论文部分内容阅读
为了标定激光平面以实现三维形貌测量,提出一种亚像素级精度激光光条提取算法.该算法包括两侧边检测和中心线提取2个部分.首先,利用基于主成分角的渐进概率霍夫变换检测2条侧边并依赖该两侧边的距离获取光条宽度;然后,应用二维泰勒展式提取具有亚像素级精度的光条中心线并依据重建的三维坐标标定激光平面.实验结果表明,所提算法光条中心提取速度较快,平均约为78 ms/帧,光条平面共面误差较低,限制在0.3 mm以内.因而所提算法能够满足提取光条的两侧边和中心线的需要,且快速可靠、精度高和抗干扰能力强.
In order to calibrate the laser plane to realize the three-dimensional topography measurement, a sub-pixel-level laser light extraction algorithm is proposed, which includes two sides detection and center line extraction.First, using the progressive probability Huo And then get the width of the light bar by using the two-dimensional Taylor expansion to extract the center line of light bar with sub-pixel-level accuracy and calibrate the laser plane according to the reconstructed 3D coordinates. The results show that the proposed algorithm can extract the center of the light bar faster, with an average of about 78 ms / frame, and the plane-plane co-planar error is low, limited to within 0.3 mm. And the center line needs, and fast and reliable, high precision and anti-interference ability.