Ellipsoidal Features Extraction for Planetary Image Registration

Using greyscale texture features recently become a new trend in supervised machine learning crater detection. Need to be analysed image data preferably by automatic technique as the data is in huge amount. Automatic feature extraction method is proposed and utilised for earth remote sensing images. These are not always applicable to planetary data which is having low contrast and uneven illumination features. Proposed a new method which is unsupervised for different ellipsoidal feature extraction for planetary image registration. Approach is based on combination of various techniques including Hough Transform and watershed segmentation technique. This is mainly applicable for geometrically compact shapes rocks, craters, and other geological features