Remote Sensing Based Approach for Photovoltaic Power Capacity Estimation Using High-Resolution C-2S/3 Imagery

Worldwide emphasis on renewable energy has increased enormously, to limit climate change. India ranks 4<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> worldwide for solar-power production. It’s an important component of the Indian National-Infrastructure-Pipeline (NIP), complemented by the PM-GatiShakti-Plan. There is an urgent need to monitor the Solar-power-plantdevelopment and solar-energy-generation automatically using High-Resolution (HR) satellite data. An integrated-remote-sensing approach is presented for estimation of solar-energy-generation by Solar/ Photovoltaic (PV) power plants. A novel method of implementing Dimensionality-Expansion (DE) of spectral bands which are combined with textural information with Artificial-Neural-Network based Self-Organizing-Map (SOM) classification is developed for automatic-detection of solar-panels using Cartosat-2S/3-HR-datasets. This technique doesn’t need any explicit model-training and pre-labelling of samples, being unsupervised in nature. Estimation of ‘technical-potential of energy generation’ is derived from the segmented PV-panels surface area and global solar irradiation. The results are crosschecked with ‘open-source’ energy capacity, which indicates that a reliable solar energy estimation on open-area as well as on rooftop of large-buildings can be provided.