激光与光电子学进展, 2017, 54 (10): 101004, 网络出版: 2017-10-09
角膜地形图仪图像的边缘检测及形貌重构
Image Edge Detection and Shape Reconstruction in Corneal Topography
摘要
为了实现所研制的角膜地形图仪应用于临床检验, 将Klyce经典的三维形貌重构模型改成光线逆追踪法进行建模: 获得24环每环256个点的亚像素级二维数据; 将这些数据由最小二乘法估算出原始假设椭球面; 对该椭球面进行修正, 得到最符合真实角膜的曲面。先采用Canny算法得到像素级的二维数据; 然后对数据的灰度进行平滑降噪; 之后对数据的灰度梯度进行3次拟合以获得亚像素级数据; 最后, 用标准的角膜地形图仪获得的清晰图片实现三维重构。通过分析两者获得的曲率半径及屈光度数据, 发现其误差小于0.9%, 表明该方法具有可行性。
Abstract
In order to achieve the application of corneal topography instrument in clinical examination, the traditional Klyce three-dimensional reconstruction of the surface model is changed into the ray tracing method. 24 rings and 256 sub-pixel data per ring are obtained. The original assumed ellipsoid is estimated by the least squares method. The ellipsoid is corrected to obtain the most realistic corneal surface. First of all, the pixel level data information is obtained by the Canny algorithm. Then, the gray value of the data is smoothed to reduce the noise. And the gray gradient of the data is fitted 3 times to obtain the sub-pixel data. Finally, the clear images obtained by standard corneal topography are used for the three-dimensional reconstruction. By analyzing the curvature radii and diopter data obtained by both methods, the results show that the errors are less than 0.9%, and it proves that the proposed method is feasible.
徐丹阳, 高建勋, 隋成华, 周明华. 角膜地形图仪图像的边缘检测及形貌重构[J]. 激光与光电子学进展, 2017, 54(10): 101004. Xu Danyang, Gao Jianxun, Sui Chenghua1, Zhou Minghua. Image Edge Detection and Shape Reconstruction in Corneal Topography[J]. Laser & Optoelectronics Progress, 2017, 54(10): 101004.