Research On Visual Inspection System For Weld Feature Size Based On Line Structured Light

With the rapid development of the machining and manufacturing industry,welding has been widely used in the forming connection of structural parts.At present,manual methods are often used for welding and quality inspection,with low efficiency and unstable product quality,which is difficult to meet the needs of large-scale production of workpieces.With the advantages of high precision,high efficiency and high stability,the 3D visual inspection technology based on line structured light(LSL)is gradually being applied in the field of industrial automation welding.Based on the requirements of visual inspection of weld feature size and taking multi-type groove welds as research objects,this paper designed and built a visual inspection system for weld feature size based on LSL.The main research results are as follows:Faced with the requirements of visual inspection of the weld geometric size and combined with the system design technical indicators,the overall design scheme of the visual inspection system based on LSL was formulated.What's more,the system hardware modules,software architecture and image processing flow were designed.After analyzing the characteristics of line structured light stripe and common center extraction methods,an adaptive light stripe sub-pixel center extraction algorithm is proposed,which effectively solves the problems of changing width and uneven gray distribution.In addition,the multi-dimensional accuracy evaluation indicators were introduced,and the verification experiments of the proposed algorithm's validation and robustness were completed.In order to extract the groove light stripe of the stripe image,a background filtering algorithm based on the region of interest division and differentiated image enhancement was designed.After the light stripe image was divided into different regions,differentiated convolution kernels were constructed for enhancement processing.The open operation and false target removal methods were used to effectively remove the background noise of the image and accurately extract the complete target stripe.Aiming at extracting the feature points of various types of groove and weld light stripes,a fusion algorithm of type recognition and corner positioning was proposed,the groove corner points were located.According to this,the center line of the light stripe was extracted in different segments,and the corresponding groove feature points were accurately extracted by the straight line intersection method and the minimum distance method.a feature points extraction algorithm based on two-parameter threshold was used to realize the feature points extraction of many different welds.The software of the visual inspection system in this paper was realized,including system parameter configuration,system calibration,image processing and other modules.System inspection performance test experiments were carried out,error correction and error analysis were also completed.The experiment results show that: the system's detection error of groove's geometric size is 0.104 mm,the detection error of weld's size is 0.073 mm,the single measurement takes up to 109 ms,and the inspection stability and repeatability of the system is 1%,meeting the online detection requirements of practical applications.