基于可控图像分割的快速视网膜血管提取算法  

Fast retinal blood vessel extraction algorithm based on controlled image segmentation

作　　者：赖小波[1] 刘华山[2] 方纯洁[1] 

机构地区：[1]浙江中医药大学信息技术学院,浙江杭州310053 [2]东华大学信息科学与技术学院,上海201620

出　　处：《光电子．激光》2013年第5期1018-1025,共8页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金(61203337);浙江省自然科学基金(LQ12F01004);浙江中医药大学校级科研基金重点(2011ZZ10)资助项目

摘　　要：针对多数视网膜血管提取算法实时性不强和分割精度不高的问题,提出了一种基于可控图像分割的快速视网膜血管提取算法。首先,对视网膜G分量图像的灰度进行反转和自适应直方图均衡化,应用结构元素为"菱形"和"圆盘形"的形态学"开"运算平滑图像背景和增强血管对比度,消除视盘后阈值分割并二值化得到不含视盘的分割图像。其次,根据在灰度图像中检测到的视盘构建掩膜,再次对视网膜绿色分量图像自适应直方图均衡化后进行阈值分割,并和掩膜进行逻辑"与"运算得到含有掩膜的分割图像。最后,将不含视盘的分割图像与含有掩膜的分割图像进行逻辑"与"运算,并融合边界信息获得最终的视网膜血管结构。实验结果表明,本文算法能有效提取视网膜眼底图像的血管网络,有较强的实时性和较高的分割精度。Aiming at the issues that the real time is not strong and the segmentation accuracy is not high in majority of retinal blood vessel extraction algorithms,a fast retinal blood vessel extraction algorithm based on controlled image segmentation is proposed. Firstly, the intensity of the green channel image is inversed and adaptive histogram equalization is applied, the morphological ＇opening＇ operation is conducted using the ＇diamond structuring element and the ＇disc＇ structuring element to smooth the image background and highlight the blood vessels, the optical disk is removed before binarizing the image by thresholding method,and the segmented image without the optical disk is obtained. Secondly, a mask is created based on the optic disc detected in grayscale image, the green component image is also applied with adaptive histogram equalization and threshold segmentation, after which the segmented image with a mask is obtained using ＇ AND＇ logic operation with the mask. Finally, the segmented image with noise removed is compared with that with a mask according to ＇AND＇ logic operation,and the final image of blood vessels is obtained together with the border information. The experiment results indicate that the proposed algorithm can effectively detect the blood vessels＇ network of fundus image, and it also has strong real time ability and high segmentation accuracy.

关 键 词：视网膜图像 血管提取 可控图像分割 自适应直方图均衡化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]