基于模糊理论的SAR图像海上舰船检测方法研究  

A method for ship detection in SAR images based on fuzzy theory

作　　者：李长军[1] 胡应添[1] 陈学佺[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027

出　　处：《计算机应用》2005年第8期1954-1956,共3页journal of Computer Applications

基　　金：中国科学院知识创新工程资助项目(KZCX0101)

摘　　要：舰船检测是合成孔径雷达图像海洋应用的一个重要部分,针对中分辨率近岸海域SAR图像,提出了一种基于模糊理论的海上舰船检测方法。该方法先利用改进的模糊增强算法对图像进行增强处理,以改变图像灰度的分布特性,从而分离图像中海洋区域和陆地区域,并结合最大熵分割法提取海洋背景中包含候选舰船的感兴趣区域,最后,对ROI区域进行分割,提取舰船的特征,并基于模糊推理技术实现对海上舰船目标的检测。Automatic interpretation of synthetic aperture radar (SAR) images was one of the most important application fields in image processing. Focusing on the medium resolution SAR images and combining with the previous algorithms, a novel technique to detecting ship targets from coastal regions based on fuzzy theory was proposed. In this new method, the input image was first processed with improved fuzzy enhancement algorithm so as to alter the characteristics of the gray-level distribution; then, with the threshold method, the sea and land regions could be separated; following that, the maximum entropic algorithm was employed to further process the image and regions of interest which contained candidate ship targets could be extracted; Then ROIs were segmented and the features of ship targets were extracted. Finally the ship targets could be detected based on fuzzy reasoning technique.

关 键 词：SAR 模糊推理 模糊增强 舰船检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]