Target Recognition Algorithm Based on Double Contour Feature Matching

In this paper, a target recognition algorithm based on double contour feature matching is proposed. When the algorithm faces the shape of an object with nonlinear deformation (occlusion and noise), it is difficult to match and the error matching rate is high. Firstly, the effective segment of the shape is obtained, and the feature points of the segment are obtained by uniform sampling. Then, the similarity of the centroid height function between the contour points of the contour segment is measured by dynamic regularization algorithm, and the confidence characteristics of the contour segment are calculated. Finally, the segmentation confidence is combined to match and identify again. The experimental results show that this method has high shape classification accuracy when dealing with nonlinear deformation.