Research On Bearing And Synchronous Wheel Dimension Detection Technology Based On Machine Vision

In the traditional machinery manufacturing process,the quality inspection of processed parts usually uses manual inspection,which not only affects the inspection speed and production efficiency,but also may affect the accuracy of the inspection results due to the interference of sampling and subjective factors of the inspector.With the development of automated and intelligent technology,machine vision detection technology that can overcome the shortcomings of traditional manual detection and achieve fast and accurate detection is gradually applied to industrial production.The machine vision-based detection technology starts from the image that contains the size information of the part,and uses the computer's outstanding computing power to process and analyze the image accordingly,and obtain the size result of the part.Its outstanding advantages are fast measurement speed,non-contact measurement and high degree of automation.Based on the in-depth analysis of machine vision inspection technology,this paper researches and designs a detection algorithm for bearing parts and timing pulley parts based on machine vision,and tests and analyzes them.The results show that the algorithm can meet the requirements of detection.Precision requirements for the manufacture of conventional bearings and timing pulleys.The main work and innovations of this article are as follows:(1)Build a hardware platform for the detection system.Based on a detailed analysis of the hardware components of the machine vision inspection system,a reasonable selection of various parts including industrial cameras,industrial lenses,lighting systems,and computers was made to build a machine vision-based inspection device Provides a hardware platform to ensure the accuracy of subsequent experimental measurements.(2)Software system platform design.Based on Python language and OpenCV open source vision library as the main platform,Qt is used to design the image processing algorithm and the result display method.The simplicity of the Python language and the quickness of Qt Designer make development easier and more efficient.(3)The image preprocessing technique and edge feature point extraction techniques are studied in detail.On the basis of in-depth analysis of image preprocessing techniques,the processing effects of different algorithms are compared and analyzed,including image enhancement,image segmentation,filter noise reduction,edge detection and sub-pixel edge localization.On the basis of comparative analysis,the algorithm processing flow including fixed threshold segmentation,median filtering and Canny operator edge detection is selected.A fast traversal extraction method is proposed for edge feature point extraction.The results show that the method can meet the accuracy requirements.(4)Detection system calibration and error analysis.On the basis of Harris corner detection,the camera is calibrated by the pixel equivalent calibration method,which is more convenient and faster than the traditional calibration method of camera internal and external parameters.The possible sources of error in the detection system are further analyzed,including the error of image acquisition equipment,noise influence,installation error,illumination influence,algorithm error,etc.,and some solutions to reduce the error as much as possible are proposed.(5)For the standard bearing with inner diameter of 25 mm and outer diameter of 52 mm,the 25-tooth and XL-type synchronous pulleys were tested.The inner and outer diameters of the bearing and the concentricity were measured.The actual outer diameter of the synchronous pulley showed that Both can meet the accuracy requirements of use.And the further extracted edge extraction method can also be applied to the same type of parts detection,such as the outer diameter of the gear and the like,and the synchronous pulley has similar characteristics.The main innovations of this article are as follows:(1)Combining programming with Python language,OpenCV library,and Qt,a software platform for the inspection system is designed,and algorithms for part inspection are written.(2)A fast traversal method for extracting edge feature points was proposed and applied to the detection of bearings and timing pulleys,which verified the feasibility of the method.(3)Based on the designed detection system,a method of classifying parts based on tolerances is proposed.The tolerance range is divided into different sections.When the parts are used together,different degrees of cooperation are achieved to meet different requirements.And the part size data operation is designed.