Skip to Main Content
Crater detection from images is a challenging problem due to variations in the geometry shape, illumination, and scale. An algorithm with part based features to automatically detecting craters on landing surfaces is presented in this paper. It is build up a coarse-to-precise approach by learning pyramid histogram of oriented gradient features (PHOG) with part based crescent like structure, whose simplicity combined with an original learning strategy leads to a fast and high accuracy detect results. The approach is verified with images data sets from Mars captured by NASA.