基于指数小波分形特征的SAR图像特定目标检测  

Special target detection of the SAR image via exponential wavelet fractal

作　　者：刘冬[1,2] 张弓[1] 

机构地区：[1]南京航空航天大学信息科学与技术学院,江苏南京210016 [2]中航工业第一飞机设计研究院,陕西西安710089

出　　处：《西安电子科技大学学报》2010年第2期366-373,384,共9页Journal of Xidian University

基　　金：教育部留学回国人员科研启动基金资助项目;航空科学基金资助项目(04D52032)

摘　　要：提出一种SAR图像特定尺寸、方向目标检测方法.根据指数小波对SAR目标边缘的选择性强化特性,对SAR图像进行方向指数小波滤波.基于滤波后的目标特性,定义指数小波分形特征(EWF).该特征依据SAR成像过程中目标尺寸、方向等先验性信息,实现对SAR图像特定目标检测.指数小波分形特征与传统的目标检测特征不同,它对目标尺寸、方向、背景对比度同时具有敏感性.基于MSTAR数据库的检测实验结果说明了EWF特征对尺寸、方向性SAR目标检测的有效性.This paper proposes a new approach for the SAR image detection of the target with a certain size and direction.We filter the SAR image with exponential wavelet directionally,based on its selective enhancement of SAR target edges.Then a novel term Exponential Wavelet Fractal（EWF） feature is defined considering the filtered target＇s characteristics.EWF can detect the SAR target using the priori information on the target size and direction.Different from most traditional SAR detections,EWF not only distinguishes target pixels from the background only based on the contrast,but also is sensitive to the object direction and size.The validity of the EWF method for the detection of targets with a known size and direction is evaluated based on the MSTAR data.

关 键 词：SAR图像处理 目标检测 合成孔径雷达 指数小波 特定目标检测 

分 类 号：TN959.1[电子电信—信号与信息处理]