一种快速光栅条纹中心亚像素坐标提取方法

光栅条纹中心的精确提取会直接影响到线结构光视觉的测量与重建精度。为了解决传统的Steger算法计算量大,且在光条不均匀时提取不精确等问题。基于改进的Steger算法,提出了针对光栅条纹中心的快速亚像素坐标提取方法。通过高斯滤波、形态学等方法得到去复杂背景的激光条纹图像,用自适应阈值法缩小ROI区域,用霍夫变化求取光栅边缘线,求出边缘线中点,然后直线拟合,最后用改进算法提取中心坐标。实验结果表明:与模板法、传统Steger算法等光栅条纹中心提取算法相比,算法的鲁棒性更强,最大提取误差在0.3个像素,平均误差为0.1个像素,平均运行时间相比传统算法减少0.4 s。

A method for extracting sub-pixel coordinates of grating fringe centre

The accurate extraction of the grating fringe centre will directly affect line structured light vision's measurement and reconstruction accuracy.To solve traditional Steger algorithm problems,such as a large amount of calculation and inaccurate extraction when the light stripe is not uniform.An improved Steger algorithm for fast sub-pixel coordinate extraction of the grating fringe centre is proposed.Through Gaussian filtering,morphology and other methods to get the laser stripe image with a complex background,using adaptive threshold method to reduce ROI region,using Hough change to get the edge line of the grating,to find the midpoint of the edge line,and then to fit the straight line,and finally to extract the centre coordinates with the improved algorithm.The experimental results show that:compared with the template method,the traditional Steger algorithm and other algorithms,the algorithm is more robust,the maximum extraction error is 0.3 pixels,the average error is 0.1 pixels,and 0.4 s reduce the average running time compared with the traditional algorithm.