Research of image recognition technology about moving object based on low-resolution camera

In this paper, we focus on the issue of accuracy in the process of motion pattern detection, and propose a new multi-frame superposition motion detection algorithm on the base of traditional adjacent frame difference method. First, we deal the captured image with blurring, color separation process and so on, then implement the multi-frame superposition processing technology and calculate the average coordinates of movement points to analyze the image. And the experiments prove that this algorithm maintains the traditional algorithm´s efficiency while has a great improvement on accuracy and versatility at the same time.