Depth camera based remote three-dimensional reconstruction using incremental point cloud compression

Abstract

Real-time three-dimensional reconstruction is a fundamental technology applied by robots in unknown environments, especially in remote operating systems. This paper proposes a remote indoor environment reconstruction method based on incremental point cloud compression, which improves the effectiveness of teleoperation-based 3D mapping in the context of hazardous area work environments. First, due to the influence of data bandwidth in remote map building, this paper adds point cloud compression to the keyframe-based building method to form a new reconstruction technique using incremental point cloud compression, so as to realize the mapping of the remote environment with small data transmission. Second, a teleoperation system that integrates a remote robot platform and a local operation platform is constructed. Remote robot equipped with a depth camera can better complete the environmental reconstruction under the guidance of the local operator. Finally, we build the actual scene for 3D construction and perform task planning in the reconstruction model. Experimental results and data prove the practicality of our system.