Detection of photovoltaic solar panels using satellite image processing and fusion techniques: The case of Konya Cumra

The increasing global demand for energy, coupled with the depletion risk of fossil fuels and the necessity to reduce environmental impacts, has significantly emphasized the importance of renewable energy sources. Among these, photovoltaic (PV) panels have emerged as a prominent and environmentally friendly alternative for sustainable energy production. However, the widespread deployment of this technology necessitates the accurate detection and monitoring of the geographic distribution of solar panel installations. In the existing literature, the majority of studies have focused on site suitability analysis for PV installation and fault detection of operational systems. This thesis presents a comparative analysis of classification accuracies in the detection of PV panels by integrating spectral and spatial resolutions through image fusion techniques, using the NSPI as a spectral indicator and object-based classification approaches. PRISMA and Sentinel-2 satellite datasets were used to generate pansharpened images, upon which NSPI was computed and analyzed. The results indicate a decline in classification accuracy due to spectral distortion introduced by fusion processes, while the highest accuracy was achieved using the original hyperspectral data. This study represents one of the first academic efforts in Turkey to evaluate NSPI in conjunction with image fusion methods based on spectral and spatial resolution integration for PV panel detection. The findings contribute to the planning and management of solar energy projects, offering valuable insights for both national and international applications.projects.