Research On Vehicle Light Bulb Detection System Based On Machine Vision

In order to ensure the product quality of the automotive lamp,improve safe coefficient and enhance user experiences,General Administration of Quality Supervision,Inspection and Quarantine and Standardization Administration of China issued the mandatory national standards(GB 15766.1-2008)about the dimension of the automotive lamp and photoelectric property.However,the existing measurement methods in the factory are basically contact measurement which the process is complex,not only consume too much time but also can cause some damage to the lamp itself.In relative terms,machine vision measurement have the features of non-contact,better flexibility,high accuracy,high efficiency and intelligent that it is a good way to satisfy the detection requirements for the automotive lamp.Therefore,this paper aims to use machine vision technology to detect the automotive lamp.For the sake of designing the detection system based on machine vision reasonably,this paper describes the hardware design and algorithm implementation of a typical machine vision inspection system with practical experience in detail.For the part of the hardware,it mainly introduces how to select the important components such as light source,lens,camera and so on.For the part of the algorithm,it mainly discusses some basic image processing algorithms,including image filtering,image segmentation,feature extraction and so on,and several popular machine vision software tools on the market are introduced properly.In this paper,the detection system of P21/5w automotive lamp and H3 automotive lamp produced by Baoding Life Automotive Lighting Group Company Limited are explored and designed,which introduces in detail how to choose suitable hardware and correct algorithms effectively according to the measured object.The detection system of P21/5w automotive lamp is designed and implemented to detect all of the size content that stated by the GB15766.1-2008.The system captures the images from different directions to ensure the convenience and accuracy of detection,which starts with the specific requirements of the project.For the image processing,the system processes the images captured in the two viewpoints independently,compares and analyzes various algorithms,adopts a relatively suitable algorithm,and obtains good results finally.In order to detect the diameter and peak of lamp housing,the curvature restrict estimation based on ellipse fitting(CREF)is proposed in this paper.The algorithm can be divided into two steps,firstly,ellipse fitting is utilized to extract the edge of lamp housing,secondly,curvature restrict estimation is utilized to get the required data points,and the effects of inspecting on the field are well.The detection system of H3 automotive lamp,mainly for the practical needs of the factory,through the measurement of the lamp foot length to complete the online real-time detection and judgment,uses PLC and external servo mechanism to classify and reject the unqualified lamp.The system makes full use of the relationship between the lamp pin and the contact surface to lock and judge,which has high precision and runs well.