Research On Real-time Detection Of Automotive Tubing Surface Defects Based On Ni Vision

At present,in the automobile manufacturing industry,the main types of automobile tubing production are the surface galvanized PVF series and surface galvanized PA12 series of double layer pipe,they are widely used in the automobile transmission system.Although the production and processing technology of this type of tubing has been relatively mature,in the actual processing process,due to the impact of processing equipment and production environment,surface defects exist in part of the tubing,and the defects often pose a serious threat to personal and property safety.Considering the above problems,based on the analysis of defects in the process of automobile tubing in the actual production and processing characteristics and regularity,combined the actual factory assembly line production characteristics,puts forward the design scheme of the system,and made the following research:Firstly,according to the characteristics and the rules of the defects on the surface of the automobile tubing,the paper systematically analyzes and summarizes the application of machine vision technology in detecting the surface defects of the automobile tubing.Secondly,aiming at the characteristics of fast moving in tubing in assembly line and the need for real-time detection of tubing this technical requirements,with four array industrial camera in the same shaft surface installation way,effectively solved the problem of the detection system for image acquisition.At the same time,for the online detection of tubing surface defects of different pipe diameters,the technology research before image processing is carried out.The edge extraction,center point extraction and other algorithms are used to identify the tubing part of the image.According to the extracted center point,the ROI division,affine transformation,image segmentation and other algorithms are used to extract the tubing sample image,effectively improving the efficiency of image processing and analysis.Through dynamic test and comparison,image processing algorithms such as enhancement,filtering and wavelet transform of tubing surface defect images were screened,so as to effectively improve image quality and enhance detection reliability.Finally,through the analysis of the characteristics of tubing surface defects,combined with the actual situation of tubing defects and movement characteristics,the combination of SVM and template matching algorithm is selected to identify the defects on the actual pipeline.The hybrid detection and recognition method can effectively improve the detection efficiency and improve the anti-interference performance of the system.Finally,in the PC end,the software development of tubing surface defect detection system is carried out by using LabVIEW platform.The image data acquisition module,image processing module,image recognition module,serial communication module and other functions required by the detection system are realized,which can effectively output recognition results and carry out defect report and marking.The system related to practical methods and techniques,compiled program can reach 25 FPS detection rate and the accuracy of 90%.The system can help the inspection personnel quickly detect whether there are defects on the surface of the tubing,and improve the efficiency of the factory line.