TARGET RECOGNITION ALGORITHM BASED ON SALIENT CONTOUR FEATURE SEGMENTS

A target recognition algorithm based on salient contour feature segments is presented. The algorithm segments the contour according to the contour division scheme and evaluates the value of the contour segment from three aspects: curvature difference, fluctuation amplitude, and bending ratio. The segments of unreasonable division are merged. The importance of the segment is evaluated from the aspect of the length of the contour segment relative to the overall contour length. On this basis, the reasonable segmentation result of the contour segments is obtained. The similarity measures are performed between the reasonably divided contour segment and the contour segment database to obtain the best matching result of the target. Experimental results show that the proposed algorithm determines the parameters of significance evaluation of contour segmentation, which ensures that the segmented contour segment has significant features and improves the target recognition rate.