Elastic Computing Platform Design Based on Lunar Rover Stereo Vision System

Abstract

In the second stage of Chinese lunar exploration program, lunar rover will be launched for the moon. Stereo vision system of the lunar rover is used for navigation and exploration. Binocular stereo vision cameras collect stereo image pairs of surrounding environment which are transmitted to ground station then used to reconstruct lunar 3D terrain environment. 112 image pairs are collected at one site and terrain data increases rapidly as the rover explores more unknown area so that sequential process system is difficult to generate 3D terrain environment in time. To solve this problem, an elastic parallel computing platform is designed and built using OpenMP parallel computing framework on a 16-processor shared-memory machine. In this paper, architecture of the elastic computing platform is introduced and experiments based on real terrain datasets show that the maximum speedup of the platform can reach 6.44.