基于视觉识别的多层多道路径规划修正  

Correction of Multi-Layer Path Planning Based on Vision Recognition

作　　者：杨乘东[1] 钟继勇[1] 陈玉喜[1] 陈善本[1] 

机构地区：[1]上海交通大学材料科学与工程学院,上海200240

出　　处：《上海交通大学学报》2015年第3期297-300,共4页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金项目(51105252);上海市科委科技计划项目(11111100302)资助

摘　　要：针对传统大厚板手工电弧焊和半自动焊工艺复杂、焊接效率低下、质量稳定性差等问题提出大厚板焊接的新方法:双机器人双面熔化极活性气体保护焊(MAG),打底焊采用双机器人双面MAG错位焊接,填充焊采用双机器人双面MAG对称焊接.多层多道焊接过程中累积误差和焊接变形会对多层多道路径规划精度造成影响,准确的多层多道路径规划是实现大厚板机器人自动焊接的关键因素.提出基于视觉识别的多层多道路径规划修正方法,通过图像处理结合图像三维信息恢复获得各焊道实时填充信息,对多层多道路径规划进行修正,提高多层多道路径规划精度.A novel technology of double sided arc welding （DSAW） in which the gas metal arc welding （GMAW） was employed for root welding and filler passes respectively was proposed in this paper. The correction of multi-layer path planning based on vision recognition （CMLPPBVR） was proposed to de- crease the accumulated error and increase the precision of path planning. Laser stripe was used to scan the filler condition of the groove after each welding bead was welded. The real time filler information of each welding bead can be obtained through the image processing and 3D information restoration to correct the multi-layer path planning, which improves the precision of multi-layer path planning.

关 键 词：双面双弧焊 路径规划 视觉识别 三维信息恢复 

分 类 号：TG409[金属学及工艺—焊接]