Optimal Design of Monocular Stereo Vision System

Abstract

To compensate for the deficiencies of current monocular and binocular stereo vision systems, a novel monocular stereo vision system configuration that simulates the human vision system but surpasses its functions is designed based on Stewart platform. For the sake of improving the searching range and ranging accuracy, throughout analyzing the kinematic and dynamic models, nondominated sorting genetic algorithm II (NSGA-II) is adopted to conduct the multi-objective optimization of the performance indexes including the kinematic performance index, the maximum inclination of the upper platform as well as the dynamic performance index under mechanical and dimensional constraints. The prototype was built according to the optimized structural parameters. Actual measurement data show that the monocular stereo vision system designed can meet the requirements.