Study On Shape Recognition And Detection System For Revolving Parts Based On Image Processing

Energy conservation and emissions reduction is one of the important contentsof our current country economy sustainable development, and Remanufacturingengineering is an important direction of energy saving and emission reduction inremanufacturing process.Due to the study object of the Remanufacturing processis the spare parts of the waste products. therefore, the old parts recognition anddetection become the key of the remanufacturing process.The recognition and detection of mechanical parts based on image processingis an very practical non-contact detection technology. It overcome themeasurement error caused by the contact in contrast to the traditional way ofartificial detection,which,to a large extent,improve the detection efficiency andaccuracy and realize the remanufacturing detection automation.The research idea of this paper is combining image processing techniques andpattern recognition to realize the recognition and detection of different parts.Themain research contents include:(1)The study focus on the helicoid part which accounts for larger percentagein the scrap parts, processing the collected images.Through contract imagesmoothing algorithm,image enhancement algorithm and edge detectionalgorithm,and finally uses gray level transformation, median filter, thresholdsegmentation and contour extraction, and obtained the relatively clear boundary.(2)This paper respectively from two aspects-based on the area and based oncontour-to describe the shape and feature of the object. Finally, it is characterizedby the area of the target area, perimeter, invariant moment and Fourier descriptor,and obtain their corresponding characteristic parameters and construct thecharacteristic vector.(3)The support vector machine (SVM) is applied to the identification process,in view of the component characteristics, design a suitable training algorithm andrecognition algorithm, finally completed the identification and classification ofspare parts.The image processing technology applied in remanufacturing for old partsdetection is the development trend of remanufacturing,.It overcome thedisadvantages of the traditional artificial recognition and detection,and therecognition accuracy can reach95%.