云层背景中目标的相关识别  

Detection of targets under complicated background of cloud

作　　者：易亨瑜[1] 叶一东[1] 张卫[1] 季云松[1] 

机构地区：[1]中国工程物理研究院应用电子学研究所

出　　处：《强激光与粒子束》2002年第5期693-696,共4页High Power Laser and Particle Beams

基　　金：中国工程物理研究院应用电子学研究所所长基金资助课题

摘　　要：针对云层中目标的识别问题 ,以计算机模拟光学相关运算给出了识别结果。当目标亮度一定时 ,在输入场景和匹配滤波器有无边缘检测的几种情况下 ,研究了背景强弱变化对目标识别性能的影响。结果显示 ,当背景强度较弱时 ,输入场景与匹配滤波器进行光学相关运算就可以找出目标位置 ;当背景强度较强时 ,输入场景的边缘检测是提高云层中目标识别率的关键 。This paper presents how to recognize targets under complicated background of cloud in sky by stimulating optical correlation. Some factors influencing recognition are calculated, such as changes of background intensity, edge detection of input image or matched filter and so on. Simulation results show that targets can be found in weaker background by direct optical correlation of input image and its matched filter without edge detection. When background becomes stronger, edge detection of input image is the key to improve target recognition capacity, but it weakens the intensity of correlative peak.

关 键 词：云层 光学相关 目标识别 边缘检测 计算机模拟 图像识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] O438[自动化与计算机技术—计算机科学与技术]