摘要
针对动态场景下因背景补偿效果欠佳而不能准确检测运动目标的问题,提出一种基于精确背景补偿的运动目标检测算法。算法采用加入对称约束的SURF特征点匹配算法,以获得稳健的匹配点对。同时利用自适应外点滤除法去除目标点对全局运动估计的影响,显著地提高了背景补偿的精度。最后用帧差法准确地检测出运动目标。实验结果表明,该算法具有很好的鲁棒性,能够在背景复杂且摄像机运动的环境下准确地提取出运动目标。
As the imprecise background compensation result will lead to inaccurate detection of moving objects in dynamic scene, we put forward an algorithm of detecting moving objects which is based on accurate background compensation. This algorithm applies speeded up robust feature （SURF） points matching method with symmetry restriction added for obtaining robust matching points. Meanwhile it employs an adaptive outliers filtering method to remove the influence of objective points on global motion estimation. As a result, the accuracy of the background compensation is improved obviously. At last the moving objects are detected by frame difference method. Experimental results show that with favourable robustness, the algorithm is able to extract the moving objects accurately even in complex background while the camera is moving.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第10期139-141,198,共4页
Computer Applications and Software
基金
河北省科技支撑计划项目(11213518D)
关键词
动态场景
SURF
背景补偿
运动目标检测
Dynamic scene Speeded up robust feature （SURF） Background compensation Moving objects detection
