UAV-Borne Mapping Algorithms for Low-Altitude and High-Speed Drone Applications

Jincheng Zhang,Artur Wolek,Andrew R. Willis

DOI: https://doi.org/10.3390/s24072204

2024-03-30

Abstract:This article presents an analysis of current state-of-the-art sensors and how these sensors work with several mapping algorithms for UAV (Unmanned Aerial Vehicle) applications, focusing on low-altitude and high-speed scenarios. A new experimental construct is created using highly realistic environments made possible by integrating the AirSim simulator with Google 3D maps models using the Cesium Tiles plugin. Experiments are conducted in this high-realism simulated environment to evaluate the performance of three distinct mapping algorithms: (1) Direct Sparse Odometry (DSO), (2) Stereo DSO (SDSO), and (3) DSO Lite (DSOL). Experimental results evaluate algorithms based on their measured geometric accuracy and computational speed. The results provide valuable insights into the strengths and limitations of each algorithm. Findings quantify compromises in UAV algorithm selection, allowing researchers to find the mapping solution best suited to their application, which often requires a compromise between computational performance and the density and accuracy of geometric map estimates. Results indicate that for UAVs with restrictive computing resources, DSOL is the best option. For systems with payload capacity and modest compute resources, SDSO is the best option. If only one camera is available, DSO is the option to choose for applications that require dense mapping results.

Robotics,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper mainly discusses the mapping algorithm problem used in low-altitude high-speed flight of unmanned aerial vehicles (UAVs). The study created a highly realistic experimental environment to evaluate the performance of three different mapping algorithms: Direct Sparse Odometry (DSO), Stereo DSO (SDSO), and DSO Lite (DSOL), in combination with the AirSim simulator, Google 3D map model, and Cesium Tiles plugin. The experimental results compared the algorithms based on geometric accuracy and computational speed, providing a basis for researchers to choose mapping solutions suitable for specific applications.

The paper points out that challenges such as sensor performance, feature visibility, and geometric distortion are increased during low-altitude flight, while high-speed flight requires handling synchronization and integration issues between large-distance sensor readings. The authors recommend that DSOL should be chosen for UAVs with limited computing resources; SDSO is the best choice for systems with payload capacity and moderate computing resources; and DSO should be selected if only one camera is available and dense mapping results are needed.

In addition, the paper analyzes the development of mapping algorithms, including traditional photogrammetry and advanced SLAM techniques, emphasizing the impact of flight altitude on algorithm performance and comparing the advantages and disadvantages of Structure-from-Motion (SfM) and stereo reconstruction methods. Finally, the paper proposes a theoretical foundation and technical approach for sensor selection in low-altitude high-speed UAV mapping applications, as well as an innovative simulation method for extracting high-fidelity geometric models and analyzing algorithms.

In summary, this paper aims to address the problem of selecting and optimizing mapping algorithms for low-altitude high-speed UAV applications, providing valuable references for related research areas.