不同云层背景中目标的识别  

Target recognization under different complicated background of cloud

作　　者：易亨瑜[1] 吕百达[1] 季云松[2] 

机构地区：[1]四川大学激光物理与化学研究所,成都610064 [2]中国工程物理研究院应用电子学研究所,绵阳621900

出　　处：《激光技术》2004年第2期133-136,140,共5页Laser Technology

摘　　要：针对云层中目标的识别问题 ,选择Sobel算子对输入图像进行了边缘检测 ,并以计算机模拟光学相关运算给出了识别结果。结果显示 ,在积状云和层状云背景中的目标识别较容易 ,在波状云中的目标识别较难。背景强度较强 ,目标亮度过强或过弱 ,或采集器件等因素导致图像中输入图像模糊 ,也会影响目标的识别率。另外 ,强目标相关峰比较尖锐 。How to recognize targets under complicated background of cloud is studied in this paper by stimulating optical correlation. Some factors influencing recognition are calculated, such as the change of shape of cloud, target intensity, clearness or faintness of input image and so on. Simulation results show that targets can be found more easily in cumulus or stratus than in undulance cloud, because undulance cloud is more abrupt change in its intensity. The analyses also show that the clearer the input image is, the more easily the target is recognized;and the more intensive the target is, the more easily it is recognized.

关 键 词：目标识别 光学相关 边缘检测 SOBEL算子 

分 类 号：O438[机械工程—光学工程]