Abstract: Iris location, an important preprocessing procedure in iris recognition, is to detect inner and outer boundaries of iris in an iris image. The traditional iris location algorithm is based on simple Hough transform, but this method owns time consuming operation and low location accuracy. According to the grayscale features of eye images, a novel iris location algorithm, which combines mathematical morphology with Hough transform, is presented. The eye image is preprocessed by using morphological filter. The binary image is extracted and processed by morphology for pupil separation. By means of Sobel operator and least square method, the inner-boundary-points of iris are determined and the inner boundary is fitted. The outer boundary is also determined by Hough transform with morphology and the apriori knowledge about iris image. The test results show that the proposed algorithm has better performance in location accuracy and operation speed, and can be suitably used as practical preprocessing algorithm in iris recognition.

Key words: iris location, Hough transform, mathematics morphology, least square method

摘要： 虹膜定位是在虹膜图像中确定虹膜的内外边界，是虹膜识别过程的首要环节。Hough变换是虹膜定位的经典算法，但对原始图像质量要求高，算法运算时间长。依据人眼图像的灰度特性，结合形态学处理提出一种改进的Hough变换定位新算法。对图像进行灰度二值化运算后进行形态学处理分离出瞳孔，结合Sobel算子边缘检测出瞳孔边界点，通过最小二乘法拟合定位出虹膜内边界；在先验知识和形态学处理的基础上对图像进行Hough变换，定位出虹膜的外边界。实验表明所提出的算法性能比传统Hough变换有较大提高，可用于实际虹膜识别的预处理过程中。

关键词: 虹膜定位, Hough变换, 数学形态学, 最小二乘法