【摘要】 为了提高虹膜识别的效率和准确性,提出了一种将粗定位和精确定位相结合的虹膜边界提取预处理方法。通过对预处理后的二值化虹膜图像进行粗定位,基于瞳孔边界点控制的最小二乘拟合方法确定虹膜内边界,并采用微积分算子精确检测定位虹膜外边界。最后用CASIA虹膜库中的756幅虹膜图像对算法进行测试,与基于Hough变换的虹膜定位方法关于虹膜内外边界的定位时间和虹膜定位的准确度进行比较。结果表明,该算法具有很好的准确性和识别速度。更多还原

【Abstract】 In order to improve the efficiency and accuracy of iris recognition,it presents a preprocessing method for defining the boundaries of the iris and precise positioning.By coarsely positioning the preprocessed binary iris image,it uses least-squares fitting method to determine the iris inner boundary based on pupil boundary points,detects and locates accurately the outer boundary of iris by operator calculus.It designs the algorithm and tests 756 iris images based on the database CASIA,compares with the iris location method based on Hough transform about positioning time and accuracy of iris location.The results show that the algorithm has good accuracy and recognition speed.更多还原