红外序列图像弱小动目标检测的研究  

Detection of weak point moving targets in infrared image sequences

作　　者：许建忠[1,2] 王祖林[1] 赵毅寰[1] 郭旭静[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100191 [2]河北大学物理科学与技术学院,河北保定071002

出　　处：《光学技术》2010年第3期460-463,共4页Optical Technique

摘　　要：针对云天背景下的红外运动弱小目标,提出了一种融合检测的方法。根据视觉的固视微动特性进行单帧图像的目标增强,利用帧差分方法检测多帧图像的动目标,最后将这两种处理结果进行基于调制的融合,并经过图像分割,从而达到动目标检测的目的。仿真结果表明该检测方法能够有效地实现复杂背景下的弱小动目标检测。A fusion detection method that detects moving targets in infrared image sequences under clouds and sky background is proposed.By making use of the eye movement mechanism in visual fixation,targets detection is carried out in the single-frame image.Frame-difference method is used for moving target detection of multi-frame images.Finally,carrying on fusion with modulation for these two kinds of test result,image division is realized and the goal of moving target detection is achieved.The simulation indicates that this examination method can carry on weak point moving target detection effectively under the complex background.

关 键 词：固视微动 帧差分 融合 目标检测 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]