摘要
为了提高图像分割的精度和效率,提出了一种基于遗传算法的二维双阈值最大类间方差法的图像分割算法。该方法用两个阈值同时考虑了图像像元点灰度信息的范围和像元点之间相关信息,并对未处理区域做了后处理,以提高分割的精度和鲁棒性。为了提高分割速度,给出了一种改进遗传模拟退火算法。实验结果表明,文中提出的方法比传统的二维最大类间方差法更加完整、快速地分割出复杂背景中的图像。
To improve the accuracy and speed of image segmentation, a genetic algorithm based on the threshold of two-dimensional two-Otsu method of image segmentation is put forward. The method is to use two thresholds at the same time to consider the reference of image pixel grayscale point and information point between the pixel information. Finally, a post-processing has been done on the untreated region. In order to increase the speed of segmentation, an improved genetic simulated annealing is given. The experimental result shows that the proposed method is more complete and can divide a complex background image with greater rapidity than the traditional two-dimensional Otsu.
出处
《电子科技》
2009年第11期35-39,共5页
Electronic Science and Technology
关键词
二维双阈值最大类间方差法
图像分割
后处理
遗传模拟退火算法
two-dimensional dual-threshold Otsu law
image segmentation
post-processing
genetic simulated annealing
