Synchronization and calibration of a stereo vision system

摘要

In this paper, we design a stereo image and video acquisition system with adjustable baseline and relative angle. Stereo vision has been used both for field robotic navigation and 3D wave reconstruction. However, there has been inadequate documentation about the synchronization performance of such system. For a dynamically moving scene, such as the wave, any delay between the two cameras will affects the accuracy of the disparity map. The present study presents a physical process to measure the delay between the two cameras. Firstly, we obtain the maximum frame rate of camera when it is triggered by external signal. Then we adopt a free fall experiment to measure the delay between trigger intervals at maximum frame rate. The relationship between baseline distance, calibration checker box tile size and target distance is discussed. Two calibration tools, i.e. MATLAB and WASS are compared using a range of distances.