The Desargues theorem to build matching graph for N images

In this paper, we propose a new approach for matching image points by exploiting minimal geometric knowledge and in a correspondenceless way. The idea underlying our approach is to use a generalization of the Desargues theorem in images sequence context. This approach allows the matching image points over a sequence without using correlation techniques. The corresponding points are calculated directly using the Desargues invariants. Examples on real and synthetic images are presented. Results are illustrated and discussed.