Quality Detection Of Automobile Windshield Surface Based On Spectral Confocal Sensor

Automobile windshield is an essential part in the production of automobile.In the process of glass surface quality inspection,it needs to rely on the way of manual modeling for sampling inspection.In the process of inspection,it is easy to be affected by manual subjectivity,so it is difficult to ensure the accuracy of inspection,and the manual inspection takes a long time,which affects the production efficiency.In this paper,aiming at the actual engineering situation,the spectral confocal sensor shape quality detection system for automobile windshield is studied,and the online real-time measurement and inspection of different types and sizes of windshield are realized.According to the characteristics of the windshield surface and the expected technical indicators,the overall architecture of the system is determined,and the selection of key hardware such as spectral confocal sensor,industrial camera and light source is completed according to the actual situation.An image registration algorithm based on windshield edge contour is proposed.The image denoising algorithms are compared,and the median filter is used for denoising the windshield edge image to reduce the influence of noise on edge extraction.Then Canny operator is used to extract the edge contour of windshield.After obtaining the point cloud data of the glass edge contour,ICP registration is carried out to obtain the rotation displacement matrix from the source point cloud to the target point cloud.This paper studies the quality inspection method of automobile windshield surface.Combined with the ratio of actual distance and pixel distance obtained by camera calibration and the result of hand eye calibration,the positioning error of windshield in the world coordinate system is calculated.After that,the error compensation is carried out for the calculated positioning error,the measurement trajectory of the sensor is replanned,and the data measurement and quality evaluation are carried out.The experiment and result analysis of the system are completed.The accuracy of ICP registration algorithm is verified by experiments,and the error is analyzed.After that,the influence of the change of the position of the measuring point of the glass surface on the sensor data measurement is analyzed,which shows that the ICP registration has high accuracy and can be used to compensate the positioning error of the windshield.Then,it is verified by experiments that the sensor can accurately measure the points to be measured on the surface after each replanning of the measurement trajectory of the sensor.