Vision-based UAV Altitude Estimation using Deep Learning: A ResNet50 Approach on Nadir Images

Year 2026, Volume: 14 Issue: 1, 46 - 54, 31.01.2026

Ahmet Ertuğrul Arık

https://doi.org/10.21541/apjess.1730867

Abstract

This study proposes a vision-based deep learning approach for unmanned aerial vehicle (UAV) altitude estimation as an alternative to traditional methods such as GPS, barometric sensors, and laser altimeters, which are often sensitive to environmental disturbances. A large-scale dataset of 303,710 nadir images was collected using Mavic 2 Pro and Mavic 2 Zoom platforms under diverse weather, illumination, and terrain conditions. Each image was labeled with above-ground-level (AGL) altitude by integrating EXIF-based GPS altitude with a 30 m Digital Elevation Model (DEM) through coordinate transformation and terrain subtraction. A pre-trained ResNet50 model, originally designed for image classification, was reconfigured as a regression network and fine-tuned for 200 epochs using the Adam optimizer and mean squared error (MSE) loss. The proposed model achieved a mean absolute error (MAE) of 4.09 m in urban areas and 6.06 m in rural areas, with R^2 scores of 0.9981 and 0.9804, respectively. Comparative experiments with alternative CNN architectures show that the adapted ResNet50 provides favorable accuracy versus complexity trade-off. These results indicate that the proposed monocular, nadir-image-based framework can serve as a reliable complement or alternative to conventional altitude sensors in UAV operations.

Keywords

UAV Altitude Estimation , Deep Learning , Digital Elevation Model (DEM); , Nadir Aerial Images; , ResNet50 for Regression;

References

• Valavanis, K. P. and Vachtsevanos, G. J., eds. Handbook of unmanned aerial vehicles. 2015.
• Colomina, I. and Molina, P. Unmanned aerial systems for photogrammetry and remote sensing: A review. ISPRS Journal of Photogrammetry and Remote Sensing, 2014. 92 :79–97. doi:10.1016/j.isprsjprs.2014.02.013.
• Çeti̇n, B. T., Aygün, C., Toprak, O., and Çeli̇K, S. Design of a multi-purpose vertical take-off and landing unmanned aerial vehicle. ASREL, 2024. :2. doi:10.56753/ASREL.2024.2.3.
• Agrawal, P. S., Jawarkar, P. S., Dhakate, K. M., Parthani, K. M., and Agnihotri, A. S. Advancements and challenges in drone technology: A comprehensive review. 2024 4th international conference on pervasive computing and social networking (ICPCSN), 2024. :638–644. doi:10.1109/ICPCSN62568.2024.00106.
• Czarnecki, T., Stawowy, M., and Kadłubowski, A. Cost-effective autonomous drone navigation using reinforcement learning: Simulation and real-world validation. Applied Sciences, 2024. 15(1) :179. doi:10.3390/app15010179.
• Eling, C., Klingbeil, L., Wieland, M., and Kuhlmann, H. A PRECISE POSITION AND ATTITUDE DETERMINATION SYSTEM FOR LIGHTWEIGHT UNMANNED AERIAL VEHICLES. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., 2013. XL-1/W2 :113–118. doi:10.5194/isprsarchives-XL-1-W2-113-2013.
• Barshan, B. and Durrant-Whyte, H. F. Inertial navigation systems for mobile robots. IEEE Trans. Robot. Automat., 1995. 11(3) :328–342. doi:10.1109/70.388775.
• Siwiec, J. Comparison of airborne laser scanning of low and high above ground level for selected infrastructure objects. Journal of Applied Engineering Sciences, 2018. 8(2) :89–96. doi:10.2478/jaes-2018-0023.
• Raxit, S., Singh, S. B., and Al Redwan Newaz, A. YoloTag: Vision-based robust UAV navigation with fiducial markers. 2024 33rd IEEE international conference on robot and human interactive communication (RO-MAN), 2024. :311–316. doi:10.1109/RO-MAN60168.2024.10731319.
• Gharsa, O., Touba, M. M., Boumehraz, M., Abderrahman, N., Bellili, S., and Titaouine, A. Autonomous landing system for a quadrotor using a vision-based approach. 2024 8th international conference on image and signal processing and their applications (ISPA), 2024. :1–5. doi:10.1109/ISPA59904.2024.10536784.
• Piyakawanich, P. and Phasukkit, P. An AI-based deep learning with k-mean approach for enhancing altitude estimation accuracy in unmanned aerial vehicles. Drones, 2024. 8(12) :718. doi:10.3390/drones8120718.
• Hrabar, S. 3D path planning and stereo-based obstacle avoidance for rotorcraft UAVs. 2008 IEEE/RSJ international conference on intelligent robots and systems, 2008. :807–814. doi:10.1109/IROS.2008.4650775.
• Arafat, M. Y., Alam, M. M., and Moh, S. Vision-based navigation techniques for unmanned aerial vehicles: Review and challenges. Drones, 2023. 7(2) :89. doi:10.3390/drones7020089.
• Zhang, X., He, Z., Ma, Z., Jun, P., and Yang, K. VIAE-Net: An end-to-end altitude estimation through monocular vision and inertial feature fusion neural networks for UAV autonomous landing. Sensors, 2021. 21(18) :6302. doi:10.3390/s21186302.
• Kupervasser, O., Kutomanov, H., Levi, O., Pukshansky, V., and Yavich, R. Using deep learning for visual navigation of drone with respect to 3D ground objects. Mathematics, 2020. 8(12) :1–13. doi:10.3390/math8122140.
• Fragoso, A. T., Lee, C. T., Mccoy, A. S., and Chung, S.-J. A seasonally invariant deep transform for visual terrain-relative navigation. 2021.
• Unlu, E., Zenou, E., Riviere, N., and Dupouy, P.-E. Deep learning-based strategies for the detection and tracking of drones using several cameras. IPSJ T Comput Vis Appl, 2019. 11(1) :7. doi:10.1186/s41074-019-0059-x.
• Zheng, H., Rajadnya, S., and Zakhor, A. Monocular depth estimation for drone obstacle avoidance in indoor environments. 2024 IEEE/RSJ international conference on intelligent robots and systems (IROS), 2024. :10027–10034. doi:10.1109/IROS58592.2024.10802577.
• Ghasemieh, A. and Kashef, R. Advanced monocular outdoor pose estimation in autonomous systems: Leveraging optical flow, depth estimation, and semantic segmentation with dynamic object removal. Sensors, 2024. 24(24) :8040. doi:10.3390/s24248040.
• Gurram, A., Tuna, A. F., Shen, F., Urfalioglu, O., and López, A. M. Monocular depth estimation through virtual-world supervision and real-world SfM self-supervision. 2022.
• Navarro, D., Antoine, R., Malis, E., and Martinet, P. Towards autonomous robotic structure inspection with dense-direct visual-SLAM. 2024 32nd european signal processing conference (EUSIPCO), 2024. :2017–2021. doi:10.23919/EUSIPCO63174.2024.10715467.
• Pilartes-Congo, J. A., Simpson, C., Starek, M. J., Berryhill, J., Parrish, C. E., and Slocum, R. K. Empirical evaluation and simulation of the impact of global navigation satellite system solutions on uncrewed aircraft system–structure from motion for shoreline mapping and charting. Drones, 2024. 8(11) :646. doi:10.3390/drones8110646.
• Mienye, I. D. and Swart, T. G. A comprehensive review of deep learning: Architectures, recent advances, and applications. Information, 2024. 15(12) :755. doi:10.3390/info15120755.
• Wells, D. and others. Guide to GPS positioning, lecture notes. 1999.
• Szypuła, B. Digital elevation models in geomorphology. Hydro-geomorphology - models and trends, 2017. doi:10.5772/intechopen.68447.
• Snyder, J. P. Map projections: A working manual. 1987.
• DJI Mavic 2 Product Information. 2018.
• He, K., Zhang, X., Ren, S., and Sun, J. Deep residual learning for image recognition. 2015. doi:10.48550/arXiv.1512.03385.
• Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. ImageNet: A large-scale hierarchical image database. 2009 IEEE conference on computer vision and pattern recognition, 2009. :248–255. doi:10.1109/CVPR.2009.5206848.
• Simonyan, K. and Zisserman, A. Very deep convolutional networks for large-scale image recognition. 2015. doi:10.48550/arXiv.1409.1556.
• Szegedy, C. and al., et. Going deeper with convolutions. 2014. doi:10.48550/arXiv.1409.4842.