基于图像块相关的弱小目标检测算法研究  

Dim-small target detection based on correlation of image blocks

作　　者：郑利苹[1] 毛征[1] 陆天舒[1] 彭超[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《国外电子测量技术》2012年第5期32-35,共4页Foreign Electronic Measurement Technology

摘　　要：针对复杂云层背景下的弱小目标检测问题,提出了一种基于图像块相关的弱小目标检测算法。首先对图像进行灰度值反转,然后利用相邻块之间相关性进行云层背景抑制的预处理,从而削弱复杂云层背景对弱小目标检测的干扰,最后对图像进行灰度值拉伸处理,增强弱小目标,提高对目标的检测准确率。实验结果表明,提出的算法可抑制云层背景对弱小目标干扰,实现复杂背景下弱小目标的检测。In this paper,it is studied how to detect dim-small target under complex cloud background,and a novel method based on the correlation of image blocks is proposed.Firstly,the gray value of image is reversed,then the correlation of adjacent blocks is used for suppressing cloud background,sequentially the disturbance of cloud background is reduced.At last,the gray value of image is stretched,in order to enhance dim-small target,and improve the correct probability of the target detection.The experiments show that the novel method can suppress the disturbance of complex cloud background to dim-small target detecting,and realize the detection for dim-small target stably.

关 键 词：弱小目标 图像块相关 背景抑制 光电目标检测 

分 类 号：TH873.7[机械工程—仪器科学与技术]