Research On Visual Inspection System For Copper Bar Cable Defects In Heat Hole Of Large Motor

With the rapid development of power industry in recent years,people gradually begin to pay close attention to the security and dependability of industrial motor.Due to the defects of the processing technology,the quality of the motor rotor copper row cable product is not up to standard.The wall of the heat dissipation hole of the motor copper row cable Burr is a major hidden danger that affects safe production.Manual visual inspection can not meet the requirements of current production scale and accurate inspection because of its high cost,lax quality control,heavy workload and other reasons.Because the visual inspection has the characteristics of non-contact,online,high speed and high precision,the non-contact machine vision inspection technology is introduced into the field of industrial product defect detection.The main research work of this paper is the detection technology of the surface burr of the heat dissipation hole wall of the large-scale motor copper row cable.In order to verb the test correctness and efficiency of the burr,the test rate of the detection system is not high due to the random shape and different size of the burr generated during the processing of the heat dissipation hole of the copper cable,and the incomplete shooting of the heat dissipation hole wall by a single camera.,This article analyzes and examine the actual state and technical requirements of copper row cable workpieces on industrial assembly lines,and designs a set of copper row cable heat dissipation hole superficies burr defect detection system based on binocular vision.The study target of this article is the burr flaw on the surface of the heat dissipation hole of the copper bar of the motor.First,the camera standardization of the binocular vision inspection system is carried out,and the visual image is obtained and corrected.Perform preprocessing such as denoising and segmentation for the corrected image.Secondly,image registration is performed based on the improvement of the SURF（Speeded Up Robust Features）algorithm,and the weighted fusion algorithm is used to complete the image fusion.Finally,on the basis of analyzing the burr feature attributes and comparing the commonly used defect detection algorithms,a burr feature extraction algorithm combining Otsu threshold segmentation with mathematical morphology and mask optimization algorithms is proposed to construct the detection ring zone area.And perform image calculation on it to complete the extraction of the burr feature on the wall of the heat dissipation hole,then traverse the area,and complete the detection and judgment of the burr by setting the threshold.The experimental results indicate that contrasted with the commonly used defect detection algorithms,the algorithm in this paper is more accurate and has a shorter detection cycle.It has better detection results for the burrs generated during the processing of copper cable heat dissipation holes.The detection rate is 98.2%,which can meet the requirements of bristle detection of the heat dissipation holes of copper row cable.