Prompt Strategies for Multimodal LLM-Based Rooftop Solar Panel Detection

This dataset contains structured prompt templates used for multimodal large language model (LLM)-based detection of rooftop solar photovoltaic (PV) panels in grayscale satellite imagery. The dataset includes four classification strategies: 1. Feature-Based Baseline Classification 2. Example-Based Binary Classification 3. Classification Based on Stepwise Feature Analysis 4. Hypothesis-Based Classification The dataset supports methodological transparency, reproducibility, and comparative evaluation of prompt engineering strategies in remote sensing applications. All prompts are provided in Markdown format.