Satellite Image-Based Rooftop Solar Assessment Using CV and Geospatial Analysis

Abstract

The Solar Roof Measurement and Energy Estimation System analyzes satellite photos of roofs. The premium version uses A.I. to determine if a roof is suitable for solar panels and calculates its area. The system has a FastAPI backend and a React frontend. This setup makes it easy to process information and interact with users. There are no expensive commercial applications in the algorithm pipeline for the Solar Roof Measurement and Energy Estimation System. Instead, it relies on free and openly available technology to create solar estimates. It uses Nominatim geocoding, Leaflet maps, OpenCV, and pvlib. The system measures the roof’s size and pitch, and calculates its orientation and solar energy potential. With resources like Google Earth imagery and Google Maps, it provides a scalable and easily accessible way to plan renewable energy. It gives homeowners, businesses, and solar providers a fast, simple method to check if their roof is suitable for solar panels. The Solar Roof Measurement and Energy Estimation System includes image analysis, map, or direction search. The Solar Roof Measurement and Energy Estimation System is particularly helpful for users of energy that is efficient, sustainable, and environmentally friendly. This product quickly and affordably analyzes roofs, estimates energy, and aids in developing smart energy solutions.

Introduction

The text presents a Solar Roof Measurement and Energy Estimation System that uses Artificial Intelligence, computer vision, and machine learning to automate rooftop solar potential analysis. Due to increasing demand for clean and sustainable energy, rooftop solar systems are becoming important, but traditional rooftop surveys are costly, time-consuming, and not scalable. The proposed system addresses this by using satellite imagery and open-source tools (OpenCV, FastAPI, ReactJS, pvlib) to analyze rooftops automatically.

The system performs key tasks such as rooftop detection, area calculation, shadow analysis, orientation and tilt estimation, and solar energy prediction. Techniques like YOLOv8, OpenCV contour detection, K-means clustering, PCA, and Hough Transform are used to extract rooftop boundaries and assess usable solar areas. Shadow detection helps determine the effective installation space, while photovoltaic models estimate energy generation using standard equations involving area, irradiance, and performance ratio.

The system also provides financial and environmental analysis, including installation cost, savings, payback period, and CO? reduction. A dataset is generated with geometric, environmental, and economic parameters to support analysis.

Overall, the study proposes a low-cost, automated, and scalable AI-based solution for accurate rooftop solar potential estimation, improving decision-making for renewable energy adoption and supporting sustainable development.

Conclusion

The proposed Solar Roof Measurement and Energy Estimation System successfully integrates Computer Vision, Machine Learning, photovoltaic energy modelling, and financial analysis into a unified platform for automated rooftop solar assessment. The system performs rooftop detection, segmentation, shadow analysis, rooftop orientation estimation, photovoltaic energy prediction, and economic feasibility analysis using satellite imagery and Artificial Intelligence techniques. The proposed methodology combines YOLOv8 segmentation, OpenCV contour analysis, polygon-based rooftop area computation, irradiance estimation, and K-Means clustering to generate accurate solar suitability analysis for residential and commercial buildings. Experimental analysis demonstrated satisfactory rooftop segmentation accuracy and realistic photovoltaic energy estimation results. The clustering model successfully categorized rooftops into different solar suitability groups based on rooftop geometry, shadow conditions, and energy generation characteristics. The generated outputs include rooftop area, usable rooftop space, photovoltaic system capacity, annual energy generation, installation cost, annual savings, and payback period. The proposed framework provides a scalable, low-cost, and reproducible solution for rooftop solar viability assessment using open-source technologies such as Python, FastAPI, OpenCV, PyTorch, PVLib, and ReactJS. Although the proposed system achieved satisfactory results, future improvements may include integration of 3D rooftop geometry reconstruction, seasonal shadow analysis, weather forecasting models, and battery storage optimization for more accurate photovoltaic energy prediction. Overall, the proposed AI-based rooftop solar analysis system provides an efficient and intelligent solution for renewable energy planning, smart city development, and sustainable rooftop solar deployment.

References

[1] G. Bradski, “The OpenCV Library,” Dr. Dobb’s Journal of Software Tools, 2000. [2] OpenWeather, “Solar Irradiance and Energy Prediction API,” 2023. Available: https://openweathermap.org/api. [3] R. Castello et al., “Quantification of Suitable Rooftop Area for Photovoltaic Installation using Convolutional Neural Networks,” Journal of Physics: Conference Series, vol. 2042, no. 1, 2021. [4] F. Andrieux, “Predicting Solar Potential of Rooftops using Segmentation,” namR Blog, 2020. [5] J. Redmon et al., “YOLO: Unified, Real-Time Object Detection,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016. [6] M. Ahmed et al., “Rooftop Detection with YOLOv8 for Photovoltaic Assessment,” International Journal of Artificial Intelligence, vol. 14, no. 3, 2025. [7] J. Zeng et al., “Intelligent Segmentation of Urban Building Roofs and Solar Energy Potential Estimation,” Journal of Imaging, Sep. 2025. [8] S. Egbe et al., “Large-Scale Rooftop Solar Energy Estimation Using Deep Learning,” BMVC Workshops, 2025. [9] D. Bautista et al., Automatic Solar Panel Detection using YOLO and FastAPI, GitHub Repository, 2024. [10] D. Arthur and S. Vassilvitskii, “K-Means++: The Advantages of Careful Seeding,” Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms, 2007. [11] F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011. [12] W. McKinney, “Data Structures for Statistical Computing in Python,” Proceedings of the 9th Python in Science Conference, 2010. [13] SolarPerformanceInsight, “FastAPI and PVLib Based Solar Energy Estimation Framework,” GitHub Repository, 2025. [14] T. M., “Solar Roof Measurement with Computer Vision,” Roboflow Blog, Dec. 2024. [15] T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016.

Copyright

Copyright © 2026 Dr. Ujwala M. Patil, Leena M. Gavande, Vedashri M. Borgaonkar, Bhagyashri N. Bhadane. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.