Research On Machine Vision System For Cable Surface Defect Inspection

Cables can be seen everywhere in production and life,including wire and cable for power transmission,optical cables for optical communication,etc.The inner layer transmission medium is mostly wrapped by an outer protective cover,which will affect the quality of the cable when the outer protective cover is damaged.In recent years,with the development of computer image processing technology and the continuous breakthrough of machine vision detection technology,more and more enterprises and scientific research units have used machine vision technology in the detection of product appearance defects,which is used to replace traditional manual detection methods.This thesis studies the machine vision system for cable surface defect detection,which satisfies the demand for online real-time detection of cable surface defects produced by enterprises.The main research content includes four parts.First,the overall design of the cable surface defect detection system.According to the user's needs,the system's functions and performance indicators were designed.The overall structure of the system was designed.The main components of the system were compared and analyzed.The design ideas of this thesis were put forward.Next,a cable surface image acquisition unit is designed.The design structure of the three-camera 360°surround cable cylindrical surface is adopted to realize the panoramic image acquisition of the cable surface.At the same time,the camera,the light source,the counting and synchronization module and its connection scheme of the system are designed.The experimental results show that the image acquisition unit can collect clear cable surface images,and the image background interference is reduced by the improved scheme for the image background,which reduces the design difficulty of the defect detection algorithm.Then,the cable surface defect detection algorithm was researched and implemented.After comparing different design schemes,the detection process suitable for the system is designed according to the actual requirements of this thesis.The main steps are divided into cable surface image preprocessing,cable area extraction,cable surface defect marking and cable surface defect feature extraction and recognition.A comparative analysis of the splicing and defect connected domain marking methods of the three cable diagrams was carried out,and the training support vector machine was used to identify the pinholes and dust on the cable surface.Finally,test the system.After the image acquisition unit function,algorithm code implementation and interface design are completed,the system function and performance index,maximum acquisition speed,minimum detection defect size and detection diameter accuracy and range are tested respectively,and the function and performance index of the system are obtained.After testing,the maximum detection speed of the system reached 124.8 m/min,the minimum defect detection area was0.1mm~2,the detection diameter ranged from 1 to 33.3mm,the detection rate of the system for cable surface defects reached 100%,and the recognition rate of surface pinhole reached 70%.The overall performance is close to the Taymer SI3100,meeting the actual needs of industrial production.