摘要
提出基于属性散射中心重构的合成孔径雷达(SAR)图像目标识别方法。该方法采用邻域匹配算法构建测试图像散射中心集与对应模板散射中心集的对应关系。并分别利用所有的测试散射中心以及匹配的模板散射中心基于属性散射中心模型重构测试图像和模板图像。在此基础上,设计重构图像之间的相似度度量并根据最大相似度的准则判定目标类别。利用MSTAR数据集在多种条件下进行了目标识别实验,验证了所研究方法的有效性。
This paper proposes a Synthetic Aperture Radar(SAR)image target recognition method by target reconstruction based on attribute scattering centers.A neighbor matching algorithm is firstly used to build the correspondence between the scattering center sets from the test image and its corresponding template scattering center sets.Then,all the test scattering centers and matched template scattering centers based on attribute scattering center model are used to reconstruct the test images and template images,respectively.A similarity measurement of thereconstructed images is designed and the object type is determined by the criterion of maximum similarity.Target recognition experiments are conducted on the public MSTAR dataset under various operating conditions and the effectiveness of the proposed method is validated.
作者
来雨
LAI Yu(Xi’an University of Architecture and Technology,Xi’an 710055,China)
出处
《火力与指挥控制》
CSCD
北大核心
2021年第2期46-52,共7页
Fire Control & Command Control
基金
陕西省年度基金资助项目(2017S014)。
关键词
合成孔径雷达
目标识别
属性散射中心
邻域匹配
目标重构
Synthetic Aperture Radar(SAR)
target recognition
attributed scattering center
neighbor matching
target reconstruction
