Research On Weld Seam Inspection Technology For Large Structural Parts Based On Line Structural Light Vision

In the field of large industrial equipment manufacturing,welding as a core material processing technology is widely used,the quality of welding directly affects the safety performance of the components,so the quality of the weld inspection of structural components is indispensable.At present,due to the large number of large structural members weld seams,complex spatial location,the direction of the spatial curve type,so the quality of post-weld inspection is mainly visual and key position weld sampling way to carry out.This inspection method is inefficient,labour-intensive,non-traceable and has a low degree of automation,and is no longer able to meet the global monitoring of the quality of welds on large structural components.To address these issues,this paper builds an automated inspection platform for boom weld seams based on line structured light vision sensors,extracts the weld seam streak features collected by the line structured light sensors and investigates the process of automated weld seam inspection in the context of engineering applications.The project begins with a technical analysis of the inspection difficulties of boom welds of large structural components and the design of a hardware system for boom weld quality inspection with a structured light vision sensor as the core.The hardware system consists of a robot,a position changer,an industrial control machine and a linear structured light sensor.The weld seam measurement model of the structured light vision sensor is established,and the calibration is completed;the characteristics of the weld seam streak images collected by the structured light vision sensor during the weld seam inspection process are analyzed,based on which different image processing methods are studied and an image pre-processing process is established to accurately segment the weld seam streak images.Then,based on the study of various stripe centerline extraction algorithms,the stripe centerline extraction algorithm based on width variation is proposed.The extraction of weld seam feature points is a key step to complete the measurement of weld seam geometry.The effect of the improved slope analysis method of weld seam feature point extraction is studied and the effect of different thresholds on feature point extraction is analyzed;a feature point extraction algorithm based on the straight line feature of the weld seam contour is proposed,which can calculate the weld seam feature points more accurately and improve the accuracy of weld seam detection.Finally,in order to realize automatic detection of the boom weld seam,an automatic detection software system is developed,which is capable of achieving accurate evaluation of the weld seam residual height,melt width and weld foot size.On the basis of the basic weld seam dimension measurement,an automatic inspection process was established,and the nodes of the inspection process were developed so that automatic inspection of a single weld seam could be achieved once the software system communicated with the hardware equipment.Using a standard block to simulate butt weld size measurement,the software system has a stable test accuracy of 0.012 mm for the residual height and 0.068 mm for the melt width,which shows that the stability and accuracy of the software system meets the requirements of engineering applications.On the basis of the inspection of a single weld seam,the automatic inspection process of the global weld seam of the boom is planned,and the hardware and software realize the joint movement,which initially achieves the goal of automatic inspection.The national standard for weld seam appearance inspection is studied,a quality assessment model for the boom weld seam is established,weld seam inspection data is extracted from the software system,and the assessment of the surface quality of the boom weld seam can be achieved according to the assessment standard.