基于分数阶积分算子的红外弱小目标检测  

Infrared dim target detection based on fractional integral operator

作　　者：陈炳文[1] 王文伟[1] 秦前清[2] 

机构地区：[1]武汉大学电子信息学院,武汉430079 [2]武汉大学测绘遥感信息工程国家重点实验室,武汉430079

出　　处：《控制与决策》2012年第1期147-151,共5页Control and Decision

基　　金：国家973计划项目(2006CB701303);湖北省自然科学基金重点项目(2009CDA141);中央高校基本科研业务费专项资金项目(212274467)

摘　　要：针对现有背景抑制算法不能有效抑制背景而导致目标检测率低的问题,提出一种基于分数阶积分背景抑制的弱小目标检测算法.首先通过研究分数阶积分理论,分析分数阶积分非线性、尖锐的低通频率特性,结合分数阶积分和警戒环技术来抑制背景;然后依据目标和背景边缘的空间特征,采用模板均差法来提取候选目标;最后结合目标的运动连续性进行多帧轨迹关联,从而检测出真实目标.实验结果表明,该算法能有效抑制背景,提高信噪比,进而有效检测出弱小目标.In order to solve the problem that the current algorithms cannot eliminate the background effectively and lead to the low detection rate, a infrared dim target detection approach based on fractional integral operator is proposed. Firstly, through studying the fractional integral theory and analyzing its predominant sharp low-pass frequency character, the fractional integral operator is introduced to combine with the guard ring technology to eliminate the infrared background. Then, by analyzing the space features of targets and background edges, the difference of template average algorithm is introduced to detect the candidate targets. Finally, by combining with the continuous motion character of targets, the multiframe trajectory related algorithm is used to eliminate the false targets. Experiment results show that the proposed approach can eliminate the background effectively so as to enhance the signal-to-noise value, and the targets can be detected effectively in the low signal noise ratio background.

关 键 词：弱小目标检测 分数阶积分算子 警戒环 模板均差法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]