UAV Photogrammetry under Poor Lighting Conditions-Accuracy Considerations

Area of Science:

• Geomatics Engineering
• Remote Sensing
• Photogrammetry

Background:

• Traditional photogrammetry relies heavily on daytime imagery.
• Night-time drone photogrammetry is an under-explored area due to lighting challenges.
• Understanding factors affecting night-time data quality is crucial for expanding applications.

Purpose of the Study:

• To investigate the feasibility and accuracy of Unmanned Aerial Vehicle (UAV) photogrammetry using night-time imagery.
• To analyze the geometric quality of night-time photogrammetric models compared to day-time models.
• To identify key factors influencing the accuracy and quality of night-time photogrammetric products.

Main Methods:

• Development and comparison of day- and night-time photogrammetric models.
• Extensive analysis of geometric accuracy and error sources.
• Evaluation of camera calibration, ground control point (GCP) distribution, and lighting conditions.
• Assessment of commercial photogrammetric software performance and camera impact on interpretative quality.

Main Results:

• Non-uniform GCP distribution significantly degraded night-time model geometry.
• Poor lighting conditions increased camera calibration errors (intrinsic orientation and distortion parameters).
• Uniformly illuminated night-time images resulted in lower reprojection errors and higher tie point precision.

Conclusions:

• Night-time drone photogrammetry is viable but sensitive to lighting and GCP distribution.
• Camera calibration and geometric accuracy are negatively impacted by low-light conditions.
• Recommendations are provided for optimizing night-time photogrammetric data acquisition and processing.