首页 | 本学科首页   官方微博 | 高级检索  
     检索      

卫星红外图像中单帧目标检测算法研究
引用本文:吴超,安玮,龙云利,林两魁.卫星红外图像中单帧目标检测算法研究[J].电子对抗,2008(5).
作者姓名:吴超  安玮  龙云利  林两魁
作者单位:国防科技大学电子科学与工程学院,长沙410073
摘    要:在有复杂云层背景的卫星红外图像中准确地检测出存在的运动目标是目前图像处理的研究热点。DBT类的运动目标检测算法有单帧目标检测和多帧轨迹关联两大步骤。为更好抑制起伏背景和背景结构边缘,文章提出一种新的单帧内背景抑制算法——邻域最小差值滤波,该算法与典型的背景预测算法(中值滤波和高通线性模版滤波)相比具有更好效果;最后针对高斯噪声条件下,利用统计分割方法对抑制背景结果进行阈值分割得到候选目标,从而形成一套针对卫星红外图像的快速有效的单帧目标检测算法。

关 键 词:卫星红外图像  点目标  邻域最小差值滤波  统计分割

Research on Target Detection Method in Single Satellite IR Image Frame
Wu Chao An Wei Long Yunli Lin Liangkui.Research on Target Detection Method in Single Satellite IR Image Frame[J].Electronic Warfare,2008(5).
Authors:Wu Chao An Wei Long Yunli Lin Liangkui
Institution:Wu Chao An Wei Long Yunli Lin Liangkui(School of Electronic Sci., Eng.,National Univ.of Defence Tech.,Changsha 410073,China)
Abstract:The exact detection of moving target from Satellite IR Image with complex cloud background is a hotspot in image processing research.DBT detection method contains two steps,which are target detection in single images frame and track association in sequence of images.For restraining the undulating background and the edge of background,neighbor minimal subtrahend filter is given as a new method which receives better outcome in contrast with typical background forecasting algorithm.Then the statistical segment...
Keywords:Satellite IR Image  point target  neighbor minimal subtrahend filter  statistical segmentation  
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号