摘要
虹膜定位算法是虹膜识别的关键步骤之一,定位的速度和准确度是衡量虹膜定位算法优越性的标准。本文根据瞳孔区域独特的灰度特征提出了一种快速虹膜定位算法,即先粗定位瞳孔区域,再利用形态学和几何学方法精确定位瞳孔,最后利用微分积分算子在瞳孔定位后分割的区域内进行外圆定位,对中科院虹膜库Version1.0中的756幅图像进行定位,准确率达到98.80%,平均定位时间为1.5s,比直接应用Hough变换定位虹膜在速度上有了很大提高。
The algorithm of iris location is a pivotal step in iris recognition and the speed and accuracy of location is the standard to evaluate the effectiveness of the algorithm. According to the grayscale characteristic of the pupil area, a fast iris location algorithm is presented based on image sub block. Firstly, find an approximate location area of the pupil roughly. Then, locate it accurately by using the methods of morphology and geometry. Finally, locate the outer circle with integral differential operator in the delimited area after pupil location. An experiment on iris location with 756 images in iris image library of version 1.0 of Chinese Academy of Sciences has been completed. The results with 98.80% accuracy and 1.5s average positioning time shows that the algorithm improves recognition speed notably .
出处
《光电工程》
CAS
CSCD
北大核心
2010年第3期127-132,共6页
Opto-Electronic Engineering
基金
教育部"新世纪优秀人才支持计划"(NCET2006348)
江苏省"333工程"(2007-16-59)资助项目
关键词
虹膜定位
分块统计
圆检测
HOUGH变换
iris location
sub block statistics
circle detection
Hough transform