基于分水岭模型的光照不均匀图像分割  

Non-Uniform Illumination Image Segmentation Method Based on Watershed Model

作　　者：王宏文[1] 宁乐[1] 雷盼雲 郭章亮 

机构地区：[1]河北工业大学控制科学与工程学院,天津300130

出　　处：《中南民族大学学报（自然科学版）》2016年第3期85-91,共7页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：天津市科技支撑项目(14ZCDZGX00818)

摘　　要：针对图像采集过程中存在光照不均的问题,首次利用分水岭模型将图像分割为不同的光照区域,再利用Sauvola算法自适应地获得局部的二值化阈值,从而达到了良好的分割效果.同时Sauvola算法使用积分图像计算出局部像素值的和与局部像素值平方和来加快算法的时间.实验结果表明:该算法能够在无人工干预的情况下克服光照的影响,并且图像分割时间复杂度大大减小.Aiming at the problem of uneven illumination in the process of image acquisition, in this paper,we firstly use the watershed model to divide image into different light regions, and then use the Sauvola algorithm to obtain the local binary threshold value adaptively, so as to achieve a good segmentation. Meanwhile we use Sauvola algorithm to calculate the local pixel value sum and the quadratic sum by using Integral Image to reduce the running time of the algorithm. Experimental results show that in the absence of human intervention, the algorithm can overcome the effects of light and image segmentation time complexity is greatly reduced.

关 键 词：分水岭模型 光照不均匀 二值化 积分图像 猴王遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]