基于投影约束和随机采样Hough变换的虹膜分割方法  

Iris Segmentation Method Based on Projection Constraints and Random Sample Consensus Hough Transform

作　　者：刘卓 刘晓敏[2] 刘金明[1] LIU Zhuo;LIU Xiaomin;LIU Jinming(College of Information and Electrical Engineering,Heilongjiang Bayi Agricultural University,Daqing Heilongjiang 163319,China;School of Information and Electrical Technology,Jiamusi University,Jiamusi Heilongjiang 154007,China)

机构地区：[1]黑龙江八一农垦大学信息与电气工程学院,黑龙江大庆163319 [2]佳木斯大学信息电子技术学院,黑龙江佳木斯154007

出　　处：《佳木斯大学学报（自然科学版）》2024年第3期21-26,共6页Journal of Jiamusi University：Natural Science Edition

基　　金：黑龙江省自然科学基金项目(LH2022E099);黑龙江省自然科学基金项目(LH2022F052)。

摘　　要：非合作式虹膜图像的采集导致图像中具有大量干扰和噪声,因此,提出基于投影约束和随机采样Hough变换的虹膜分割方法提取虹膜有效区域。对虹膜二值图像进行投影粗定位虹膜的外轮廓圆心和半径,通过圆心和半径确定虹膜的大致区域,在该区域上,使用随机采样的Hough变换确定虹膜的内外轮廓,从而提高了虹膜分割的准确性。实验证明,与现有方法相比虹膜分割方法错误指标E1和E2最多提高8%和20%,具有更低的错误率和较高的时间效率。The acquisition of non-cooperative iris images results in a large amount of interference and noise in the images.Therefore,a iris segmentation method based on projection constraints and random sample consensus Hough transform is proposed to extract the effective area of the iris.By projecting the binary image of the iris,the approximate area of the iris is determined by the center and radius.In this area,the random sample consensus Hough transform is utilized to determine the inner and outer contours of the iris improving the accuracy of iris segmentation.Experimental results have shown that compared with existing methods,the error metrics E1 and E2 of iris segmentation methods can be improved by up to 8%and 20%,which have the most accurracy rate and appropriate time efficiency.

关 键 词：虹膜分割 图像二值化 随机采样 HOUGH变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]