Large-scale environment mapping and immersive human-robot interaction for agricultural mobile robot teleoperation

14 Aug 2023  ·  Tao Liu, Baohua Zhang, Qianqiu Tan ·

Remote operation is a crucial solution to problems encountered in agricultural machinery operations. However, traditional video streaming control methods fall short in overcoming the challenges of single perspective views and the inability to obtain 3D information. In light of these issues, our research proposes a large-scale digital map reconstruction and immersive human-machine remote control framework for agricultural scenarios. In our methodology, a DJI unmanned aerial vehicle(UAV) was utilized for data collection, and a novel video segmentation approach based on feature points was introduced. To tackle texture richness variability, an enhanced Structure from Motion (SfM) using superpixel segmentation was implemented. This method integrates the open Multiple View Geometry (openMVG) framework along with Local Features from Transformers (LoFTR). The enhanced SfM results in a point cloud map, which is further processed through Multi-View Stereo (MVS) to generate a complete map model. For control, a closed-loop system utilizing TCP for VR control and positioning of agricultural machinery was introduced. Our system offers a fully visual-based immersive control method, where upon connection to the local area network, operators can utilize VR for immersive remote control. The proposed method enhances both the robustness and convenience of the reconstruction process, thereby significantly facilitating operators in acquiring more comprehensive on-site information and engaging in immersive remote control operations. The code is available at: https://github.com/LiuTao1126/Enhance-SFM

PDF Abstract

Datasets


  Add Datasets introduced or used in this paper

Results from the Paper


  Submit results from this paper to get state-of-the-art GitHub badges and help the community compare results to other papers.

Methods


No methods listed for this paper. Add relevant methods here