations in a compact form. Finite projective planes of order q, such as PG(2, q), relate to affine planes of the same order through the addition of a "line at infinity," which eliminates parallels by connecting all directions at infinity.[70][71] In design theory, finite projective planes provide prototypical examples of symmetric 2-(v, k, 1) designs, where the points form a set of size v = q^2 + q + 1, each block (line) has k = q + 1 points, and every pair of distinct points appears in exactly one block. This combinatorial interpretation underscores their utility in constructing balanced incomplete block designs (BIBDs) with λ = 1, facilitating applications in statistical experimentation and coding.[71][72]

Non-Desarguesian Planes

Non-Desarguesian projective planes are those that do not satisfy Desargues' theorem, meaning they cannot be coordinatized by a division ring (or skew-field).[73] These planes arise in both infinite and finite settings, highlighting pathologies in axiomatic projective geometry where standard theorems fail. Unlike Desarguesian planes, which embed into higher-dimensional spaces over fields, non-Desarguesian examples rely on alternative algebraic structures like near-fields or semifields.[74] A classic infinite example is the Moulton plane, originally an affine plane constructed by modifying the Euclidean plane: lines with positive slope remain straight, while those with negative slope bend at the y-axis, doubling their slope below it. This modification preserves affine axioms but violates Desargues' theorem due to the altered parallelism and intersection properties. The projective completion of the Moulton plane, obtained by adding points at infinity, yields a non-Desarguesian projective plane that is a Baer plane, satisfying a restricted form of Desargues but not the full theorem.[75] Finite non-Desarguesian planes include Hughes planes, constructed in 1957 using Dickson near-fields of rank 2 over their kernel sub-division ring. In a Hughes plane of order q2q^2 (where qq is a prime power), points and lines are defined via the near-field's addition and multiplication, with lines as cosets of subgroups; the non-associativity of the near-field multiplication ensures failure of Desargues' theorem, distinguishing these planes from Desarguesian ones over finite fields. Hughes planes are not translation planes and exhibit unique collineation groups.[76] Hall planes, introduced by Marshall Hall Jr. in 1943, form another family of finite non-Desarguesian planes, constructed as translation planes using ternary rings derived from finite fields with modified multiplication tables. For order qnq^n ( qq prime power, n>1n > 1), a Hall plane replaces the field multiplication with a non-associative operation on a vector space over Fq\mathbb{F}_q, ensuring the plane admits a transitive translation group but fails Desargues due to the underlying ring not being a division ring. More generally, translation planes arise from semifields—non-associative division-like structures—where the semifield coordinates the affine plane, and its projective extension is non-Desarguesian unless the semifield is a field.[73] All known non-Desarguesian projective planes of order n>2n > 2 are non-Arguesian, meaning they fail the Argues configuration (a higher-order incidence theorem implying Desargues in Desarguesian planes). This property underscores the rarity of Arguesian non-Desarguesian planes, with no examples known beyond order 2.[77]

Applications and Extensions

In Algebraic Geometry

In algebraic geometry, projective varieties serve as the primary objects for studying the global properties of algebraic sets, defined as the common zero loci of a finite collection of homogeneous polynomials in the projective space Pkn\mathbb{P}^n_k over an algebraically closed field kk. Unlike affine varieties, which may exhibit non-compact behavior, projective varieties are compact in the classical topology and incorporate points at infinity, enabling a unified treatment of asymptotic phenomena and intersections. This definition ensures that projective varieties are closed subsets of Pkn\mathbb{P}^n_k invariant under scalar multiplication, facilitating the use of homogeneous coordinates to describe them rigorously.[78][79] A central method for embedding affine varieties into the projective setting is homogenization, which transforms non-homogeneous polynomials from affine space An\mathbb{A}^n into homogeneous ones by introducing a new variable zz and multiplying terms by appropriate powers of zz to equalize degrees. For example, the affine hyperbola defined by xy=1xy = 1 in A2\mathbb{A}^2 homogenizes to the equation xy−z2=0xy - z^2 = 0 in P2\mathbb{P}^2, where the line at infinity z=0z=0 intersects the curve at the point [1:0:0][1:0:0] and [0:1:0][0:1:0], completing the affine picture. This process yields the projective closure of the affine variety, preserving algebraic structure while adding infinite points, and is essential for applying projective techniques to affine problems.[80][81] Bézout's theorem exemplifies the power of projective geometry in intersection theory, stating that two plane curves of degrees dd and ee in P2\mathbb{P}^2, with no common irreducible component, intersect in exactly dede points, counted with appropriate multiplicity. This result relies on the compactness of projective space to guarantee all intersections are finite and accounts for multiplicities via the local ring at intersection points, providing a foundational tool for enumerative geometry. In higher dimensions, generalizations to complete intersections maintain this degree-based counting, highlighting projective space's role in resolving affine limitations.[82][83] In modern algebraic geometry, particularly within scheme theory, the Proj construction formalizes projective varieties by associating to a finitely generated graded commutative ring SS over a base ring the scheme Proj(S)\mathrm{Proj}(S), which consists of homogeneous prime ideals not containing the irrelevant ideal (S+)(S_+). This functorial approach extends classical projective varieties to schemes, accommodating non-reduced structures, torsion, and relative situations over arbitrary base schemes, and underpins advanced topics like moduli spaces and cohomology. The construction is detailed in Hartshorne's foundational monograph, where it unifies affine and projective schemes via the distinguished open cover by Spec(S(f))\mathrm{Spec}(S_{(f)}) for homogeneous elements ff.[84][85]

In Computer Vision and Graphics

In computer vision, the pinhole camera model serves as the foundational representation of image formation, modeling the projection of three-dimensional world points onto a two-dimensional image plane as a projective transformation. This model assumes an ideal point-like aperture through which light rays pass without distortion, mapping a 3D point X=(X,Y,Z)T\mathbf{X} = (X, Y, Z)^T in homogeneous coordinates to a 2D image point x=(x,y,1)T\mathbf{x} = (x, y, 1)^T via a 3×4 projection matrix P\mathbf{P}, such that x∝PX\mathbf{x} \propto \mathbf{P} \mathbf{X}. The intrinsic parameters (focal length, principal point) and extrinsic parameters (rotation and translation) are encoded in P\mathbf{P}, enabling the handling of perspective effects inherent to projective geometry.[86] Homographies extend this framework to planar scenes or image-to-image mappings, represented by a 3×3 invertible matrix H\mathbf{H} that induces a projectivity between two planes: x′∝Hx\mathbf{x}' \propto \mathbf{H} \mathbf{x}. In vision tasks, homographies arise when projecting a planar surface from one view to another, preserving straight lines and incidence relations while allowing for perspective distortion. This matrix has 8 degrees of freedom (up to scale), estimated from at least four point correspondences using linear methods like the direct linear transformation (DLT). Homographies are crucial for rectifying images or aligning features in scenarios where depth variations are negligible.[87][88] Epipolar geometry captures the projective relationship between two uncalibrated views, constraining corresponding points to lie on epipolar lines defined by the fundamental matrix F\mathbf{F}, a 3×3 matrix of rank 2 satisfying x′TFx=0\mathbf{x}'^T \mathbf{F} \mathbf{x} = 0. This matrix encodes the essential projective structure, with its null space revealing the epipole (projection of one camera center onto the other image). Derived from the relative pose and intrinsics of the cameras, F\mathbf{F} has 7 degrees of freedom and is computed from at least eight point correspondences via eigendecomposition or iterative optimization. Epipolar constraints reduce the search space for stereo matching and facilitate 3D reconstruction by imposing geometric consistency across views.[87][88] These concepts underpin key applications in computer vision and graphics. In panorama stitching, homographies align overlapping images captured from a rotating camera, warping them into a cohesive wide-field view by estimating H\mathbf{H} from feature matches (e.g., SIFT keypoints); this projective alignment handles parallax minimally for near-planar scenes, enabling seamless blending. For augmented reality calibration, the DLT algorithm estimates camera parameters from known 3D-2D correspondences, solving P\mathbf{P} via singular value decomposition to register virtual overlays onto real environments with sub-pixel accuracy in controlled setups. Such methods ensure robust pose estimation, vital for real-time AR systems like head-mounted displays.

Connections to Other Geometries

Projective geometry provides a unifying framework for various classical geometries by embedding them as subgeometries, where affine, hyperbolic, elliptic, and conformal structures arise through specific choices of metrics or subspaces within projective space.[89] Affine geometry emerges from projective geometry by removing the line at infinity, which consists of all ideal points corresponding to directions of parallel lines in the affine plane. In this construction, parallel lines in the affine plane intersect at points on the line at infinity in the full projective plane, thereby recovering the parallel postulate absent in pure projective geometry. This embedding allows affine transformations to be viewed as projective transformations that preserve the line at infinity.[90] Hyperbolic and elliptic geometries are realized within projective geometry through the Beltrami-Klein model, where the geometry is represented inside a projective plane bounded by a conic section serving as the "line at infinity." In the hyperbolic case, points lie inside the conic disk, with straight lines as chords, and distances defined via the projective metric induced by the conic, ensuring hyperbolic parallelism. For elliptic geometry, the model covers the entire projective plane, with the conic identifying antipodal points to form a metric of constant positive curvature. This projective embedding, developed by Eugenio Beltrami and Felix Klein, demonstrates the consistency of non-Euclidean geometries by reducing them to projective properties. Conformal geometry connects to projective geometry via Möbius transformations, which preserve angles and can be interpreted as projective transformations in one higher dimension. Specifically, Möbius transformations on the Riemann sphere correspond to linear fractional transformations in the complex projective line, but in three dimensions, they act as projective maps on the sphere embedded in projective 3-space, mapping circles to circles and preserving the conformal structure. This higher-dimensional projective realization unifies conformal mappings with projective invariance.[91] A fundamental result states that all classical geometries—affine, Euclidean, hyperbolic, elliptic, and conformal—can be embedded as subgeometries of projective geometry, with their transformation groups as subgroups of the projective linear group. This embedding theorem highlights projective geometry's role as a foundational structure encompassing these geometries through appropriate choices of subspaces and metrics.[89]

References

  1. [PDF] Basics of Projective Geometry - UPenn CIS
  2. None
  3. [PDF] Projective Geometry - Fall 2022 - R. L. Herman - UNCW
  4. [PDF] Projective Geometry - Georgia College & State University
  5. Key Concepts and Properties of Projective Geometry
  6. [PDF] An Introduction to Projective Geometry for computer vision
  7. [PDF] PROJECTIVE GEOMETRY Contents 1. Basic Definitions 1 2. Axioms ...
  8. https://www.sciencedirect.com/science/article/pii/B9780444515605500166
  9. Projective Geometry | SpringerLink
  10. https://www.sciencedirect.com/science/article/pii/B9780444511041500034
  11. [PDF] An Introduction to Projective Geometry (for computer vision)
  12. [PDF] math 145 notes - Arun Debray
  13. Linear Perspective: Brunelleschi's Experiment - Smarthistory
  14. Alberti's revolution in painting - Smarthistory
  15. The History of Perspective - Essential Vermeer
  16. Kepler's Optical Part of Astronomy (1604) - MIT Press Direct
  17. Kepler's Optical Part of Astronomy (1604): Introducing the Ecliptic ...
  18. [PDF] Desargues' Brouillon Project and the Conics of Apollonius
  19. Girard Desargues - Biography - MacTutor - University of St Andrews
  20. Blaise Pascal - Biography - MacTutor - University of St Andrews
  21. Pascal's mystic hexagram, and a conjectural restoration of his lost ...
  22. Mathematical Treasure: Poncelet's Projective Geometry
  23. Jean Victor Poncelet, Traité des propriétés projectives des figures ...
  24. Polemics in Public: Poncelet, Gergonne, Plücker, and the Duality ...
  25. Hermann Grassmann (1809 - 1877) - Biography - MacTutor
  26. Karl von Staudt - Biography - MacTutor - University of St Andrews
  27. Projective geometry : Veblen, Oswald, 1880 - Internet Archive
  28. [PDF] The Vision, Insight, and Influence of Oswald Veblen, Volume 54 ...
  29. Projective Geometry - H.S.M. Coxeter - Google Books
  30. Geometric Algebra : Artin,E. : Free Download, Borrow, and Streaming
  31. [PDF] Error Correcting Codes and Finite Projective Planes
  32. [PDF] 3. The axioms of projective geometry
  33. The axiomatics of ordered geometry: I. Ordered incidence spaces
  34. [PDF] Projective Spaces
  35. [PDF] The real n-dimensional projective space, RPn is ... - UC Davis Math
  36. [PDF] Real Projective Space: An Abstract Manifold
  37. [PDF] Reading 15 Real projective space - Hiro Lee Tanaka
  38. Analytic Projective Geometry – Projective spaces and linear varieties
  39. [PDF] MATH431: Real Projective 3-Space - UMD MATH
  40. [PDF] Projective Geometry: From Foundations to Applications
  41. [PDF] Hilbert's Axioms - IME-USP
  42. [PDF] Lecture 1: Projective planes
  43. [PDF] Formalization, Primitive Concepts, and Purity
  44. [PDF] DESARGUES' THEOREM Two triangles ABC and A ... - OSU Math
  45. [PDF] The Theorem of Pappus and Commutativity of Multiplication
  46. [PDF] Affine and Projective Planes - BearWorks - Missouri State University
  47. Projective Planes - jstor
  48. [PDF] A Simple Non-Desarguesian Plane Geometry - IME-USP
  49. [PDF] Foundations of Projective Geometry
  50. [PDF] The fundamental theorem of projective geometry - Academic Web
  51. Harmonic_Perspectivity
  52. [PDF] Projective Geometry - TU Berlin
  53. [PDF] CLASSICAL GEOMETRIES 13. The cross ratio - Cornell Mathematics
  54. [PDF] MULTIDIMENSIONAL PROJECTIVE GEOMETRY
  55. Duality Principle -- from Wolfram MathWorld
  56. [PDF] Duality - DPMMS
  57. [PDF] 18.900 Spring 2023 Lecture 26: Projective Geometry
  58. [PDF] Oriented Projective Geometry - CMU School of Computer Science
  59. [PDF] 3 Polarities and forms
  60. [PDF] GEOMETRY REVISITED H. S. M. Coxeter S. L. Greitzer - Aproged
  61. projective line configurations - PlanetMath.org
  62. [PDF] Incidence Geometry - G Eric Moorhouse
  63. Projective Geometry | SpringerLink
  64. [PDF] Manifold Theory Peter Petersen - UCLA Mathematics
  65. Harmonic ovals | Journal of Geometry
  66. Real plane separating (M-2)-curves of degree d and totally ... - arXiv
  67. Finite Projective Geometries - jstor
  68. [PDF] An Introduction to Finite Geometry
  69. https://math.libretexts.org/Bookshelves/Combinatorics_and_Discrete_Mathematics/Combinatorics_(Morris
  70. [PDF] FINITE PROJECTIVE PLANES An incidence structure π = (P,L) is ...
  71. None
  72. [PDF] Affine planes, ternary rings, and examples of non-Desarguesian ...
  73. https://doi.org/10.4153/CJM-1961-035-6
  74. Nearfield Planes - Magma Computational Algebra System
  75. Survey of Non-Desarguesian Planes - American Mathematical Society
  76. [PDF] introduction to algebraic geometry, class 8
  77. [PDF] Math DRP Spring 25 Summary - Algebraic Geometry
  78. [PDF] INTRODUCTION TO ALGEBRAIC GEOMETRY
  79. [PDF] Projective Geometry - Purdue Computer Science
  80. [PDF] Introduction to Bezout's Theorem - UChicago Math
  81. [PDF] Bézout's Theorem for curves - UChicago Math
  82. Algebraic geometry : Hartshorne, Robin - Internet Archive
  83. [PDF] Math 632 Notes Algebraic Geometry 2
  84. [PDF] Camera Models and Fundamental Concepts Used in Geometric ...
  85. Epipolar Geometry and the Fundamental Matrix (Chapter 9)
  86. [PDF] Epipolar Geometry and the Fundamental Matrix
  87. None
  88. 404 Not Found
  89. [PDF] Möbius Transformations Revealed
User Avatar
No comments yet.