Paper Title
APPLYING NO-CODE AI FOR THE DETECTION AND TRACKING OF SOLAR PANELS FROM SATELLITE IMAGES

Abstract
Among various renewable energy sources, solar energy is one of the most promising in many countries. Consequently, localizing and monitoring the development of solar panel plants is crucial for management, sustainability, maintenance, and strategic planning in the energy sector. The main objective of this research is to develop an efficient approach for detecting and tracking Photovoltaic (PV) solar panels using satellite imagery. We present a data-centric approach utilizing modern AI techniques, specifically a no-code object detection provided by Landing AI platform. This method allows researchers to focus on the selection and quality of training and development data. Our approach was evaluated using a dataset of 2101 PV solar panels and arrays extracted from 49 Google Earth satellite images from various regions in Dubai, UAE. The model achieved an F1-score of approximately 90% on both validation and testing datasets, demonstrating its effectiveness in detecting and tracking PV solar panels. Future work will focus on evaluating the model on larger datasets, which is a necessary step before system deployment. Keywords - PV Solar Panel, Satellite imagery analysis, No-coding object detection, ML in Energy Management.