摘要
针对传统粒子群算法的机器人焊接路径精度低、速度慢等问题,提出一种改进粒子群的焊缝轨迹高效识别及精确路径规划算法。首先,通过CCD相机拍照,采集焊接工件,并将图像传递至上位机,上位机对采集的图像进行灰度化、提取目标区域、阈值分割、腐蚀膨胀、去干扰等图像预处理操作;其次,利用Canny算子对预处理后的图像提取出焊接件外轮廓边缘像素坐标信息,并根据焊缝拟合曲线函数残差模的最小值拟合出焊缝的曲线函数;最后,提出一种改进粒子群的指数函数惯性权重动态更新策略,其惯性权重随着迭代次数的增加而减小,以实现前期快速全局搜索,后期局部寻优的目的,实现焊缝精确路径规划用于弧焊机器人的快速、高效焊接。实验结果表明,改进粒子群的机器人焊接系统可对复杂路径的焊接件进行路径高效识别与准确规划,可降低机器人焊接路径点的复杂度,焊接误差值不超过0.30 mm,相较传统粒子群的焊接误差超过0.50 mm,机器人焊接精度提高了20%。
Improved particle swarm optimization algorithm is proposed for efficient recognition and precise path planning of weld trajectory,in response to the problems of low accuracy and slow speed of robot welding paths in traditional particle swarm optimization algorithms.Firstly,a CCD camera is used to take photos of the welding workpiece,and the image is transmitted to the upper computer.The upper computer performs image preprocessing operations such as graying,extracting the target area,threshold segmentation,corrosion expansion,and de interference on the collected image;Secondly,the Canny operator is used to extract the pixel coordinate information of the outer contour edge of the welded joint from the preprocessed image,and the curve function of the weld joint is fitted based on the minimum residual modulus of the weld joint fitting curve function;Finally,an improved exponential function inertia weight dynamic update strategy based on particle swarm optimization is proposed,which reduces the inertia weight as the number of iterations increases to achieve rapid global search in the early stage and local optimization in the later stage,and to achieve precise path planning for welding seams for fast and efficient welding of arc welding robots.The experimental results show that the improved particle swarm optimization robot welding system can efficiently identify and accurately plan the path of complex welded parts,reduce the complexity of robot welding path points,and the welding error value does not exceed 0.30mm.Compared with traditional particle swarm optimization,the welding error exceeds 0.50mm,and the robot welding accuracy is improved by 20%.
作者
罗辉
崔亚飞
LUO Hui;CUI Yafei(Key Laboratory of Intelligent Manufacturing at Yongzhou Vocational and Technical College,Yongzhou,Hunan 425100,China)
出处
《自动化与仪器仪表》
2023年第10期191-195,共5页
Automation & Instrumentation
