人以正常步行速度通过采集装置会出现非理想的虹膜图像.本文针对非理想虹膜中的角度偏离图像,提出了一种基于椭圆投影的虹膜定位算法.该算法首先计算双眼相对位置来决定头部的旋转角度,之后找到瞳孔内一点,从该点出发根据梯度变化搜索到虹膜的内边界点,采用最小二乘法拟合椭圆形内边界.然后依据椭圆与圆形的近似度检测人眼凝视方向,使用投影变换将角度偏离的虹膜图像转换成正视角度.最后在已经校正到正视角度的虹膜图像上,使用边界点优选方法搜索到虹膜的外边界点,并利用三点确定圆的方法定位外边界.实验结果表明本算法可以在较短的时间内对角度偏离的非理想虹膜有效地定位.
There will be non-ideal iris images when the person is moving at a normal walking pace through an iris capture device.An iris localization method based on elliptical projection is proposed,which solved the localization problem of non-ideal off-angle iris images.Firstly,relative position between two eyes is calculated to determine the rotation angle of the head.Then a point within pupil is easily found.Starting from that point,iris inner boundary points are searched according to gradient variation.An elliptical inner boundary is fit using the least squares.Secondly,eye gaze direction is detected by the circularity of ellipse,and an off-angle iris image is converted into a frontal view iris image adopting projection transform.Thirdly,iris outer boundary points are searched using the boundary point optimal selection method in the frontal view iris images,and outer boundary is located by the means of three points fitting a circle.Experimental results show that the proposed algorithm can effectively locate off-angle iris in a very short time.