Identification And Tracking Of Gas-liquid Two-phase Flow And Flow Velocity Measurement In Small Pipeline Based On Machine Vision

Gas-liquid two-phase flow widely exists in nature and industry,and due to the improvement of manufacturing technology and the development of small-scale industrial equipment,the research on the mechanism of gas-liquid two-phase flow in small pipes has attracted more and more scholars’ attention,and the research on parameters’ measurement methods has important theoretical significance and application value.However,due to the complexity of gas-liquid two-phase flow in small pipes,the existing detection methods still need to be improved.Based on the technology of computer vison,this thesis designs a flow pattern detection,identification and tracking scheme for the gas-liquid two-phase flow in a small pipeline,and realizes the flow velocity detection.The main contents and contributions of this thesis are as follows:(1)Three types of gas-liquid two-phase flow pattern recognition methods based on Support Vector Machines(SVM)in small pipes are compared.The Hu moments of images obtained by preprocessing are used as features to establish the recognition model,and the identification of four common flow patterns is realized.(2)Based on the improved Canny algorithm,a bubble target detection method for gas-liquid two-phase flow in small pipes is presented.(3)A target tracking and velocity detection method for two-phase flow in a small pipe is presented.The position of bubble and slug obtained is located.By combining the rotation angle of the pan-tilt and the pixel compensation processing technology,bubble in different pictures is tracked and velocity measurement of bubble is carried out.In this thesis,the flow rate measured by this method is in agreement with the experimental calibration flow rate,and a better tracking effect is achieved.(4)An experimental device is built for vertification of presented methods.The experimental results show that the presented methods are effective.