摘要
本文提出了一种新颖的虹膜图像预处理方法,该方法打破了传统的定位思想,并且将噪声检测的过程由空间域转到频域来完成。首先,在原始图像上填充瞳孔内的光斑;其次,将人眼图像展开成矩形,在矩形图像上以点线检测代替传统的圆检测确定虹膜的内外边界,并进行归一化;最后在归一化后的虹膜图像上利用2D-Gabor滤波器的方向选择性检测眼毛和眼睑,标记干扰区域。实验证明,该方法可以解决内边界不是理想圆造成的瞳孔遗留或纹理损失问题,而且显著降低了虹膜定位时间,提高了虹膜识别的准确性。
A new iris pre-processing method was proposed, which breaks the traditional localizing principle. The process of noise detection was achieved by converting the spatial domain into the frequency domain. The detailed steps were as follows： Firstly, the light spot within the pupil was filled in the original image. And then the image was unfolded into a rectangle, the circle detection was substituted by the point and line detection in the rectangle image to find the inner and outer edge, and the normalization was also performed. Finally, eyelash and eyelid were detected and disturbed region was marked as well by taking advantage of the direction selectivity of 2D-Gabor filter in the normalized iris image. The results prove that this method not only solves the problem of pupil carry-over or the texture loss caused by the inner edge which is not an ideal circle, but also sharply reduces the location time and enhances the accuracy of iris recognition.
出处
《光电工程》
CAS
CSCD
北大核心
2009年第4期133-139,共7页
Opto-Electronic Engineering
基金
国家自然基金资助项目(60672078
60472088)
关键词
图像预处理
虹膜定位
真实内边界
去除噪声
归一化
image pre-processing
iris location
real inner edge
de-noising
normalization
