作　　者： ; ; ; ;

机构地区： 西安电子科技大学电子工程学院

出　　处： 《光子学报》 2003年第12期1502-1505,共4页

摘　　要： 提出了一种基于代价函数和模糊熵的图像分割方法 该方法先用代价函数最小化法对退化的图像作预处理 ,之后 ,利用模糊熵作进一步的处理 实验结果表明 :新方法和一般的阈值分割方法相比 ,不但分割图像的错分率较小 。 Image segmentation is not realized by normal threshold segmentation when objective and background heavily overlap each other. In order to solve this problem, a novel image segmentation approach based on fuzzy entropy and cost function is presented. Firstly, the new method makes pre-processing to corrupted image by method of cost function minimum, secondly, makes further processing using fuzzy entropy. Experimental result shows that the new method not only gives superiority on visual sense effect of image compared with normal thresholding segmentation, but also makes the error classify rate of segmentation image become small.

关 键 词： 代价函数 模糊熵 图像分割

领　　域： [电子电信] [电子电信]