一种基于模型的合成孔径声呐图像目标快速识别方法

针对基于合成孔径声呐(SAS)图像目标识别的先验模板获取困难、运算复杂度高的问题，该文提出一种基于模型的改进型相关快速识别方法。首先，基于构造凸壳估计目标姿态角，实现目标成像几何关系的估计；其次，提出改进的基于隐藏点移除的目标图像快速生成方法，可实时得到各备选目标对应成像几何关系的仿真图像；进而基于图像相关实现目标图像识别；最后，仿真实验证明了算法的有效性。仿真实验结果表明，相比于常规的直接模板识别方法，该方法识别率高，计算速度快。

Fast Model-based Automatic Target Recognition Method for Synthetic Aperture Sonar Image

A modified model-based method is proposed to obtain sufficient prior templates and reduce the computational complexity on Synthetic Aperture Sonar (SAS) automatic target recognition. First, a quick method based on build convex hull is proposed to estimate the target pose quickly as well as the SAS imaging geometry for the specified target. Second, an improved method based on Hidden Point Removal (HPR) algorithm is proposed to obtain the target SAS simulation image effectively. Accordingly, the target recognition is realized by the correlation between the test image and the simulated image. Finally, the effectiveness of the proposed method is verified by the simulation experiment. It is shown that the proposed method can achieve higher computational efficiency than the conventional direct templet-based method, but remain the same high recognition rate.