Last update, Dec.10, 1998

BACK


Publications on Projective Geometry :

Publications on Perspective Geometry :

BibTeX references .


The Cross Ratio

by Jim Blinn (MSR - Graphics)
IEEE Computer Graphics & Applications, v.18(6), Nov/Dec. 1998, pp.78-80, Jim Blinn's Corner.

Notes

Explains the concept of cross ratio in projective geometry and proves it (easy explanation).


La Place de J.-H. Lambert (1728-1777)
dans l'Histoire de la Perspective

par Roger Laurent, Cedic/Nathan: Paris, 1987, 316 pages.

Notes

Restitution en géométral d'une perspective ou des règles inverses de la perspective

Problème: étant donné une perspective, où en général ne figure pas la position de l'observateur par rapport au tableau, retrouver les dimensions géométrales du dessin ainsi représenté, à une échelle donnée. (Peut-on aussi reconstruire un bâtiment ainsi?)

La résolution de ce problème pré-suppose que le tableau soit perpendiculaire au plan géométral et que l'on puisse déduire les 4 données suivantes (qui caractérisent une perspective):

  1. la ligne d'horizon;
  2. le point de l'oeil sur le tableau;
  3. la distance de l'oeil au tableau;
  4. la hauteur de l'oeil au-dessus du plan géométral.

Problem 21

"Find the Horizon Line, the Eye Station and the Principal Distance, from the 2PtP of a square."

Solution:

Let abcd be the perspective image of the square ABCD in Ground Plane G. The parallel sides of the square, ab & cd define the VP m, while the parallels ad & bc define the VP M. These 2 VPs thus define the Horizon Line (HL).

Furthermore, mbM must be a right angle (in space). The same is true of the diagonals acN and bdn. Hence, the VPs m & M and the diagonal intercepts with the HL, n and N, determine two circles upon which the Eye Station Q must lie. This determines Q as the intercept of both (half-)circles. N.B.: These 2 half-circles are called "arcs capables d'angles droits" in French.

Finally, the orthogonal line from the Eye Station, Q, to the Horizon Line determines the Pincipal Point , P. The (scaled) length QP is then the Principal Distance, f (or focal distance).


Square in 2PtP: finding the eye (Q) & PP (P).


Resolving the square in 2PtP.

Problem 22

"Find the Horizon Line, the Eye Station and the Principal Distance, from the 2PtP of a rectangle, knowing the ratio of its two sides."

Solution:

Assume the ratio AD/AB is known for the rectangle ABCD. Then, the internal angle alpha = DBA is also known (since DAB is a right angle): alpha = arctg(AD/AB). This is also the angle associated to dab, the perspective projection of DAB.

Consider now the VPs m & n associated to the directions BA & BD (or ba & bd in perspective), respectively. These 2 VPs determine the Horizon Line, as usual.

Furthermore, Q, the Eye Station, must be on the half-circle through m and M as usual, where M is the VP for the other side of the rectangle. i.e., BC (bc).

Then, alpha = mQn, since Qm is parallel to BA (ba) and Qn is parallel to BD (bd). Thus, we have the following geometrical construction to retrieve the locus of Q on the "cercle capable":

  1. On the circle of diameter mM, find the point H, s.t. the angle HMm = alpha (=arctg(mH/MH)).
  2. Then, construct the line through Hn, which intercepts the circle at Q.


Resolving the rectangle in 2PtP.


Brook Taylor's work on Linear Perspective

by Kirsti Andersen,

In "Sources in the History of Mathematics and Physical Sciences", vol.10, Springer-Verlag, 1992.

Notes

Introduces a particular form of "cross-ratios", where one of the four points is a vanishing point (p.28).

Uses an angle-scale on a vanishing line and the semi-circle including the eye to find how to draw 2D figures making an angle with respect to the viewer (pp.30-33).

Considers certain Inverse problems of perspective.

VPs IKH & Station Point (O) generate the
visual tetrahedron. S is the ortho. projection
of O & orthocenter of triangle IKH .
Resolving the visual tetrahedron,
given a cube in 3PtP.

Mention of John Hamilton's "Stereography or a Compleat body of Perspective" (London: 1738, 1749), where the curve of foreshortening is studied (an hyperbola):


Page created & maintained by Frederic Leymarie, 1998-2001
Comments, suggestions, etc., mail to: leymarie@lems.brown.edu