Integrating GNSS/IMU and DEM data for precise aerial triangulation: Insights from airborne hybrid systems in upper Egypt

Digital photogrammetry primarily aims to extract three-dimensional coordinates (X, Y, Z or E, N, H) of feature points, which is crucial for mapping applications. The Aerial Triangulation (AT) process for aerial images must be adjusted with high precision to achieve accurate measurements. Enhancing the accuracy of Global Navigation Satellite System (GNSS) and Inertial Measurement Unit (IMU) sensors significantly improves the AT process. Additionally, Airborne Light Detection and Ranging (LiDAR) data can produce a high-resolution Digital Elevation Model (DEM), which aids in initializing the aerial triangulation process. Modern services, such as Real-Time eXtended (RTX), are also used for GNSS/IMU corrections, further refining their accuracy.

The novelty of the current research is based on an end-to-end procedure for enhancing AT accuracy, especially in variable terrain height regions, using a hybrid airborne system. The scope is to use GNSS/IMU data coupled with a DEM from airborne LiDAR to initialize the AT process. The study cases were based in Maghagha City, Minia Governorate, Egypt, where a flight mission was carried out in 2017 using the Trimble AX60 system. This system integrates a photogrammetric camera and laser scanner with GNSS/IMU sensors. The aerial triangulation of the images was processed using MATCH-AT software. The accuracy of the results was evaluated using checkpoints. The findings indicate that AT using GNSS/IMU corrected data yields the best accuracy in AT, particularly in the Z direction, with an accuracy enhancement in check points residuals, compared with AT without using GNSS/IMU. Consequently, the final Root Mean Square (RMS) improved from 0.25 m to 0.17 m in E, from 0.2 m to 0.17 m in N, and from 3 m to 0.5 m in H. That demonstrates the significant benefit of incorporating GNSS/IMU data in improving the precision of three-dimensional spatial measurements. In addition, the DEM initialization improved the RMS slightly, also, the matching between aerial images during the triangulation process gets better values along the iteration time.