基于背景差分的运动目标检测方法

针对静止摄像机下的运动目标检测问题,提出了一种基于背景减法的运动目标检测算法.通过对一组连续视频进行处理,从中得到不含运动目标的背景图像.再利用背景差分的方法提取出运动目标.在确定比较阈值的过程中,一改以往通过实验不断调整的做法,提出了动态阈值的概念,从而增强了检测效果,提高了算法的可实施性.融入了高斯模型关于背景更新的算法,克服了由于背景突然改变而造成的误检测.实验结果表明,通过背景差分与高斯模型相结合的方法,在有诸多不确定性因素的序列视频中构建背景有较好的自适应性,能迅速响应实际场景的变化,为准确地检测出运动目标提供了必要的基础.

Moving objects detection based on background subtraction method

A detection algorithm of moving objects based on background subtraction method was proposed for a video surveillance system with static camera. Background without moving objects was got from video sequences, and then moving objects were detected based on background subtraction method. To define the threshold, instead of using the experimental method which needs continual adjustment, a dynamic threshold concept was proposed to enhance detection effect and thus increase the possibility of implementation. The detection that integrates Gauss model concerning background renewal of calculation, overcomes the error resulted from the sudden change of background. Test result indicates that by combing background subtraction with Gauss model, the proposed detection method has better adaptability when constructing background in the serial video frequencies which have many indefinite factors. This method can respond to actual changes of the scene rapidly, providing necessary basis for accurately detecting the moving objects.