Study On The Image Processing And Matching Technique Of Multi-Bubbles In The Gas-Liquid Two Phase Flow

With the development of science technology, gas-liquid two-phase flow has been widely applied in many fields such as petroleum, chemical industry, metallurgy, pharmacy and nuclear energy. The effect on the engineering application depends largely on the bubble motion modality and the interaction between the discrete phase and the continuous phase. In order to study the inner mechanism of gas-liquid two-phase flow deeply, describe the interaction characteristics of bubbles accurately, gain more ideal application effect in the engineering practice, the extraction of three-dimensional parameters of gas-liquid two-phase flow is of great significance.In the paper, virtual binocular stereo vision measuring system is built based on the theory of binocular stereo vision and the construction design is optimized. As to the multiple bubbles objects, digital image processing, name matching and movement matching technique are studied. The details are as follows:1. Binocular stereo vision theory is further studied, and virtual binocular stereo vision measuring system is built using a single high-speed camera and two sets of front surface reflection mirrors. The construction design is optimized.2. The multiply bubbles parameters of gas-liquid two-phase flow are extracted with the help of digital image processing technique.3. In regard to the multiply bubbles images gained by virtual binocular stereo vision measuring system, two-way epipolar line threshold matching method is proposed considering the quality of the bubbles and the scene particularity. Combined with ordinate compatibility constraints, minimum distance constraints projection compatibility constraints and reverse epipolar line constraint, the name matching is finally completed. The experimental results show that name matching can be achieved well with the help of two-way epipolar line threshold matching method.4. Typical movement matching algorithms are summarized and studied., Algorithm application degree is analyzed with the multiply bubbles images characteristics of two continuous frames acquired by virtual binocular stereo vision measuring system., multiply bubbles matching of two consecutive frames images is realized based on the polar coordinate correlation algorithm. The experimental results prove that movement matching can be well done with the improved algorithm. 5. As to the multiply bubbles images in the gas-liquid two-phase flow, the three dimensional reconstruction of the multiply bubbles movement is finally accomplished using multiply bubbles name matching technique, multiply bubbles movement matching, calibrated camera internal and external parameters and sensor parameters. And the measurement accuracy of sensor is studied. The experimental results show that the measurement spatial distance error is better than 0.48mm, and the relative error is better than 1.15%.