基于互补曲线拟合的格雷码亚像素边缘定位  

Gray Code Sub-pixel Edge Localization Based on the Complementary Curve Fitting

作　　者：于微波[1] 王国秀 李洋 李岩[1] 刘克平[1] YU Wei-bo;WANG Guo-xiu;LI Yang;LI Yan;LIU Ke-ping(School of Electrical and Electronic Engineering, Changchun University of Technology, Changchun 130012, China)

机构地区：[1]长春工业大学电气与电子工程学院,长春130012

出　　处：《科学技术与工程》2021年第21期8999-9004,共6页Science Technology and Engineering

基　　金：吉林省发改委省级产业创新专项(2019C010);吉林省发改委项目(2020C018-1)。

摘　　要：精确定位格雷码光条纹边缘是结构光三维重建系统的关键问题之一。提出了一种互补曲线拟合的光条纹亚像素边缘定位算法,在对互补编码图像进行光强成分分析并二值化处理的基础上,通过边缘检测算法获得准确的初始像素边缘,然后映射回互补编码图像并对其边缘过渡区域进行拟合,最后通过计算互补曲线的交点得到格雷码光条纹的亚像素边缘点。采用平均相对等宽性误差进行评价,得到的最小边缘定位误差为1.97277×10^(-3),分别对不同投影方向和条纹数量的图像进行处理,都保持了较小的定位误差。实验结果表明,提出的算法能够很好地解决初始边缘定位困难的问题,具有较好的准确性和鲁棒性,提高了光条纹定位精度。Precisely locating the edges of gray code optical stripes is one of the key problems in structured light 3D reconstruction systems.An algorithm for light-stripe subpixel edge localization by complementary curve fitting was proposed.Based on the analysis of the light intensity components of the complementary encoded image and binarization processing,the accurate initial pixel edge was obtained by the edge detection algorithm,then the complementary coded image were mapped back to the complementary coded image and the edge transition region was fitted,finally the subpixel edge point of the gray code light stripe was obtained by calculating the intersection point of the complementary curve.The minimum edge localization error obtained by using the average relative equal-width error for evaluation was 1.97277×10^(-3),and the images with different projection directions and number of stripes were processed separately,and all maintained small localization errors.The experimental results show that the proposed algorithm can solve the problem of difficult initial edge localization with good accuracy and robustness,and improve the accuracy of optical stripe localization.

关 键 词：三维重建 结构光 格雷码 互补编码 亚像素检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]