图象阈值分割的统计迭代算法  

The Image's Segmentation by the Threshold Method

作　　者：王耀明[1] 董建萍[1] 严炜[1] 陈以[2] 

机构地区：[1]上海师范大学,200234 [2]桂林电子工业学院计算机分院

出　　处：《桂林电子工业学院学报》2000年第2期6-8,共3页Journal of Guilin Institute of Electronic Technology

摘　　要：图象分割是图象分析和计算机视觉等学科中最基本也是最困难的课题之一。直方图技术是图象分割的重要技术 ,而灰度图象的象素灰度分布近似符合高斯概率分布 ,本文是把直方图技术和高斯统计分布函数有效地结合起来 ,作为求取图象二值化阈值的理论基础。并获得其理论表达式。由此 ,通过对该表达式的分析 。Image segment is one of the most fundamental and difficult problems in image analysis and computer vision. Histogram technique is an important technique of image segment. The distribution of a gray level image is often similar to that of Gaussian distribution. So the article combines the histogram technique with Gaussian distribution function, achieving a theory basis of image segment.From the analysis of the theoretical expression of this theory,the article gets the iterative formulas and steps on how to work out the threshold in image segment.At last,the article uses two examples to prove this method.

关 键 词：灰度直方图 图像分割 阈值法 统计迭代算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]