基于改进差分算法的运动目标检测与跟踪  

作　　者：王镇西[1] 王冬青[1] 张惠娟[1] 

机构地区：[1]同济大学,上海200240

出　　处：《工业控制计算机》2010年第10期71-72,共2页Industrial Control Computer

摘　　要：针对视频序列中运动目标检测与跟踪的问题,改进传统的差分方法,提出一种基于二次插值的差分算法。在相邻两帧图像中插入若干图像使所得图像的背景近似为不动的背景,在用差分算法对图像检测与跟踪。实验结果表明,该方法能有效提高跟踪的效率和精度。Moving object detection in video sequences and tracking problems,improve the traditional difference method,proposed a differential method based on quadratic interpolation.Two adjacent frames in a number of images to insert a background image from approximately fixed background image in the use of differential detection and tracking algorithm.Experimental results show that the method can effectively improve the tracking efficiency and precision.

关 键 词：边缘提取 运动目标检测 人体运动跟踪 差分图象 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]