Matching color uncalibrated images using differential invariants

This paper presents a new method for matching points of interest in stereos copic, uncalibrated color images. It consists in characterizing color point s using differential invariants. We define additional first order invariant s, using color information, and we show that the first order is sufficient to make the characterization accurate. The characterization thus obtained i s invariant to orthogonal image transformations. In addition, we make it ro bust enough for affine illumination transformations. We go on to present a generalization of a gray-level corner detector to the case of color images. Thirdly, we propose a robust and fast incremental technique for matching p oints of interest in uncalibrated cases, which works robustly and rapidly w hatever the number of points to be matched. Our matching scheme is evaluate d using stereo color images consisting of many points, with viewpoint and i llumination variations. The results obtained clearly show the relevance of our approach. (C) 2000 Elsevier Science B.V. All rights reserved.