一种改进的适用复杂场景的运动目标检测算法

Improved Algorithm of Motion Detection in Complicated Environment

针对传统运动目标检测算法中存在的"拖影"、光变干扰、阴影等问题,提出了一种改进的更具鲁棒性的检测算法。基于背景边缘检测差并通过两次结合帧间差分法以及颜色偏差用以消除噪声和减低运动目标边缘断裂现象,从而获取运动目标的完整轮廓,同时采用双向模板填充算法进行运动目标的分割,最后通过数学形态学滤波和连通域分析来进一步去除噪声和填补空洞,获得完整理想的运动目标区域。实验结果证明,相对于传统的帧差与背景差分检测算法,能够有效地克服阴影和光扰所产生的噪声问题,可以在复杂背景下准确地检测分割出运动目标,并满足实时性要求。

As the traditional algorithms of detecting moving target cannot solve the problems such as ghost,illumination change and shadows,an improved robust algorithm for motion detection from complicated background in video sequences is proposed in this paper.We get the edge map of foreground object using the new algorithm based on background edge-characteristic difference combining inter-frame difference and chromaticity distortion to reduce noise and edge gaps.Furthermore,foreground object can be segregated completely by filling the edge map.Finally,mathematical morphological post-processing and connectivity analyzing methods are introduced to reduce noise and gaps.Proved by the results of experiments on standard and real video sequences,the proposed algorithm can get exact moving object from complex background and eliminate disturbing of background and the illumination changes efficiently in real-time.