ERW Welding Condition Monitoring Technology Based On Digital Image

Recently,the use of ERW pipes has taken a more ative role in the energy developing fields. ERW pipesis now gradually replacing those seamless pipes that are currently being used for deeper wells,exreme highand low temperatures,high pressures and highly corrosive environment. China had a lot of ERWmanufacturing equipment in the world advanced level,but control technology gap is larger compare withother countries,which leaded to ERW pipes quality is not stable. Therefore, the design and development areal-time monitoring and control system of ERW welding states,which could improve our ERW pipequality. And it have the important theory significance and engineering application value.This paper mainly researched on digital image of heated area and taking the concept of ImageEngineering into the control algorithmic hierarchy,which were used to research ERW welding conditionmonitoring technology based on digital image,and design a computer real time system for monitoring andcontrol ERW welded. The main research contents are as follows:(1) an adaptive threshold segmentation method which named Advanced OTSU was proposed in thispaper.It had overcomed the effect of image noise which generated by light and moisture in fieldenvironment.The problem of the foreground and background separation is solved effectively by this method.Through experiments the method of the feasibility and reliability was verified.(2) Studying on the technology of Freeman chain code to find the edge of target and the method oforganize hierarchical relationship and distinguish holes and contours,which named suzuki,were used toovercome the problem of fake edges and extracted results unclosed by using gradient algorithm andpolygonal approximation. And used a square as threshold to solve the problem of extract foreground underenvironment exist interferences.(3) Progressive probabilistic Hough transform was researched in order to find linear information inthe image, and designed a screening algorithm for specified linear targets.Those methods had overcomedlinear information extraction in the complex conditions, and solved the problem of screening specified linein the dense linear clusters.(4) the expert system including heating area, welding speed, V-convergency point, V-convergence angle and angle bisector was established, through the learning and accumulation the expert system couldprocessing and feedback welding state in real-time, and control electrical equipment timely and effective.(5) The relation of image information and controlled variable is established by extracting intrinsicinformation monitoring image, which could monitor and estimate the weld quality.And structured acomputer monitoring system for ERW manufacture by using image processing, image analysis and imageunderstanding, It could enhence weld quality and improve the degree of automation and efficiency...