边缘检测眼底图像屈光度测量的研究

Research on diopter measurement of detecting edge of eye retina image

为了解决人眼屈光度测量精度问题,提出了一套基于眼底视网膜图像处理的眼睛屈光度参数测量方法。该方法将从眼底视网膜上反射成像在CCD上的六点红外图像采集到计算机中处理。首先定位六点区域,然后利用边缘CANNY检测算子对六个区域分别图像去噪、平滑处理和边缘检测,最后再用最小二乘拟合的方法计算测量图像的中心,求出眼睛屈光度。实验用标准模拟眼进行了检测,并与二值化法进行了比较,精度明显提高。实验结果表明,该方法使测量精度能达到±0.25D,达到国家计量部门的要求。

In order to solve the precision of diopter measurement,a new parameters measurement method is proposed based on processing image of eye retina.Computer processes six dots infrared image displayed in CCD reflected from eye retina. Firstly,six dots regions are located.Secondly,the image with six dots is respectively processed with CANNY operator such as image denoising,smoothing and edge detection.Thirdly,the image center can be obtained by least-squares fitting method.Finally,the parameter of diopter is acquired by calculating equation of ellipse.The experiment is carried through by using the standard analogue eye,and comparison with result of binary method.Results show measurement error is less than ± 0.25D which is minimum error of diopter prescribed by National Metrology Bureau.