提出了一种快速而准确的虹膜识别算法,基本思想是:通过图像预处理,确定虹膜的位置和大小,并将直角坐标系下的环形虹膜展开成极坐标系下的矩形虹膜;利用寻找局部灰度极小值点的方法寻找图像中每条扫描线上的特征点,去掉伪特征点以后,得到固定长度的虹膜编码;计算两个虹膜编码之间的海明距离,根据海明距离给出识别结果。
This paper described a rapid and accurate iris recognition method. Firstly, got the location and size of the iris through the image preprocessing, and then unwrap the localized iris to a rectangular texture block with a fixed size. Secondly, detected all of the feature points on each scan lines in the image utilizing the method of finding local gray minimum points, and then abandoned the false feature points and got a fixed length iris code. Thirdly, calculated the Hamming distance between two iris codes and then obtained the recognition result according to the Hamming distance.