基于方位弯折的高分辨距离像雷达目标识别

High resolution range profile radar target recognition based on aspect warping

针对高分辨雷达距离像的方位敏感性问题,将应用于语音处理的动态时间弯折技术引入高分辨距离像雷达目标识别领域,提出了一种动态方位弯折技术,该技术通过将测试数据与模板特征序列进行非线性伸缩(弯折),获得测试帧与模板之间最小匹配距离对应的最佳匹配路径,实现对测试数据的目标识别。对ISAR实测飞机数据的分类实验及与模板匹配法和RBF识别方法的比较,获得了良好的识别结果,表明了该方法的可行性和有效性。

Radar high-resolution range profile （HRRP） is very sensitive to target-aspect variation. Dynamic aspect warping is presented to tackle this aspect sensitivity for high resolution range profile radar target recognition. The method can be used to find the best matching route through nonlinear warping between test data and a template sequence where the matching distance between the test data and the template sequence becomes the smallest for achieving the target recognition. Experimental results of the proposed method for real-life ISAR data target recognition together with a comparison to the RBF and the MCC-TMM method, indicate effectiveness and efficiency of the proposed method.