基于极坐标系下二维直方图的图像分割算法

Image Segmentation Algorithm Based on Two-Dimensional Histogram in Polar Coordinate System

基于二维直方图的图像分割算法存在明显误分,且利用二维Renyi熵求解最佳阈值计算量过大.为解决这些问题,提出基于极坐标系下二维直方图的图像分割算法.首先,将图像各像素点表示在极坐标系中,根据各点的极角区分噪声点和非噪声点;然后对噪声点进行平滑处理,处理之后图像各像素点都集中在极坐标系中极角为45°的极径附近.由于滤噪后各像素点的极角差别很小,所以仅利用各点的极径信息即可进行分割阈值的选取,由此将二维问题转化为一维问题,以减少计算量.实验结果表明,该算法分割效果良好,尤其适用于受噪声污染较严重的图片,而且与传统二维算法及其改进算法相比,运行速度有很大提高.

There is obvious wrong segmentation in the image segmentation algorithm which is based on the two-dimensional histogram,and the computational load of solving the optimal threshold by using two-dimensional Renyi entropy is too large. To solve these problems,an image segmentation algorithm based on two-dimensional histogram in polar coordinate system was proposed. Firstly,the pixels of the image wererepresented in polar coordinate system,and the noise points and non-noise points were distinguished according to their polar angles;then the noise points were smoothed. After this procedure,all the pixels of the image wereconcentrated around the polar axis with a polar angle of 45 degrees. Since the differences among the polar angles of these pixels are very small,the segmenta-tion threshold can be selected by using the polar radius information of each point. Thus,the two-dimensional prob-lem is converted into a one-dimensional problem to reduce the computational load. The experimental results show that the algorithm is effective in image segmentation,especially for images with serious noise pollution. Moreover,compared with the traditional two-dimensional algorithm and its improved algorithm,the running speed of this algo-rithm has been greatly improved.