Research And Development Of Hose Quality Inspection System Based On Machine Vision

Automotive rubber conduits are the key components for transmitting oil,water,gas,and hydraulic power in automobile engines,and their quality directly affects the engine’s operating stability and automobile driving safety.Strict quality inspection is required during the production of automotive rubber conduits.The inspection of the quality index of dimensions is currently performed manually by using contact measurement tools.Due to the deformable nature of rubber materials,the accuracy of contact measurement is subjectively affected Large and low inspection efficiency,cannot fit the requirements of modern production.This dissertation is devoted to the development of an automatic inspection system for the size and quality of hoses based on machine vision technology.The purpose is to replace manual inspection and improve inspection accuracy and efficiency.This thesis first analyzes the requirements of hose size measurement and proposes the overall scheme and design principles of the system.In the design of the hardware,the key parameters and specific models of the equipment,such as the light source,camera,and frame grabber,are determined.In terms of software,the functional design of each module and formulated the overall algorithm flow.To improve the accuracy of hose size detection,this thesis proposes a method for extracting measurement points based on subpixel edges and polar coordinate axes.This method first uses the camera calibration to obtain pixel equivalents;then performs preprocessing and ROI selection on the original image to obtain pixel-level edge of the hose cross-section without fiber layer interference;this edge serves as a coarse positioning,and sub-pixel edge location can be obtained by fitting through Zernike moment sub-pixel edge detection method.A series of dimensional measurement points are obtained by establishing multiple polar axes with centroid as the origin and the intersection of the boundary curve.Compared with the traditional detection algorithm,the sub-pixel operator can automatically obtain the optimal value of step gray,which improves the robustness and measurement accuracy of the system.Through software platform construction,use the measurement mentioned above algorithm to carry outsize measurement and quality inspection classification on the hose.In summary,this thesis applies machine vision technology to the dimensional measurement link and designs a reliable and high-precision dimensional measurement system.After experiments,the measurement accuracy is higher than 0.05 mm,and the accuracy rate is 97.87%.The efficiency is about three times higher than that of manual inspection.It meets the requirements of enterprises for the quality inspection of hoses and has portability for the measurement of irregularly shaped annular workpieces,which has certain practical application value.