基于图像处理的模糊控制埋弧焊缝跟踪研究

焊接工艺由于其本身加工过程及环境的复杂性,使得焊缝难以精确跟踪.针对埋弧焊系统,采用视觉传感器进行焊缝跟踪,得到误差信号后,使用分段拼接控制的方法,使误差快速收敛,且减小超调.首先对焊缝图像进行预处理,使用自动化的阈值进行分割得到焊缝区域,应用边缘检测和Hough变换得到焊缝边缘,计算得到焊缝中心位置、焊矩和焊缝偏差;然后采用模糊增量式分段PI控制处理误差,当误差较大时采用增量式PI控制消除误差,当误差较小时应用模糊控制,设计了隶属度表以及模糊推理机制,采用小误差的模糊控制有效地抑制了超调,使误差快速收敛.最终仿真结果显示,使用PI混合模糊控制后,超调量被控制在0.2个单位以内,调节时间小于1 s,基本实现了对焊缝的快速准确跟踪.

Research on Fuzzy Incremental Control for Submerged ARC Weld with Image Processing

Because of the complexity of the welding crafts and the environment,the weld is hard to be accurately tracked.Submerged arc welding system that uses visual seam tracking sensor,after the error signal is obtained,sub-splicing control is used to make the error convergence fast and the overshoot reducing.Firstly,the weld seam image is preprocessed,after median filter for removing noise,automatic threshold segmentation is used to get the weld area.Then,edge detection and Hough transform is used to get weld edge.As a result,the central position of welding seam,the welding torch and seam deviation are calculated and obtained.Then the fuzzy segmented incremental PI control is used to handle the error.When the error is large,incremental PI control is used to eliminate the error,when error is small,the fuzzy control is used to design fuzzy reasoning mechanism.The fuzzy control effectively inhibits the overshoot and error convergence is completed quickly.Finally,simulation results show that,after using mixed PI fuzzy control,overshoot is controlled within 0.2 units,adjustment time is less than 1s,a fast and accurate tracking of the weld is achieved basically.