基于形态学滤波和Hough变换的掌纹主线提取方法  

A Palm-Print Principal Line Extraction Based on Morphological Filter and Hough Trans-Formation

作　　者：郑艳清[1] 裘正定[2] 吴介[3] 

机构地区：[1]哈尔滨师范大学计算机科学与信息工程学院,哈尔滨150025 [2]京交通大学计算机与信息技术学院,北京100044 [3]北京交通大学计算机与信息技术学院,北京100044

出　　处：《北京交通大学学报》2008年第5期107-110,共4页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家自然科学基金资助项目(60773015);教育部博士点基金(20060004007);黑龙江省自然科学基金(F200605)

摘　　要：提出一种结合形态学滤波和Hough变换提取掌纹主线的方法,在对方向性掌纹线特征子图像做形态学滤波后,将滤波图像转换至Hough变换参数空间,利用Hough变换具有的点线对偶性,检测参数空间内累加数组的峰值来确定主线的空间位置和走向,并展开邻域检测得到主线.本文算法更符合主线的自然生长规律,在成功提取主线的同时消除了粗大掌纹皱纹在主线提取中易产生的干扰.This paper proposed a principal line extraction method of directional palm-print line by combining morphologically filter and Hough transformation. After the directional palm-print line feature images were morphologically filtered and converted to Hough transformation space, and then by using the point-line dualitv to detect the location and direction of the principle lines in the line features image through finding the peak values in the transformed space, the neighborhood searching was also developed to ensure pixels on each principal line. Principal lines derived by the proposed method were natural and clear. The method also eliminated the disturbance caused by thick wrinkles during the extraction of principal lines.

关 键 词：掌纹 主线 特征提取 形态学滤波 HOUGH变换 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]