基于灰度特性的海天背景小目标检测  

Detection of sea-sky line in complicated background based on grey characteristics

作　　者：董宇星[1,2] 刘伟宁[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京100039

出　　处：《中国光学与应用光学》2010年第3期252-256,共5页Chinese Optics and Applied Optics Abstracts

基　　金：国家863高技术研究发展计划资助项目(No.2006aa701410)

摘　　要：针对复杂海天背景下的小目标检测存在海浪、云层干扰等问题,提出了先提取海天线,然后利用一维最大熵阈值分割法对出现在天空、海面或者海天线附近特定区域的目标进行检测的算法。该算法主要利用天空海面行灰度均值特性,结合梯度运算和形态学运算在海天线的潜在位置中检测边缘,进而用强鲁棒性的Hough变换直线检测法拟合海天线,实现对海天线的准确定位。实验处理分辨率为256 pixel×256 pixel的位图时,定位海天线需时4.1 ms,检测到目标需时5.3 ms,完全满足高帧频图像处理的实时性要求。实验结果表明,该算法能够快速、准确地检测出小目标,大大降低了虚警率。To detect precisely small and dim targets in a complicated background with ocean waves and cloud clutters, a new algorithm is presented which picks up the sea-sky line at first, then efficiently detects the targets on the special region under sky, ocean, and nearby the sea-sky line. On the basis of the line gray charac- teristics of sea and sky, this algorithm combines the gradient method and morphologic operation to segment the sub-band image in a potential area of sea-sky line, further to fit the sea-sky line by Hough transform to implement the precise location of the sea-sky line. The experiment result shows that this method can extract the seasky line from the complicated sea-sky background and can detect the small targets fast and efficiently. For a 256 pixel × 256 pixel file, the sea-sky line can be located in 4. 1 ms and a small target can be detected in 5.3 ms, which demonstrates that the algorithm can meet the requirements of high frame image processing in real time.

关 键 词：小目标检测 海天线 灰度特性 背景分析 图像处理 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]