Research On Key Technologies For Defects Detection Inside Small Pipes Based On Machine Vision

Small size and thin features of pipes or boresare used more and more widely inthe aerospace, automotive, energy and chemical industry.It is very important toachieve the defect detection and measurement of thin pore wall morphology formodern industrial manufacturing, quality and safety controlling. Among currentmethods, such as endoscopes, optical probes and micro-pipe robot, there arealwaysdeficiencies such as needed for manual intervention for sampling, high structuralcomplexity or low efficiency, which ishard to meet the actual requirements. In thispaper, a way to achieve a flexible micro-pore wall defects, automatic, rapid,high-precision measurement method is researched. With the use of special opticaltransmission components, light source importing and image exportingis completed.Combined with the principle of machine vision and image processingalgorithms,pores on the inner defect size and location of the high-precision detectionisachieved. Through modeling and experimental analysis,system performanceevaluation is completed. The main work of this paper is as following:1. Existing defect detection method inner pores surface is analysd.With theobjective needs of industrial production and quality control, a flexible, fast,comprehensive, high-precision measurement method based on machine vision isresearched and applied to thin pore defect detection.2. The method of attitude adjustment of optical transmission component isresearched. By constructing a mathematical model of image transmission andfeedback control based on image processing, the alignment between opticaltransmission component and the measured pores is completed to ensure accuracy andsafety of detection system.Experimentsare done to verified the feasibility of attitudeadjustment.3. Adaptive illuminator systems based on imageevaluation is researched. MCU isused on high-power white LED light source for embedded control to ensure theoriginal image clarity and stabilityunder different environment. Make it easy toautomate measuring and extending the scope of application of the system.4. Digital image processing technology on defect detection is researched.Throughthe study on the different images at different stages, image algorithms are designed andoptimized.The errors theory of image processing is analysised and theexperiments are completed forevaluating system performance.5.The error factors of defect detection system are analyzed from theoreticalanalysis and experimental verification on multiple aspects.Error combination andevaluation are completed for the system performance.