| The engine high-pressure fuel pipe is an important component that connects the fuel injection pump and the fuel injector,and is responsible for delivering high-pressure fuel to the engine.However,during the production process,surface defects such as scratches,pits,and burrs may occur on tubing upsets due to force majeure factors such as machining tools,artificial environments,and so on.These defects can become the source of fatigue cracks,leading to tubing cracking and oil leakage.Once these happens,it will not only have a negative impact on the reliability of the entire engine,but also may lead to serious fire accidents.Therefore,this paper uses machine vision technology to study the surface defect detection and classification system for high-pressure oil pipes in automotive engines.This article aims to build a hardware platform for the characteristics of high-pressure oil pipes in automobiles,design the structure of the hardware platform,build an image acquisition and lighting system based on the actual size and shape characteristics of the oil pipes,complete the omni-directional image acquisition of the oil pipes,use machine vision algorithms for real-time image processing of oil pipe part drawings,and use NI LabVIEW software to achieve hardware manipulation and real-time display of processing results,Complete the detection and classification of defects such as scratches,pits,and burrs on the outer surface of the tubing.Using image graying,denoising,and other methods to preprocess the tubing image,and comparing the effects of various image segmentation methods,the maximum entropy method is proposed to segment the collected image.Finally,the image is processed by particle morphology using expansion and corrosion,thereby realizing edge detection and size measurement of the tubing pier.In this study,image analysis was conducted for different types of defects in automotive oil pipes,and different defect points were extracted.Methods such as gray level matching and linear gradient operator were used for defect detection and classification,with a comprehensive defect detection rate of 97%.Select 500 samples of long pier,short pier,and unprocessed pier in tubing,and conduct defect detection experiments using integrated software.Based on the comprehensive detection results,the average single detection time is less than 500 ms,with an accuracy rate of 97.79% to determine whether the product is qualified;conduct 100 measurements of the size of the upper and lower protrusions of the long and short pier heads of the tubing,collect data,and prepare an Xbar-R control chart.Analyze the control chart using statistical process control(SPC)theory,and analyze the process capability of the long and short pier heads.The process capability index of the lower protrusion of the long pier head is 1.56,and the process capability index of the upper protrusion of the short pier head is 1.39,indicating sufficient process capability;the process capability index for the upper protrusion of the long pier head is 1.89,and the process capability index for the lower protrusion of the short pier head is 2.07.The process capability is relatively high,and the detection requirements can be appropriately relaxed.Use NI LabVIEW software to design the five functional modules of the software part of the system,integrating image acquisition,defect detection,size detection,and other functions.Completed defect detection,dimensional measurement,data collection of automotive oil pipes,and communication with PLC to classify and process oil pipes,providing support and theoretical basis for defect detection and system research of automotive high-pressure oil pipes. |
