复杂背景下光条中心提取算法研究

RESEARCH ON LASER STRIPE CENTRE EXTRACTION ALGORITHM UNDER COMPLEX BACKGROUND

在光切法三维表面重建过程中,光条中心坐标的提取至关重要。针对复杂背景下激光图像特点,研究激光条图像的预处理和光条中心提取算法,提出一种两帧平均再相减的改进差影法,较好地实现了复杂背景下激光条图像的分割,选用基于极值法的灰度重心法提取激光条中心线。实验表明该方法实现简单,光条图像处理复杂度低,与经典OTSU方法相比较其处理速度提高了20%左右,精度达到了亚像素级水平。

In the process of three-dimensional surface reconstruction using light section method,the extraction of coordinate of light stripe centre is of great importance.In view of the laser image features under complex background,in the paper we study the laser stripe image pre-processing and the algorithm of laser stripe centre extraction.We suggest an improved difference image method which averages the two frames first and then subtracts the above results,in this way we well achieve the segmentation of laser stripe images under complex background.Meanwhile,we choose the gray centre approach which is based on extreme value method to extract the centreline of laser.Experiments show that the proposed method is simple in implementation and has low complexity in laser stripe image processing.Compared with classic OTSU method,its processing speed improves almost 20% and its accuracy achieves the level of sub-pixel.