Research On Image Processing Technology Based On C++ In Weld Tracking System

Welding technology is a permanent material connection method.Welding technology plays a vital role in metallurgy and materials science.With the rapid development of electronic technology and computer network technology,and the extensive application of fuzzy mathematics and neural network in welding field,the seam tracking technology based on visual sensing has become a hot topic in the current research.At present,the application of laser vision automation welding equipment in the actual production is more and more mature.However,laser vision tracking technology has limitations.It does not apply to the weld seam which is without groove.In order to solve this problem,this paper has developed a set of weld seam tracking system for closely connected curve weld.The system consists of four parts:image acquisition unit,information processing unit,walking mechanism and correction mechanism.In the welding process,the image acquisition unit move forward along the Welding direction Driven by the walking mechanism.At the same time,images acquired by the CCD camera will sent to the IPC for data processing.And then send the appropriate control instructions to the correction mechanism according to the data processing results.The torch is always aligned with the weld center under the control of corrective mechanism,so that the goal of tracking is achieved.In the aspect of image acquisition,this paper adopts mechanical protective cover and light filter to filter out the interference of arc and splash.In the aspect of image processing,this paper adopts a series of image processing algorithms to extract the weld center point and calculate position deviation between welding torch and weld seam.The image algorithm includes local adaptive binarization,morphological filtering(open operation),regional connectivity filtering,gravity method extracts the weld centerline,least squares fitting curve and so on.In the aspect of tracking control,this paper adopts industrial computer and PLC to realize the real-time tracking during the welding process.And the tracking accuracy can reach ± 0.2mm.This paper has completed the debugging of weld tracking system and experimental verification work.The experimental results show that the system meets the design requirements and process requirements,and the tracking of the close butt welds is successfully achieved.This system makes up the defects of weld tracking system based on the laser vision.