【作者】 陈炳文； 王文伟； 秦前清； 刘志刚；

【Author】 CHEN Bing-wen1,WANG Wen-wei1,QIN Qian-qing2,LIU Zhi-gang3(1 College of Electronic Information,Wuhan University,Wuhan 430079,China;2 State Key Laboratory for Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,Wuhan 430079,China;3 School of Geography,BEIJING NORMAL University,Beijing 100039,China)

【机构】 武汉大学电子信息学院； 武汉大学测绘遥感信息工程国家重点实验室； 北京师范大学地理学与遥感科学学院；

【摘要】 针对现有算法未能有效结合时域与空域信息,提出一种基于时空融合和粒子滤波的弱小目标检测算法。首先利用像素点的帧间相关性与邻域相关性,采用短程非参数核建模算法来检测候选目标;接着采用粒子滤波结合目标的运动特性来跟踪候选目标,剔除虚假目标得到真实目标航迹。实验结果表明,该算法能有效结合时域与空域信息,能有效检测出低信噪比背景下的目标。更多还原

【Abstract】 Aimed at fusing the time domain information and the space domain information effectively,an infrared dim target detection based on space-time integration and particle filter is presented.At first,the short non-parametric kernel model was used to detect the candidate target.Then combined with the character of target motion,the particle filter was used to track the candidate target,so as to eliminate the false target and get the track path of target.Experiments have been carried out and the results show that the proposed approach can fuse the time domain information and the space domain information effectively.It also can detect the target effectively in the low signal noise ratio background.更多还原