Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring the potential of collaborative UAV 3D mapping in Kenyan savanna for wildlife research

Vandita Shukla, Luca Morelli, Pawel Trybala, Fabio Remondino, Wentian Gan, Yifei Yu, Xin Wang

Abstract

UAV-based biodiversity conservation applications have exhibited many data acquisition advantages for researchers. UAV platforms with embedded data processing hardware can support conservation challenges through 3D habitat mapping, surveillance and monitoring solutions. High-quality real-time scene reconstruction as well as real-time UAV localization can optimize the exploration vs exploitation balance of single or collaborative mission. In this work, we explore the potential of two collaborative frameworks - Visual Simultaneous Localization and Mapping (V-SLAM) and Structure-from-Motion (SfM) for 3D mapping purposes and compare results with standard offline approaches.

Exploring the potential of collaborative UAV 3D mapping in Kenyan savanna for wildlife research

Abstract

UAV-based biodiversity conservation applications have exhibited many data acquisition advantages for researchers. UAV platforms with embedded data processing hardware can support conservation challenges through 3D habitat mapping, surveillance and monitoring solutions. High-quality real-time scene reconstruction as well as real-time UAV localization can optimize the exploration vs exploitation balance of single or collaborative mission. In this work, we explore the potential of two collaborative frameworks - Visual Simultaneous Localization and Mapping (V-SLAM) and Structure-from-Motion (SfM) for 3D mapping purposes and compare results with standard offline approaches.
Paper Structure (7 sections, 3 figures, 5 tables)

This paper contains 7 sections, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Data
  3. Methodology
  4. Offline processing
  5. Real-time processing
  6. Results
  7. Conclusions

Figures (3)

  • Figure 1: Visualization of image orientation results for dataset 1. (a) Degenerate trajectories from Metashape; (b) Degenerate trajectory recovered for agent 1 in COLMAP, shown from top and side view; (c) Trajectories in in COLMAP (SuperPoint) for both agents, shown fr top and side view; (d) OtF-SFM trajectories for both agents shown from an oblique view point; (e) OtF-SfM (SuperPoint) trajectories for both agents shown from oblique view point; (f) CCM-SLAM trajectories for individual agents shown from top and oblique front view (green point cloud represents mapping through agent 1 and red point cloud from agent 2).
  • Figure 2: Visualization of image orientation results for dataset 2. (a) Degenerate trajectories from Metashape; (b) Trajectories recovered from all three agents in COLMAP shown from nadiral and side view; (c) Trajectories recovered with in COLMAP (SuperPoint) for two agents shown from top and side view; (d) OtF-SfM trajectories for two agents shown from oblique view points; (e) OtF-SfM (SuperPoint) trajectories for two agents shown from nadiral and side view point; (f) Collaborative mapping results in CCM-SLAM for two agents (1 and 3) shown from top and oblique side view (red lines indicate the location matches in key frames).
  • Figure 3: CCM-SLAM loses tracking of features when the camera undergoes rotation movement. RQt visualization shows the reduction in tracked features as the drone turns.