提出了一种基于TV-norm的阈值迭代分割算法.该算法首先用局部阈值迭代法确定各个子图像的阈值,得到一个待处理的矩阵;然后利用TV-norm对阈值矩阵进行插值处理,使之成为与原图像像素数目相同的新矩阵,得到光滑的阚值曲面,进而实现对图像的分割;最后通过仿真试验说明该算法的可行性和有效性.
In this article, an image segmentation method based on PDE iterative is proposed. First, it takes an iterative partial threshold to determine the threshold of each sub-image. Then the TV-norm iterative is used to deal with the threshold matrix. It makes the matrix transform to a new threshold array which has same amount of pixels of original image, obtains the smooth threshold curved surface, then realization the image segmentation. Finally it proves the validity and the feasibility of these algorithms by simulation experiment.