基于二维小波变换及邻域均值滤波的虹膜定位算法  

Iris Localization Algorithm Based on Two-dimensional Wavelet Transform and Neighborhood Average Filter

作　　者：赵静[1] 

机构地区：[1]宁夏大学数学计算机学院,宁夏银川750021

出　　处：《宁夏大学学报（自然科学版）》2013年第1期22-25,共4页Journal of Ningxia University(Natural Science Edition)

基　　金：宁夏自然科学基金资助项目(NZ1137);宁夏大学科研基金资助项目(NDZR10-22)

摘　　要：为了提高虹膜定位的准确率和速度,提出了一种基于二维小波变换及邻域均值滤波的虹膜定位算法.采用阈值法分割瞳孔,使用边缘检测算子检测瞳孔区域边缘,定位虹膜内边缘;然后对人眼图像进行二维小波处理降低虹膜图像的分辨率,以减少虹膜本身的纹理对判断外边缘点时所产生的影响;最后采用邻域均值滤波进行虹膜外边缘点提取,根据所得虹膜外边缘点确定虹膜外边界.仿真结果表明:该算法定位虹膜内外边界的平均时间为1.75s,准确率为99.7%,其中虹膜外边缘定位误差小于4.2%,在虹膜识别系统中有较高的实际应用价值.An iris localization algorithm based on two-dimensional wavelet transform and neighborhood average filter is proposed to improve the accuracy and the speed of the iris localization. Firstly, the algorithm segments the pupil area of the iris by the threshold. Secondly, it locates the iris inner edge by the edge detection operator in the pupil area. Thirdly, the human eye iris images is processed by the twodimensional wavelet transform to reduce the image resolution, In order to reduce the impact of the iris texture on the judgment of the iris outer edge points. Fourthly, the algorithm extracts the iris outer edge points by the neighborhood average filter. Finally, it locates the iris outer edge by the outer edge points. The simulation results show that the algorithm locates the iris inner and outer edge with average time of 1.75 s and accuracy of 99.7%, the error of iris outer edge localization is less than 4.2%. The algorithm has a higher practical value in the iris recognition system.

关 键 词：虹膜识别 虹膜定位 二维小波变换 邻域均值滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]