Multi-camera visual SLAM for off-road navigation

• DOI: 10.1016/j.robot.2020.103505
• Publication Date: 2020-03-26T20:09:53Z
• Authors: Yi Yang; Di Tang; Dongsheng Wang; Wenjie Song; Junbo Wang; Mengyin Fu
• Keywords: Monocular

With the rapid development of computer vision, vision-based simultaneous localization and mapping (vSLAM) plays an increasingly important role in field unmanned driving. However, traditional SLAM methods based on a monocular camera only perform well simple indoor environments or urban with obvious structural features. In off-road environments, situation that encounters could be complicated by problems such as direct sunlight, leaf occlusion, rough roads, sensor failure, sparsity stably trackable texture. Traditional are highly susceptible to these factors, which lead compromised stability reliability. To counter problems, we propose panoramic vision method multi-camera collaboration, aiming at utilizing characters stereo perception improve precision environments. At same time, independence information sharing each system can its ability resist bumps, illumination, occlusion sparse texture environment, enable our recover metric scale. These ensure ground vehicles (UGVs) locate navigate safely reliably complex