一种图象归一代新算法  

A NEW ALGORITHM FOR IMAGE NORMALIZATION

作　　者：周激流[1] 吕航[1] 朱辉[1] 魏志成[1] 郭晶[1] 

机构地区：[1]四川大学电子信息学院,成都610064

出　　处：《四川大学学报（自然科学版）》2000年第1期55-58,共4页Journal of Sichuan University(Natural Science Edition)

摘　　要：介绍了一种综合利用图象力矩及图象边界提取等方法来实现对数字图象的大小归一化的算法．先对原始图象（待处理图象）和标准图象的灰度日进行边界提取，消除噪声及图象背景的影响．在此基础上，分别计算出原始图象和标准图象的力矩，即可求得大小变化因子．实验结果表明，此算法可获得较好的归一化效果．Authors introduce a new algorithm which can normalize digital images with image moments and image edge detection. First, the edges of the gray images of original image and standard image are detected in order to remove noise and the effects of image background. Secondly,the image moments of original image and standard image are computed respectiVely. Thus, the size factor with the image moments of original image and standard image can be obtained. The experiment shows that this algorithm can Set good normalization results.

关 键 词：模式识别 图象力矩 边界提取 归一化 图象识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]