系统阐述了虹膜识别的原理和实现方法,提出将小波变换与Hough变换相结合定位虹膜外边界,采用同心圆渐变展开虹膜图像为矩形图像,根据纹理分布方差对分类距离加权.新算法在保持平移和缩放不变性的同时,有效地实现了虹膜识别.
The principle and algorithm of iris recognition were expounded systematically. Three new ideas were presented, locating the outside edge effectively by combining wavelet transform and Hough transform, spreading the iris image to the rectangular image by gradually changing concentric circles, and weighting the classification distance based on the distribution variance of texture feature. This algorithm retains the characteristics of rotation, translation and scale invariants and realizes the iris recognition effectively.