红外起伏背景下运动点目标的检测方法  

SMALL MOVING TARGET DETECTION IN CLUTTER INFRARED BACKGROUND

作　　者：李吉成[1] 沈振康[1] 

机构地区：[1]国防科技大学电子工程学院ATR实验室,长沙410073

出　　处：《红外与激光工程》1997年第6期8-13,共6页Infrared and Laser Engineering

摘　　要：文中根据目标、背景干扰和噪声在红外序列图像中的差异，提出了一种基于空间高通滤波和时间域上最大递归滤波的运动点目标检测方法。该方法可以在低信噪比的情况下消除红外起伏和随机噪声的影响，有效地检测出点目标。实验结果表明，采用这种时空混合处理的方法可以得到满意的结果。A new algorithm for detecting small moving targets in noisy image se-quences is presented in this paper. The approach takes the following three steps: (1 )clutter suppressing, (2) adaptive thresholding, and (3) recursive maximum correlationfilter. AInng the process, clutter suppressing based on the gradient - inverse- weightedneighborhood averaging (GIWNA) filter and the recursive maximum filter are maincontribution of this paper. Both of them can be implemented by fully parallel algorithm,and real-time processing is possible for image sequences at high frame rate. Test resultsshow that the approach is capab1e of detecting small low-contrast moving targets withunknown position and trajectory.

关 键 词：红外序列图像 点目标 目标检测 图像处理 

分 类 号：TN211[电子电信—物理电子学] TN919.8