基于改进型最小二乘法拟合的虹膜定位

Iris localization algorithm with improved least squares fitting

针对传统虹膜定位算法计算速度慢、内存占用大的缺点,提出最小二乘法拟合圆曲线的虹膜定位算法。该算法根据虹膜区域图像和周边灰度值相差异的特点,先利用Canny算子探测虹膜图像边缘,设定合适的阈值将图像二值化;再用最小二乘法勾勒虹膜内边界和拟合外边界,并计算出虹膜内外边缘的坐标和半径。从CASIA图像数据库抽取100幅图像进行实验,结果显示所提出的算法准确率达99%,而且比传统的定位算法快150ms,减少虹膜定位计算的复杂度和内存空间。

For the traditional iris localization algorithm is slow and takes up big storage space, an iris localization algorithm based on least squares fitting circular curve is proposed in this paper. According to the difference of gray value between iris image and iris surrounding, the proposed algorithm uses Canny operator to detect image edge, makes the image binarization with a right threshold, and then outlining iris inner boundary and fitting iris image outer boundary are given by least squares approximation,and the coordinates and radii of iris＇ inner and outer edges are calculated. 100 images extracted from the image database CASIA were experimented. The results show that the accuracy of the proposed algorithm is up to 90 percent, and the speed is 150 ms faster than the traditional localization algorithm. The algorithm reduces the complexity of the calculation and saves memory space.