Research On Assembly Testing Technology Of Bearing Sealing Cover Based On Machine Vision

Bearing plays a pivotal role in mechanical manufacturing and it is an essential part for all large-scale mechanical structures.As a part of the bearing,the seal is particularly important for a qualified bearing.The quality of seal assembly directly determines the quality of bearing.Damage to the seal will lead to the impurity dust in the air entering the interior of the bearing,thus exacerbating the internal wear and shortening the life of the bearing.Similarly,for the mechanical system,minor damage to the seals affects the mechanical performance and efficiency,and serious damage can result in the scrap of the mechanical equipment.Therefore,the study of bearing seal assembly detection system is of great significance to the production and manufacturing.Aiming at the problem of sealing cover assembly detection in the existing automatic bearing assembly production process,the machine vision technology was applied to seal cover assembly detection in this thesis.The three possible nonconformities in the sealing cover assembly process were analyzed independently and three detection algorithms were studied according to their different characteristics,and the sealing cover assembly detection system was built to realize the detection of all kinds of unqualified seals assembly types.Meanwhile,each detection result is saved.When the detection record is inquired,the data of each type of detection result is summarized and drawn into an intuitive pie chart.Firstly,taking into account the application conditions of machine vision and the real-time acquisition of the visual camera when working,the possible noise types of the collected bearing images should be analyzed according to their situation,and the image denoising of different types that might be exist should be given the contrasts with Matlab,so as to select the denoising method suitable for this thesis.Secondly,appropriate processing methods are used to complete the pre-processing of the collected images in Halcon.Then,the detection algorithm based on the number of connected domains and the difference between the minimum peripheral circle and the ring width is studied.The improved Threshold segmentation of Otsu is adopted to extract the target area and improve the efficiency of the algorithm.Finally,the sealing cover assembly detection algorithm research is completed.Furthermore,the sealing cover assembly detection algorithm program is exported as C# program and compiled with Visual Studio2019.Finally,C# programming language and SQL Server database were combined to develop the bondless detection system to realize the visualization of detection results and real-time detection records.The experiment results show that the system is stable and reliable and meets the detection requirements.