The noise elimination of the image based on otsu

In the image processing, this paper proposed a new denoising algorithm for scanning images. The basic idea was that each grayscale probability was computed. In order to prevent the probability is too small, using 1 minus the grayscale probability, and with each grayscale corresponding between-cluster variance to multiply. Finally, let the gray value that made the result was the biggest grayscale as optimal threshold. Using optimal threshold to discriminate the foreground and background shadow, removed scanning images from the shadows. Experiments showed that the denoising algorithm had good effect, but also it was very well to protect the details.