An improved color separation method for object detection in complex background

An improved background subtraction algorithm based on color separation method is introduced to solve the problem of high rate of missing detection of moving object in complex background. First, the current and background color images are individually divided into three images according to RGB channels. Then the background subtraction is made, and the object with threshold in every channel is segmented. After that, three result images are combined to obtain a completed foreground image. Edge subtraction result is used to fix the foreground image so that the noise brought by light is removed. At the same time, the background image from each channel is updated adaptively, in real time and individually. The results show that this method can effectively detect the difference of color which cannot be found by tranditional method because of similar grayscale in complex background.