Video moving target binary image processing method based on OTSU

Because of some pixels of foreground image which is binarized by fixed threshold, the Gaussian mixture model can not get accurate result. So we introduce OTSU to look for the threshold which satisfy maximum-between-clusters-variance,and it will get best divisional effect.It achieve adaptive threshold image dynamic adjustment of image binarization. Then use video data to validate, the result shows that the algorithm have better result by comparing with the traditional binarization algorithm in the video image.