Humanitarian demining robot Gryphon: New vision techniques and optimization methods

摘要

Landmines continue killing and injuring people even many years after the end of conflicts. Mechanical systems and robots have been used for assisting in landmines location and improving the quality of humanitarian demining tasks. These new systems are expected to enhance safety, efficiency, time duration, etc, but many improvements are still needed in order to realize a practical system. In this paper, new vision techniques for improving the effectiveness of terrain mapping using stereo vision camera under high-contrast light conditions, i.e., mixed direct sunlight and shadowed areas, are introduced. Furthermore, trajectory optimization taking in account the obtained terrain map is also analyzed, which greatly reduces time consumption for landmine location marking tasks. These newly introduced methods greatly contribute to overall increase in reliability and efficiency of humanitarian demining tasks.