中文摘要：

在经典过零点算法的基础上,改进特征提取与匹配算法。分析各阶高通小波系数在同一算法下的识别率,以此选择小波系数来综合编码分析。采用不同的小波滤波器进行分析,考察不同的小波滤波器对算法的影响。DB3小波的识别率为99.61%,等错率为0.41%;四次B样条小波识别率为99.75%,等错率为0.25%;Coif3小波识别率为99.72%,等错率为0.29%。第五、六阶小波高通系数涵盖了虹膜纹理的主要信息,可采用第五、六阶小波系数进行编码识别。不同的小波滤波器得到的结果区别不是很大,DB3、四次B样条、Coif3小波编码识别效果好。DB3小波符合识别速度与效果的综合要求。

英文摘要：

On the classic arithmetic of the wavelets zero-crossing, found a modification of the feature exactions and matching. Firstly analyzed the CRR and the EER in different resolution. The right order would be selected to encode. Using different wavelet filters to analyze the influence on the recognition result. CRR of the DB3 is 99.61% , EER is 0.41%. CRR of the forth-order B-Spline wavelet is 99.75% ,EER is 0.25% o CRR of Coif3 is 99.72% ,EER is 0.29% . The fifth and the sixth resolution of the wavelets coefficient covered the most important information of the iris. The different wavelet filters had little influence on the result. The fifth and the sixth order of the wavelet coefficient could be used to iris recognition. DB3, the forth-order B-Spline, Coil3 wavelet have good recognition results. DB3 can fulfill the demands on the speed and the recognition result of the biologic characteristics recognition.