一种特征点和曲线相似度的目标识别方法

本文以5型常用火箭为研究对象，针对当前雷达散射截面（RCS）特征识别存在的问题，基于全空域RCS仿真库和连续RCS测量信息，开展了特征点和曲线相似度的识别方法研究。利用仿真数据和测量数据，对双频段特征点识别能力进行了评估，在4.5 dB阈值下的识别能力可达100%。研究了曲线相似度比较方法，采用动态时间规划实现了RCS曲线的相似性度量，取得较好的应用效果。提出了新的基于曲线相似度的RCS目标识别方法，该方法对于提高雷达目标识别能力具有借鉴意义。

A Target Identification Method for Feature Point and Curve Similarity

In view of the current problems of RCS feature identification, this paper, taking five types of rockets as the research objects, carries out feature point and curve similarity identification study based on full space RCS simulation library and continuous RCS measurement information. Using the simulation data and the measurement data, the identification ability of two-band feature points are evaluated, reaching 100% at the 4.5 dB threshold. The comparison method of curve similarity degree is studied. The similarity measure of RCS curve is realized by DTW (dynamic time warping) and good application results are achieved. The new curve similarity-based RCS target recognition method is useful for improving radar recognition capability.