基于VBAI的机器人视觉焊缝坡口测量系统研制

发布时间：2018-02-25 19:34

本文关键词： 弧焊机器人焊缝识别结构光视觉传感器特征点提取算法　出处：《哈尔滨理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：在现代制造行业领域,尤其是重工业中,焊接技术的发展程度往往标志着整个行业的发达程度。在焊接技术中,机器人焊接技术如今已成为关键技术之一,而其向智能化方向发展已经成为一大趋势,怎样提高焊接机器人的智能化程度已经成为主要的研发课题。传统的焊接机器人大多属于示教再现型机器人,对复杂情况的应变能力差,受环境变化影响大,对工件的适配精度要求高。一旦焊接中出现装配问题或热应力导致的变形,焊接机器人很难复现理想的示范动作,这就会影响焊接质量。由于焊接机器人没有检测焊缝位置、偏移程度的能力,因此需要为其配置一个焊缝识别系统,帮助机器人在焊接过程中实时掌握焊缝位置以便做出应对,调节自身移动轨迹,实现智能焊接。本文参考了国内外主要的焊接机器人焊缝识别系统的设计方法,采用直射式激光三角法设计了一套基于视觉图像处理的焊缝识别系统。利用智能相机实时获取焊接图像,并利用VBAI进行图像处理,获取图像中焊缝特征点的位置,之后再根据传感器模块的几何关系,以及前一幅图像中特征点的位置,计算出两幅图像中焊缝分别在竖直方向和水平方向发生的位置变化,将其反馈到上位机进行处理,进而指导机器人移动。该方案精度较高,经过实验,该方案能够将误差控制在约0.5mm以内,而通常弧焊焊接对精度要求都在2mm内。并且,由于该方案利用智能相机获取图像,而智能相机本身带有数据处理功能,已经将上位机的职责分担了一部分,如果未来能够实现利用焊接机器人自身的软件系统实现对数据的接收,则有望彻底摆脱上位机,简化焊接机器人的构成,具有发展前景。本文同时对传感器模块的重要参数给出了标定方法,在实际焊接之前需按照方法对相机内外参数、几何设计参数进行标定,以保证计算焊缝位置偏移量时的准确性。
[Abstract]:In the field of modern manufacturing, especially in heavy industry, the development of welding technology often marks the development of the whole industry. In welding technology, robot welding technology has become one of the key technologies. How to improve the intelligent degree of welding robot has become the main research and development topic. Most of the traditional welding robots belong to the teaching and reproducing robot, and how to improve the intelligent degree of welding robot has become the main research and development topic. It is difficult for welding robot to reproduce ideal demonstration action once assembly problem or deformation caused by thermal stress occurs in welding, because of poor strain ability of complex cases, greatly affected by environment change, and high requirement of workpiece adaptation precision. This affects the quality of the welding. Because the robot has no ability to detect the position and deviation of the weld, it needs to be equipped with a weld identification system to help the robot master the position of the weld in real time during the welding process in order to respond. This paper refers to the design method of welding robot weld identification system at home and abroad. A welding seam recognition system based on visual image processing is designed by means of direct laser triangulation. The welding image is acquired by intelligent camera, and the position of weld feature points in the image is obtained by VBAI image processing. Then according to the geometric relationship of the sensor module and the position of the feature points in the previous image, the position changes of the weld seam in the vertical direction and the horizontal direction in the two images are calculated, which are fed back to the upper computer for processing. The scheme can control the error within 0.5 mm, but usually the precision of arc welding is less than 2 mm. Moreover, because the scheme uses intelligent camera to obtain images, the precision of the scheme is higher than that of the robot, and the precision of arc welding is less than 2 mm. The intelligent camera itself has data processing functions and has shared the responsibilities of the upper computer. If the software system of the welding robot itself can be used to receive the data in the future, it is expected to completely get rid of the upper computer. In this paper, the calibration method of the important parameters of the sensor module is given. Before the actual welding, the camera internal and external parameters and geometric design parameters should be calibrated according to the method. In order to ensure the accuracy of the calculation of weld position offset.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【相似文献】