一种基于多阈值的灰度图像去噪算法

A Removing Noise Algorithm of Gray Image Based on Multi-Threshold

对灰度图像的去噪是图像处理技术中的重要组成部分,能够将图像中目标部分的主题信息凸显出来是一项具有重要意义的工作。提出一种灰度图像的去噪算法。本算法首先利用区域增长技术将图像分为不同的区域,这些区域里的像素之间存在一定的相似性,满足一定的相似规则;然后,根据每个区域的像素,计算出各自的适合区域特征的阈值;最后,根据每个阈值分别对每个区域二值化,得到一二值化图像。本算法区别于全局阈值法和局部阈值法,实验结果证明,本算法在提取灰度图像主题信息具有较为明显的效果。

Denoising in the grayscale image is an important part of image processing technology, and showing the prominent theme of an image is an important work. In this paper, a denoising algorithm of grayscale image is present. Firstly, the growing technique is used to divide the image into different regions, the pixels in these regions have some similarities, and meet a certain similarity rule;Then, the threshold of each region is calculated on the base of the pixels in respective region; Finally, the image is changed into a binary one by threshold. This algorithm is different from the global threshold method and the local threshold method, the experimental results are proved that the algorithm is effective in extracting information on gray image.