Online Monitoring Of The Welding Quality Of ERW Based On Image Processing

ERW pipe is widely used in petroleum, chemical and other fluid transport. In recentyears, with the stat of the project of west east gas pipeline, China ERW pipe demandconstantly increasing, ERW pipe production capacity is increasing, but ERW pipe productionequipment and the degree of automation in China is still very low, welding quality most relyon working experience to judge, because the lack of online monitoring system of the weldingquality, the contradictions that production efficiency and welding quality inspection and ourERW pipe demand more and more sharp, research and development of such a set of ERWpipe weld quality online monitoring system is of great significance.This paper research object as ERW pipe production equipment in Bohai China oil pipeequipment Yangzhou branch as the research object, combined with the domestic andinternational study on ERW welding quality monitoring, put forward a technology of ERWpipe quality online monitoring system based on digital image processing. This paper studiesthe production of domestic ERW pipe, and ERW pipe weld quality detection of status, throughanalysis of welding in each link of the welding quality, get the influence factors of weldingquality. Because the influence factors of ERW pipe weld quality are in many aspects, thisarticle put forward a method for extraction and detection method for combining multiplevalue—image processing. Based on the basic theory of the region growing algorithm makesthe image of ERW welding for segmentation, the Hough transform algorithm was used toextract the line features of the image, the statistical method was used to extract the area size ofthe image, and the method of geometry was used to extract other geometric features of theimage, Extracted from the images of characters associated with the welding quality values are:welding heating area, type V junction position, V angle, V angle on both sides of thesymmetrical degree. Welding image is obtained by welding of high speed CCD color cameramounted in ERW welding. Finally, an artificial neural network and pattern recognitiontechniques to fuse these feature values, and then judge the welding quality of ERW pipe.The whole system is accomplished in Visual Studio2008software development platform, and image processing algorithms is using C++language, and compiled by MFC dialog based human computer interface, the interface is divided into four parts: the upper left of theinterface can display the real-time welding process, left part of the lower update displaywelding change curve of characteristic value from the image, the upper right corner is thecamera control settings, the lower right corner is the program on the welding qualityevaluation results of the quality forecast warning lamp. Through the debugging and test site,test result of the system of system of qualified rate in more than86%is right. To improve theERW pipe production efficiency and welding quality inspection automation is important.