Design And Implementation Of Visual Inspection System For Small Diameter Seamless Steel Pipe Defects In Boilers

Quality control is a core competitiveness of enterprises,especially in the boiler industry.Living in coal-fired boilers winter,in the face of profit decline,excess capacity,the enterprises will control the quality more and stricter.However,the traditional method of manual detection is inefficient,has the high rate of false positives,and the types of defects and the descriptions are extremely limited.Machine vision technology is a kind of detection method which has become more and more mature in recent years,with the development of computer and automation,which has become the focus and trend of current enterprise research.In this paper,small-diameter seamless tupe in the boiler is the research object.Focus on the detection of surface defects in the machine vision.According to the site environment and business requirements,determine overall design of the system of the seamless tupe surface defect.And analyze the variety of image segmentation,enhancement and other pretreatment methods in detail,using an improved efficient median filtering method for noise smoothing.According to the characteristics of different texture caused by different surface of tupe,the parameters influencing the texture characteristics are studied.The parameters characteristic of none-detects tupe sample is used as the template,and the characteristic database is established.Through comparative analysis to detect defects,meet the industrial automation testing requirements.