Researchers have developed AI to determine the energy efficiency of buildings from a photograph or satellite image

Researchers from the University of Cambridge have created an AI algorithm that can assess the energy efficiency of any building from a single photograph. This could speed up the identification of energy-inefficient homes that are a source of greenhouse gas emissions.

Here's What We Know

The algorithm, trained on public data, identifies "hard-to-decarbonise" homes with 90% accuracy. Such buildings are difficult to retrofit due to age, construction or location.

The model identifies which elements of the house - roof, windows - lose the most heat. It also determines whether the building is old or new. In the future, the researchers plan to increase the detail and accuracy of the AI.

Houses that are difficult to decarbonise are shown in red, while those that are more energy efficient are shown in blue.

This is the first time AI has been trained to identify difficult-to-decarbonise homes from open data, the authors said. Such a model could help authorities identify priority sites to reduce carbon emissions.

The researchers are now working to improve the AI. They will add data on energy consumption, residents' income levels and thermal imagery of facades. This will increase the accuracy and detail of the analyses.

According to the researchers, AI can extract valuable insights from big data to solve complex environmental problems.

Source: TNW