Research On Automobile Oil Seal Dimension Detection Method Based On Machine Vision

With the development of modern industry,the improvement of product quality is closely related to the improvement of inspection technology.For example,precision parts made of rubber,such as automobile oil seals,have many models and large demand,and large dimensional errors will cause transmission components to leak oil.The traditional manual size detection method is not only slow in detection speed,but also unavoidable such as reading,misdetection and missed detection.In order to solve the phenomenon that the inherent elasticity of the oil seal is easy to bring great uncertainty to the artificial dimensional detection accuracy,to meet the current efficient and rapid modern parts production and highprecision dimensional detection,this paper proposes a vehicle oil seal size based on machine vision inspection technology High-precision detection method,the detection method is efficient and the cost of equipment is low,and a reasonable detection scheme is designed.The main research contents of this article are as follows:First,after researching and designing the system detection platform and overall scheme,combining the technical indicators of the detection system,the errors were analyzed to prove the feasibility of the system.Complete the research of the imaging system and hardware selection,analyze the influence of light intensity on the edge of the oil seal image through the oil seal partial maps taken under different light intensities,determine the optimal light intensity of the system and complete the image acquisition.Bilateral filtering is used to remove image noise,and the advantages and disadvantages of several traditional edge detection operators are analyzed and compared.Based on the classic PCNN edge detection,the advantages of the Canny operator and the concept of bilateral filtering weighting coefficients are drawn,and an improved algorithm based on PCNN edge detection is proposed to obtain the true edges of automobile oil seals.Combined with the cubic spline interpolation method,a sub-pixel edge detection method based on improved PCNN is proposed to obtain the sub-pixel edge point position.Analyze and study the traditional Huff transform feature detection and random Huff transform circle detection methods.Based on this system,an optimized algorithm for random Huff transform circle detection is proposed.After a comparative analysis of the experimental results,the round edges detected after the algorithm is improved.Very close to the true contour of the oil seal,which facilitates high-precision size inspection.This paper combines the size detection algorithm of oil seals with camera calibration results,randomly selects 10 automobile oil seals for size detection experiments,and calculates the system size detection error based on the results of a high-precision threecoordinate measuring instrument.It is verified by experiments that the detection accuracy of the system is less than 0.03 mm and the average detection time of a single oil seal is 2.17 s,which meets the requirements of system technical indicators.Using MFC framework to design the human-computer interaction interface of the detection system through the VS2015 platform,so that the system has functions such as real-time display of size information.