【作者】 王宏群； 方帅； 高明；

【Author】 WANG Hong Qun(Anhui vocational college of police officers)

【机构】 安徽警官职业学院； 合肥工业大学；

【摘要】 在运动目标检测过程中,若两帧图象在拍摄条件下光线强度发生突变,则运动目标检测的结果将会受到影响。据此,该文提出了抑制光强突变影响运动目标检测方法。首先,利用统计的方法建立混合高斯背景模型,并实时地对背景模型进行更新,一般光强变化不大,则背景减除检测结果是准确有效的,如变化大,则先对该帧采用分水岭算法分割出目标轮廓,然后与检测结果进行融合,互相补充,互相增强。该算法较好地解决了外界光强突变给运动目标检测带来的问题。实验结果证明,该算法对固定场景下运动目标检测是快速有效的。更多还原

【Abstract】 During the course of the moving target detection,if light intensity changes obviously in shooting the two photos,the results will be affected by the changes.Therefore,the paper proposes one way of moving target detection by restraining the changes in light intensity.Firstly,we build mixed Gaussian background model based on statistics,and update the background model from time to time.If the changes of light intensity are tiny,the results of background subtraction detection are effective.Otherwise,we segment the objects contour by link watershed algorithm,and then we combine it with the results of detection.They will be combined with each other and reinforce mutually.The algorithm can solve the problems resulted from the changes of light intensity from outsides.The experiment proves the algorithm is very useful in the moving target detection of fixed scenes.更多还原