A remote sensing ship recognition method based on dynamic probability generative model

Abstract

Aiming at detecting sea targets reliably and timely, a novel ship recognition method using optical remote sensing data based on dynamic probability generative model is presented. First, with the visual saliency detection method, prior shape information of target objects in put images which is used to describe the initial curve adaptively is extracted, and an improved Chan–Vese (CV) model based on entropy and local neighborhood information is utilized for image segmentation. Second, based on rough set theory, the common discernibility degree is used to compute the significance weight of each candidate feature and select valid recognition features automatically. Finally, for each node, its neighbor nodes are sorted by their ε-neighborhood distances to the node. Using the classes of the selected nodes from top of sorted neighbor nodes list, a dynamic probability generative model is built to recognize ships in data from optical remote sensing system. Experimental results on real data show that the proposed approach can get better classification rates at a higher speed than the k-nearest neighbor (KNN), support vector machines (SVM) and traditional hierarchical discriminant regression (HDR) method.