SolarMap-India: A Georeferenced Satellite Imagery Dataset for Solar Panel Identification and Mapping

Engineering notes

Engineering notes will be added by the Power for Solar editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为光伏效率、钙钛矿太阳能电池、储能技术、太阳能热利用、BIPV、并网技术等高价值论文补充中文说明。

Original abstract

The increasing deployment of rooftop solar photovoltaic (PV) systems has created a growing demand for reliable geospatial datasets that support automated solar panel identification, mapping, and large-scale renewable energy assessment using remote sensing and computer vision techniques. This work presents a curated geospatial satellite imagery dataset developed for solar panel detection and segmentation, with a primary focus on diverse urban and semi-urban regions across India. The dataset was constructed using geographic coordinates represented by latitude and longitude values, which served as spatial reference points for acquiring satellite imagery through the Google Maps Static API. An initial collection of 3,000 georeferenced satellite images was obtained, with each image associated with its corresponding latitude and longitude information, enabling spatial traceability and location-based analysis. The collected imagery represents geographically diverse regions, with substantial coverage across Gujarat and additional locations spanning multiple Indian states. Following systematic screening and visual assessment, approximately 2,500 images containing identifiable solar photovoltaic installations were selected for detailed annotation. The selected imagery was divided into four subsets and annotated in parallel using a standardized annotation workflow. Solar panel regions were precisely delineated and exported in the Common Objects in Context (COCO) format, providing structured image metadata, object-level annotations, segmentation information, and consistent class definitions. The independently annotated subsets were subsequently consolidated into a unified dataset through identifier remapping, category standardization, and annotation integrity verification. The resulting dataset integrates geospatial coordinate information with detailed solar panel annotations, providing a structured resource for developing and evaluating computer vision and deep learning approaches for solar panel detection, instance segmentation, rooftop solar mapping, and geospatial energy analysis. This dataset is intended to support reproducible research and facilitate future applications in photovoltaic infrastructure mapping, automated solar inventory generation, urban renewable energy assessment, and large-scale geospatial analysis.

Links and sources

Comments