基于计算机视觉的自动矫正数控系统

Computer Vision-based Automatic Correction Numerical Control System

基于机器视觉技术,利用OpenCV计算机视觉开源库作为开发函数库,以工业相机和三轴雕刻机作为实验平台,为数控系统进行视觉矫正系统开发.首先对装夹后PCB板采集Mark点,直线检测完成后,利用角点检测对Mark点进行定位.通过霍夫变换得到世界坐标系下的直线方程,以此求取当前待加工PCB板的装夹位置和姿态.最后通过图像坐标和机械坐标的变换关系对加工位置信息进行纠偏处理.针对数控系统PCB板加工中出现的装夹误差问题,设计并开发了基于机器视觉技术的数控系统视觉矫正系统.通过实验和具体误差数据证明了本系统达到了PCB板的加工精度要求.

Based on machine vision technology with using VS2009 and OpenCV computer vision open source library as a development platform and language, industrial cameras and three-axis engraving machines are used as experiment platforms to develop vision correction functions for numerical control systems. First, the Mark points on the PCB board after clamping are collected. After the straight line detection is completed, the corner point detection is used to locate the Mark points. The world coordinate straight line equation is obtained by Hough transform, so as to obtain the current clamping position and attitude of the PCB board to be processed. Finally, the processing position information is corrected by the transformation relationship between image coordinates and machine coordinates. Aiming at the clamping error in the CNC system, a vision correction system is designed and developed for the CNC system based on machine vision technology. The actual verification and specific error data prove that the system meets the machining accuracy requirements of the CNC system.