最小方差搜索圆心的快速虹膜定位算法

A Fast Iris Location Algorithm Based on Minimal Variance Searching Center

为了改善虹膜识别的实时性,提出一种新的快速虹膜定位方法。首先对虹膜图像进行去噪处理,然后采用类间方差法对图像进行阈值分割,再运用投影方法粗略得到虹膜内边缘圆心和半径,最后根据粗定位得到内边缘圆周参数,采用所提出的算法对虹膜内边缘进行精定位;对于外边缘定位,依据先验知识以及内边缘圆周参数去掉虹膜图像多余的边缘点及噪声点,缩小搜索范围,然后采用同样的算法对外边缘进行精定位。实验结果表明,该方法能够准确快速的定位出虹膜内外边缘,定位速度较传统算法提高了十倍左右,并且减少了传统定位算法搜索的盲目性。

In order to improve the real-time characters of iris identification, this paper presents a new fast iris location algorithm. Firstly, the iris image was denoised and the threshold segmentation was carried out by applying the OSTU method. Then the rough center point and radius of the inner edge of iris were obtained by using the projection method. At last, according to the obtained rough center point and radius, the inner edge of iris was located accurately by applying the presented algorithm in this paper. For the outer edge’s location, based on the prior knowledge and the gained circle parameters of inner edge, the extra edge points and noise points were removed and the searching range was reduced, and then the accurate center point and radius of the outer edge of iris was gained by using the same way with the inner location. Experimental results show that this algorithm can determine the inner and outer edge of iris quickly and accurately, the location speed is nearly ten times faster than those traditional algorithms, and reduce the searching blindness.