An Autonomous Positioning Method for Drones in GNSS Denial Scenarios Driven by Real-Scene 3D Models

Area of Science:

• Robotics and Autonomous Systems
• Computer Vision
• Geospatial Technology

Background:

• Drones require reliable positioning for military and civilian applications.
• Global Navigation Satellite Systems (GNSS) are prone to interference and signal loss.
• Enhancing drone positioning accuracy and independence from GNSS is crucial.

Purpose of the Study:

• To develop a novel autonomous drone positioning system.
• To achieve high positioning accuracy independent of GNSS.
• To validate the system's performance in complex urban environments.

Main Methods:

• Utilizing a real-scene 3D model and image point cloud reconstruction.
• Segmenting 3D models into image and point cloud datasets.
• Performing real-time image capture with a monocular camera.
• Employing image matching for preliminary position estimation and 3D point cloud reconstruction.
• Matching reconstructed point clouds with real-scene 3D point cloud data.
• Applying pose estimation algorithms for coordinate acquisition.

Main Results:

• Successful autonomous positioning of drones in complex urban settings.
• Achieved remarkable positioning accuracy of up to 0.4 meters.
• Demonstrated the efficacy of the 3D model and point cloud reconstruction approach.

Conclusions:

• The proposed method offers a robust solution for GNSS-independent drone positioning.
• High accuracy positioning is attainable even in challenging environments.
• This technology has significant implications for drone autonomy and navigation safety.