摘要
虹膜图像的边缘提取是采用Hough变换结合边缘检测进行虹膜定位的关键步骤.提出一种基于数学形态学算子的虹膜定位方法,内边界采用二值化的形态学边缘检测算子;运用多尺度边缘检测方法以获取虹膜灰度图像的外边界和眼睑轮廓图,利用连通区域的标记进一步去除随机噪声,得到较理想的图像边缘,大大减少了Hough变换搜索的时间,并与经典的定位方法进行比较.实验表明,该算法抗噪性能良好,具有一定的应用价值.
Edge detection of iris image is the key step of iris location that used Hough transform and edge detection. An iris location method based on mathematical morphology operator is proposed in the paper. Morphological edge detection operator of binary image is used for inner boundary. The outer boundary and eyelid contour are obtained by a multi-scale edge detection method, using label of regional connectivity to further remove the random noise and obtain an ideal image edge, which greatly reduce the Hough transform search time compared to the classic method of iris location. The experiment shows that the algorithm has better performance of noise reduction, which has certain application value.
出处
《微电子学与计算机》
CSCD
北大核心
2009年第6期129-132,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(60472006)
广东韶关学院校级科研项目(2007年220文件)
关键词
虹膜定位
数学形态学
多尺度
多结构
眼睑
iris location
mathematical morphology
multi-scale, multi-structure
eyelid
