基于快速搜索的运动目标检测算法  

An Algorithm of Detecting Moving Objects Based on Fast Searching Method

作　　者：苏晓星[1] 申金缓[1] 戴志强[1] 常胜江[1] 张延炘[1] 

机构地区：[1]南开大学现代光学研究所教育部光电信息技术重点实验室,天津300071

出　　处：《光电子．激光》2003年第3期281-284,共4页Journal of Optoelectronics·Laser

基　　金：天津市光电子联合研究中心资助项目(0031021)

摘　　要：对高分辨率下的运动目标检测算法提出一种基于快速搜索的改进方案,通过对本方案的理论分析与实验研究,表明当运动目标在整个扫描图像中所占比例较小时,与一般算法相比,此方案可以有效地提高运动目标的检测速度,在本文的实验中提高100多倍。结果表明该方案不仅能明显提高了检测的实时性,而且能给出更加精确的目标边界。An improved scheme of algorithm of detecting moving objects in scene images was proposed.The fast searching method was used to improve the detecting speed of the moving objects.The experimental results have shown that the detecting speed is improved more than 100 times.Theoretical analysis and experimental results have demonstrated that this scheme can increase the processing speed significantly and provide better silhouette of the moving object segmented under certain conditions.

关 键 词：算法 快速搜索法 高分辨率 运动目标检测 

分 类 号：TN247[电子电信—物理电子学]