红外系列图像运动小目标检测  

Small Moving Target Detection in Infrared Image Sequences

作　　者：陈玉丹[1] 

机构地区：[1]军械工程学院光学与电子工程系,河北石家庄050003

出　　处：《兵工自动化》2006年第12期65-66,共2页Ordnance Industry Automation

基　　金：国家高技术"863"计划项目(2003AA775093)

摘　　要：红外系列图像运动小目标检测,基于形态学滤波抑制图像系列背景方法,在图像分割基础上,利用目标在帧间运动的连续性进行目标识别。即首先通过形态学滤波进行图像预处理,用自适应阈值方法分割出候选目标,再利用目标运动特征,采用邻域判决法最终检测出目标。模拟实验表明该方法能够准确高效地检测出运动小目标。The detection of small moving targets was based on the image serial background of morphologic filtering restrain. Based on image segmentation, the movement continuity of targets in frame was adopted to carry out the target reorganization. That is, first morphologic filtering was used for preprocessing. Then segment possible targets were acquired through the self-adaptive threshold. At last neighboring region verdict was brought forward to detect target based on target movement character. Simulation result indicated that the algorithm could detect small moving targets with high efficiency.

关 键 词：红外系列图像 小目标检测 形态学滤波 图像预处理 图像分割 邻域判决 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]