胡利平, 董纯柱, 邢笑宇, 殷红成. SAR图像目标和阴影径向积分特征评估[J]. 电波科学学报, 2014, 29(2): 254-259+287. doi: 10.13443/j.cjors.2013042601
      引用本文: 胡利平, 董纯柱, 邢笑宇, 殷红成. SAR图像目标和阴影径向积分特征评估[J]. 电波科学学报, 2014, 29(2): 254-259+287. doi: 10.13443/j.cjors.2013042601
      HU Liping, DONG Chunzhu, XING Xiaoyu, YIN Hongcheng. An evaluation method of SAR images based on radial integral features of target and shadow[J]. CHINESE JOURNAL OF RADIO SCIENCE, 2014, 29(2): 254-259+287. doi: 10.13443/j.cjors.2013042601
      Citation: HU Liping, DONG Chunzhu, XING Xiaoyu, YIN Hongcheng. An evaluation method of SAR images based on radial integral features of target and shadow[J]. CHINESE JOURNAL OF RADIO SCIENCE, 2014, 29(2): 254-259+287. doi: 10.13443/j.cjors.2013042601

      SAR图像目标和阴影径向积分特征评估

      An evaluation method of SAR images based on radial integral features of target and shadow

      • 摘要: 为了评价理论建模建立合成孔径雷达(Synthetic Aperture Radar,SAR)图像模板的准确性,利用SAR图像中两个典型区域(目标和阴影),建立仿真与实测图像之间相似性的定量评估方法.该方法预先分割出目标和阴影区域,并分别提取目标轮廓、目标强度分布和阴影轮廓的极化映射径向积分特征,基于这三种特征计算仿真图像和实测图像的相关系数.对三种车辆目标(BMP2、BTR70、T72)的仿真SAR图像与运动和静止目标获取与识别(Moving and Stationary Target Acquisition and Recognition, MSTAR)实测SAR图像进行相似性比对和分类识别,结果表明,目标轮廓、目标强度分布、阴影轮廓的极化映射径向积分特征评估方法具有较好的相似性和分类性能.

         

        Abstract: A quantitative evaluation method of synthetic aperture radar(SAR)simulation image based on features of a target and its shadow is introduced to evaluate the validity of SAR templates by theoretical modeling.Firstly,target and shadow regions are segmented from the original images.Secondly,the radial integral features of polar mapping of a target contour image,a target intensity image and a shadow contour image are extracted,respectively.Finally,three correlation coefficients are computed based on these features.By computing the similarities and the correct recognition rates between the simulated SAR images and the moving and stationary target recognition(MSTAR)SAR images of three typical ground vehicles(BMP2,BTR70,and T72),we conclude that the evaluation method based on radial integral features of polar mapping is effective with higher degrees of correlation and classification rates.

         

      /

      返回文章
      返回