基于四阶矩的SAR图像目标检测  

Target detection of SAR image based on four-order moment

作　　者：王志[1] 桂任舟[1] 

机构地区：[1]同济大学电信学院,上海200092

出　　处：《信息技术》2012年第2期1-5,共5页Information Technology

基　　金：东南大学毫米波国家重点实验室开放课题项目(K200916);武汉大学测绘遥感信息工程国家重点实验室开放课题项目(09R05);卫星海洋环境动力学国家重点实验室开放课题项目(SOED1105)

摘　　要：SAR图像中车辆目标检测是一个备受瞩目的关注点,运用传统的低阶矩方法进行SAR图像中目标检测,尤其是在图像中含有乘性噪声和在分辨率降低时,普遍存在检测率不高,不能充分体现出目标相对背景杂波的特性。本文根据SAR图像中目标的特性,提出基于四阶矩的方法对SAR图像进行目标检测。实验表明,采用四阶矩方法后,在乘性噪声和分辨率降低状况下,四阶矩方法的检测概率相对传统的二阶矩方法有明显提高。未来四阶矩将更广泛地用于SAR图像中对目标进行检测,有重大的应用前景。Vehicle detection in SAR image is attractive field. Detection probability is generally low by using traditional low-order moments which is unable to distinguish target and background clutter effectively, especially when the figure is in the lower resolution and contains multiplicative noise. The article presehts a method based on fourth-order moment according to the feature of SAR image. The experiment shows that the detection probability with fourth-order moment method is much higher than loworder moment in above two conditions. In future, fourth-order moment will be widely applied in target detection of SAR image which definitely has significant promise.

关 键 词：目标检测 四阶矩 拖尾 检测概率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]