Abstract:

As the extraction of lunar craters presently suffer from insufficient quantitative information and mis-extraction, this paper proposes an algorithm of automatic extraction and recognition of lunar impact craters fusing CCD images and DEM data of Chang’e-1. (1) The images of craters comply with specific rules derived from the illumination of sun, therefore the extraction of craters in CCD images is implemented by the matching of corresponding crater feature segments; (2) the mis-extraction of craters in CCD images is then eliminated in terms of  the continuity of the aspects of crater wall, which are computed using DEM data. (3) the edge points of a crater are extracted in the light of the abnormal change between the direction of the  normal vector of an edge point and that of a non-edge point, the parameters of craters are then computed through the fitting of edge points and the types of craters are recognized according to the proportion of points on the bottom and the hatch of craters. The proposed approach is tested on CCD images and DEM data of Chang’e-1. The results show that our proposed algorithm has good performance in the high-latitude areas where the distribution of lunar impact craters is uniform.

Key words: CCD image, DEM, Impact craters, Feature matching, Aspect