Detection and recognition method for irregular craters on planetary surface

The paper focuses on the feature detection and recognition on planetary surface in the deep space exploration mission. A novel detection and recognition method for irregular craters, especially overlapped and incomplete craters on planetary surface is proposed. First, extraction of candidate crater edge is performed by using the Contextual-based Hopfield neural network. Furthermore, pseudo-edges which do not satisfy the constraints are removed by analyzing multi constraints for real crater edges. Finally, real edges containing the feature of the crater are detected by using the least median square ellipse fitting method combined with a robust least square method. Meanwhile the physical parameters are determined. Mathematical simulations demonstrate the performance of the proposed method for detection and recognition of irregular craters in complex background.