Satellite and digital image processing (SDIP) framework for enhancing community building energy model (CBEM) inputs

Urban building energy modelling has become essential for understanding and optimising building energy consumption at the city level. It involves creating virtual representations of buildings to simulate and analyse their energy performance. However, obtaining key parameters, including building footprints, heights, and window-to-wall ratios, is crucial for an accurate energy assessment, which is typically collected from conventional design drawings or site investigations. Acquiring input from such sources becomes difficult in regions, such as developing countries, where there is no regulated digital database. In response to the above limitations, the proposed framework couples AI-assisted geospatial and computer vision techniques to extract critical inputs for building energy modelling from existing satellite imagery and smartphone photography. The developed framework is deployed and validated through a comprehensive case study to determine energy use in a building community in Mumbai, India. Results show that the energy usage determined using the framework is 118.87 MWh, with a percentage error of 10.88% relative to the actual readings, demonstrating its potential to develop a scalable, automated, and precise solution for energy assessment across building communities.