Optimum design of vision sensor for the three-dimensional measurement of bubbles

We develop a novel virtual stereo vision sensor which contains just single high-speed camera and two reflector sets to measure three-dimensional feature parameters of bubbles in the gas-liquid two-phase flow. The measurement principle of the binocular stereo vision sensor and the virtual stereo vision sensor are analyzed in detail. According to the mathematical model of the virtual stereo vision sensor, the structure parameters of virtual stereo vision sensor are optimized based on the actual size of view field, and the three-dimensional measurement error is analyzed. The three-dimensional modality and motion parameters are tested on the upstream bubble producing appliance. The experimental results show that the three-dimensional distance error of the virtual stereo vision sensor is better than 0.132mm and the relative error is better than 0.5%, and the three-dimensional motion trail of bubble is reconstructed.