基于三帧差分及Canny算子的运动目标提取

针对目前常用的运动目标提取方法易受到噪声、光线变化的影响,很难提取出完整的运动目标这一问题,提出一种基于三帧差分和Canny边缘检测相结合的运动目标提取算法.该算法首先对连续三帧图像进行差分；然后对差分结果其进行区域填充,得到运动区域；再对当前帧进行Canny边缘检测得到边缘图像,二者相“与”得到运动目标的精确边缘图像；最后通过区域填充得到运动目标图像,从而实现运动目标的提取.实验表明,该算法可以实时有效地将运动物体从图像序列中提取出来.

The moving target extraction based on the three frame difference and Canny operator

According to the present commonly used method to extract moving object is susceptible to noise and the change of the light, it is difficult to extract moving target complete the problem, put forward a kind of based on the three frame difference and canny edge detection of moving object extraction algorithm. The algorithm firstly to difference of three consecutive frames, then fill in the area, the result of the difference of motion area, then the current frame for Canny edge detection to get the edge image, both get moving object with accurate edge image, and finally moving object images is obtained by area filling, so as to realize the extraction of moving objects. Experiments show that the proposed algorithm can effectively real-time moving object from image sequence extracted.