Study On Visual Detection Methods Of The Internal Hole's Coating Defects Of Screw

As a part of the foundation of manufacturing industry,the quality of screw will directly affect the future development of the industry.The excessive and lack of coating on the inner hole of the screw will affect the quality of the screw and the performance of participating in the industrial production process.At present,the commonly used detection method of screw parts is still manual detection,which is inefficient and highly subjective,and is not suitable for automatic production,and it is difficult to reach the enterprise precision standard.At present,there are few cases of using visual inspection method to detect the defects of screw inner hole coating at home and abroad.Therefore,as a new detection method,the visual detection method is urgently needed to be applied in the detection of this kind of workpiece.The technical requirement of the coating defects of screw inner hole required by the enterprises cooperating with the laboratory is that the diameter of the coating holes should be greater than 3mm.The machine vision detection technology can detect the defects of this accuracy.This research method is to conduct tests and experiments around this index.The detection device includes a translation platform A and a translation platform B,making the inner hole of the screw move to the position of the lens and collecting images on the bottom surface,side surface and top surface of the inner hole of the screw respectively.The lens,camera and light source were selected according to the testing conditions,and the establishment of the coating defect detection device for the inner hole of the screw was realized.In the software design,mainly includes the system software function realization and the flaw detection algorithm design.Among them,the defect detection algorithm recognizes and classifies defects according to their imaging features.Image processing algorithm is the key to the detection,the use of image preprocessing of image enhancement in the gray level transformation characteristics needed for the surface of the image is more obvious,the Canny edge detection algorithm and threshold segmentation of the global fixed threshold segmentation determines the underside boundaries,the boundary of the image at the center of the circle and underside global fixed threshold segmentation and global dynamic threshold method to segment the between-cluster variance of screw image of the hue(H)of the channels,and finally through opening function of mathematical morphology and closed operation accurate defect area.The detection algorithm was used to traverse the map library for experimental test,and the defect area with hole diameter over 3mm could be detected,with a detection rate of 5.4% and high accuracy.