The accurate detection of optic disc is important for the accurate detection of other features of retinal fundus images such as blood vessels and microaneurysm. This paper presents an optic disc detection method which combines brightness and circularity characteristics of optic disc. Based on the detection method, the fundus images would be preprocessed ifrstly. Then the mathematical morphology operation is introduced to remove the interference of blood vessels, and threshold segmentation is used to identify the potential optic disc regions. Then the authentic optic disc regions would be detected according to the area and circularity characteristics of optic disc. Experimental results shows that the algorithm owns high localization accuracy and fast localization speed.%视盘的准确检测对视网膜眼底图像的其他特征,如血管和微动脉瘤的精确检测非常重要。本文提出一种结合视盘亮度和圆形度特征的视盘检测方法,先对眼底图像进行预处理,然后运用数学形态学运算去除血管的干扰,采用阈值分割提取出可能为视盘的区域,最后再根据视盘的面积和圆形度特征检测出真正的视盘区域。实验结果表明,本文算法具有较高的定位精度和较快的定位速度。
展开▼