Design Of Rubber Seal Detection System Based On Machine Vision

Rubber as raw material sealing ring is most commonly used in all kinds of mechanical equipment on the oil,water,gas and other fluid seal,because of its good sealing performance,strong structure,low cost and other advantages,widely used in industrial hydraulic and pneumatic transmission system.However,the leakage caused by surface defects,edge notches and size discrepancies caused by manufacturing processes not only seriously affects the function and safety of the system itself,but also leads to the deterioration of the performance of the power system and the increase of mechanical wear,and even threatens the safety of human life and property.At present,the quality detection of rubber sealing rings mainly relies on artificial visual resolution defects,which is difficult to adapt to the requirements of high-precision and high speed detection in large quantities.This paper based on machine vision technology to expand the rubber seal quality detection system,including the following aspects.1.Because the rubber seal ring is easy to deform and the edge is uneven,the traditional visual detection technology is easy to be interfered in the edge notch detection and size measurement,thus affecting the detection accuracy.Aiming at the crux of the problem,this paper firstly defines the edge position by using the local edge trend fitting method,and then resets the boundary between interference and defect,which solves the interference encountered in detecting the edge notch of the sealing ring.Secondly,based on the least square fitting and the three-point circle fixing algorithm,a specific sealing ring size measurement method is proposed,so that the inner and outer diameters of the measuring ring are more in line with the actual size value.2.Aiming at the problem that the existing visual technology is difficult to detect the subtle defects on the surface of the sealing ring,this paper proposes a method for the detection of the sealing ring defects based on six photometric stereoscopic method and image comprehensive feature analysis.Firstly,six images with different light source angles were collected,and the surface gradient and reflectance maps were reconstructed by photometric stereoscopic method.The surface gradient maps were converted into average curvature and Gaussian curvature images.Then the two curvature maps were converted into gray maps,and the gray maps were separated by fixed threshold method and local mean and variance threshold segmentation method.Finally,the connected domain feature analysis method is used to determine whether the product has defects.3.Based on the above research,according to the actual needs of the industry for the seal ring automatic detection system,the seal ring quality detection system is developed using Proxy mode.Firstly,according to the characteristics of the sealing ring,the appropriate light source,camera,controller and other hardware is selected to build the corresponding detection platform,and the mechanical position and motor parameters are run and adjusted.Then the relevant detection algorithm tools are developed and the software interface is completed by Visual Studio graphics library Win Form.Finally formed a set of sealing ring automatic detection system with transportation,identification,sorting and other functions.