系统工程与电子技术 ›› 2019, Vol. 41 ›› Issue (3): 534-540.doi: 10.3969/j.issn.1001-506X.2019.03.11

• 传感器与信号处理 • 上一篇    下一篇

基于模糊综合评判的车辆目标SAR仿真图像评估方法

胡利平1,2, 刘锦帆1,2, 王洪叶3, 闫华1,2, 殷红成1,2   

  1. 1. 北京环境特性研究所, 北京 100854; 2. 北京环境特性研究所电磁散射重点实验室, 北京 100854; 3.中国人民解放军63629部队, 北京 100162
  • 出版日期:2019-02-25 发布日期:2019-02-27

Vehicle SAR simulation images validation method based on fuzzy comprehensive evaluation

HU Liping1,2, LIU Jinfan1,2, WANG Hongye3, YAN Hua1,2, YIN Hongcheng1,2   

  1. 1. Beijing Institute of Environmental Features, Beijing 100854, China; 2. Science and Technology on Electromagnetic Scattering Laboratory, Beijing Institute of Environmental Features, Beijing 100854, China; 3. Unit 63629 of the PLA, Beijing 100162, China
  • Online:2019-02-25 Published:2019-02-27

摘要:

为了验证理论建模计算合成孔径雷达(synthetic aperture radar,SAR)仿真图像的准确性,给出基于层次分析法(analytic hierarchy process, AHP)、模糊综合评判(fuzzy comprehensive evaluation,FCE)、图像特征与Fisher可分性及目标识别等知识综合的车辆目标SAR仿真图像评估方法。首先确定五级评语集{优秀,良好,中等,较差,很差},然后根据AHP构建评估因素集,分别利用Fisher可分性判据和目标识别率确定二级FCE的两层权重向量,再依据图像相似度评价标准区间构建模糊隶属函数,最后将权重向量和模糊关系矩阵合成计算出仿真和实测图像的FCE结果,即模糊向量,对模糊向量按照最大隶属原则给出评价结果。通过对车辆目标的仿真与实测SAR图像进行比对分析结果显示,该方法可以准确客观地反映出SAR仿真与实测图像之间的相似程度。

Abstract:

In order to evaluate the veracity of synthetic aperture radar (SAR) images by theoretical modeling, a vehicle SAR simulation images validation method is proposed by synthesizing analytic hierarchy process (AHP), fuzzy comprehensive evaluation (FCE), and the technology of image features, Fisher criterion, and target recognition. Firstly, fivedegree comments {excellent, good, fair, poor, and very poor} are determined. Secondly, the evaluation factors are established based on the AHP, and then the first-level and second-level weight vectors are obtained by the recognition rates and the Fisher criterion, respectively. Thirdly, the fuzzy membership functions are constructed according to the standard of image similarity. Finally, the result vector of FCE between the simulated and measured images is computed by integrating the weight vectors and the fuzzy matrices, and then the final evaluation result is given by the maximum rule from the vector. By comparing and analyzing the experimental results of vehicles of SAR simulated and measured images, the proposed evaluation method is effective on reflecting the similarities of the simulated and measured images.