Monocular Vision System for Fixed Altitude Flight of Unmanned Aerial Vehicles

Area of Science:

• Robotics and Automation
• Computer Vision
• Photogrammetry

Background:

• Aerial photography is crucial for terrain imaging, but unmanned aerial vehicles (UAVs) face challenges in maintaining stable flight altitude.
• Existing methods for altitude control in UAVs can be complex and may not guarantee consistent image resolution.

Purpose of the Study:

• To develop and validate a robust method for maintaining UAV flight altitude using integrated computer vision techniques.
• To enable consistent terrain image acquisition and accurate relative altitude measurement for aerial surveys.

Main Methods:

• A system combining skyline detection with a stereo vision algorithm was developed for UAVs.
• A monocular camera captured ground images, while stereo vision processed UAV velocity to determine relative altitude.
• A forward-looking camera and skyline detection algorithm were employed to ensure flight stability.

Main Results:

• The proposed system successfully enabled UAVs to maintain a fixed, low flight altitude.
• Consistent terrain image resolution was achieved throughout the flight path.
• Accurate detection of relative altitude from the ground to the UAV was demonstrated.

Conclusions:

• The integrated skyline detection and stereo vision system effectively addresses UAV flight altitude maintenance challenges.
• This method enhances the reliability and efficiency of UAV-based aerial photography for terrain imaging.
• The system facilitates stable flight patterns and consistent image data acquisition for various applications.