SAR图像舰船目标快速检测方法研究  

A Fast Algorithm for Ship Detection in SAR Images

作　　者：王智勇[1] 窦浩[1] 田金文[1] 

机构地区：[1]华中科技大学自动化学院多谱信息处理技术国防科技重点实验室,武汉430074

出　　处：《舰船电子工程》2016年第9期27-30,88,共5页Ship Electronic Engineering

基　　金：国家自然科学基金(编号:61273279;61273241)资助

摘　　要：随着SAR图像分辨率的不断提高,传统CFAR检测算法由于速率较慢,已经不能满足实际应用的需求。针对高分率SAR图像,提出了一种两级舰船目标快速检测算法。第一级采用改进的最大熵双阈值检测算法将图像分割为目标、疑似目标、背景三部分;第二级采用改进的双参数CFAR检测算法,在第一级结果基础上进行自适应二次检测,最后得到舰船目标。通过真实SAR图像数据实验表明,提出的算法具有良好的检测性能,更符合实际高分辨率SAR图像舰船目标检测的应用需求。With the improving resolution of SAR Images, the traditional Constant False Alarm（CFAR） algorithm is in-efficient for ship detecting. A new two-stage fast method is proposed for high resolution SAR images in this paper. In the first stage, the improved algorithm of entropic double-thresholds proposed by Kapur, Sahoo, Wong（KSW） is implemented. Then the whole image is divided into three parts, including target, suspected target, and background. In the next stage, the improved two parameters CFAR algorithm is applied to the pixels in the suspected target area. Combing results of two stages, ship targets in the image are detected. According to the experiment with real SAR images, the results show that the method proposed in this paper performs very well in ship detecting. The fast algorithm satisfies the demand of ship detection in high resolution SAR images.

关 键 词：合成孔径雷达SAR 舰船检测 CFAR 最大熵双阈值检测算法 快速算法 

分 类 号：TN957.52[电子电信—信号与信息处理]