快速准确的虹膜定位算法  

Fast and accurate algorithm for iris boundaries location

作　　者：周俊[1] 李王辉[2] 罗挺[3] 杨眉[4] 

机构地区：[1]后勤工程学院研究生1队,重庆400016 [2]鹤壁职业技术学院电子信息工程系,河南鹤壁458030 [3]后勤工程学院训练部模拟中心,重庆400016 [4]重庆师范大学,重庆400030

出　　处：《计算机工程与设计》2010年第18期4058-4061,共4页Computer Engineering and Design

摘　　要：针对经典虹膜定位算法速度较慢的问题,提出一种基于圆几何特征的粗定位与精定位结合的快速虹膜定位算法。基于行列扫描的方法找到瞳孔内4条弦,利用弦的中垂线结合瞳孔边界阈值定位四对切点,取切点坐标参数的均值为内边缘粗略定位参数,然后利用微积分方法对内边缘进行精确定位;利用内外边缘圆心非常接近的先验知识,缩小微积分方法定位外边缘的搜索范围,从而显著提高了虹膜定位速度。实验结果表明,该算法显著提高了虹膜定位速度且具有较高的定位精度。to the slow speed problem of classic iris location algorithm, a fast iris location algorithm is proposed, which com-bined with fine positioning and coarse positioning based on the geometrical characteristic of circle.bined with fine positioning and coarse positioning based onthe geometrical characteristic of circle. First, four strings are found in iuner boundary based on the row and column scanning method, using perpendicular bisector of strings combine boundary threshold to locate four pairs oftangency point, making use of the mean of tangent point coordinates parameters as the coarse inner edge parameters, then use the priori knowledge that the center of inner boundary and outer boundary is very close, narrowing the search range of calculus method to locate outer boundary, and thus significantly improve the locating speed. The expermental results show that the algorithm has high location accuracy and improves location speed markedly.

关 键 词：虹膜定位 圆几何特征 内边缘粗定位 外边缘定位 微积分方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]