Abstract
Accurate detection and localization of solar photovoltaic (PV) panels insatellite imagery is essential for optimizing microgrids and activedistribution networks (ADNs), which are critical components of renewable energysystems. Existing methods lack transparency regarding their underlyingalgorithms or training datasets, rely on large, high-quality PV training data,and struggle to generalize to new geographic regions or varied environmentalconditions without extensive re-training. These limitations lead toinconsistent detection outcomes, hindering large-scale deployment anddata-driven grid optimization. In this paper, we investigate how large languagemodels (LLMs) can be leveraged to overcome these challenges. Despite theirpromise, LLMs face several challenges in solar panel detection, includingdifficulties with multi-step logical processes, inconsistent output formatting,frequent misclassification of visually similar objects (e.g., shadows, parkinglots), and low accuracy in complex tasks such as spatial localization andquantification. To overcome these issues, we propose the PV Assessment withLLMs (PVAL) framework, which incorporates task decomposition for more efficientworkflows, output standardization for consistent and scalable formatting,few-shot prompting to enhance classification accuracy, and fine-tuning usingcurated PV datasets with detailed annotations. PVAL ensures transparency,scalability, and adaptability across heterogeneous datasets while minimizingcomputational overhead. By combining open-source accessibility with robustmethodologies, PVAL establishes an automated and reproducible pipeline forsolar panel detection, paving the way for large-scale renewable energyintegration and optimized grid management.