一种改进的虹膜快速定位算法  

Improved fast iris localization algorithm

作　　者：朱柳柳[1] 卢光明[1] 

机构地区：[1]哈尔滨工业大学深圳研究生院,广东深圳518055

出　　处：《计算机应用》2008年第3期674-676,共3页journal of Computer Applications

基　　金：国家博士后基金资助项目(20060390069)

摘　　要：针对现有虹膜定位算法的局限性,提出一种检测变形瞳孔的算法,实现了精确快速的虹膜定位,并通过最大类间方差法确定图像阈值。变形瞳孔近似椭圆,因此利用最长弦定位内边界。由于瞳孔虹膜近似同心,利用圆灰度梯度算子小范围搜索外边缘。此算法精确定位了变形瞳孔,避免了外边缘搜索的盲目性,提高了虹膜定位的精度与速度。For the purpose of detecting iris precisely and quickly, an ellipse pupil based detection algorithm was proposed based on some prevailing localization algorithms. After determining the threshold by OTSU algorithm, the longest-chord based method detected the boundary of pupil more effectively than the prevailing algorithms based on circular template. The outer boundary of iris based on circular template was searched in less range as the pupil and iris were approximately concentric. Compared with the usual algorithms based on circular template, this algorithm detects pupil as ellipse, reduces the blindness of the search process, and improves the speed and accuracy.

关 键 词：虹膜识别 虹膜定位 最大类间方差法 圆检测算子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]