New AI tool aims to improve building material assessment

We think of cars and factories as major contributors to carbon emissions, but buildings make up a large chunk of those emissions.

A collaborative research initiative led by Peking University and the University of Southern Denmark has developed an advanced framework integrating deep learning with remote sensing to identify building materials. The technology has the potential to reshape how cities approach sustainability, particularly in densely populated urban centers like New York.

Credit: Environmental Science and Ecotechnology. Illustration of geospatial and visual data fusion for building material classification. The figure integrates geospatial and visual data to effectively classify building materials. A. An aerial view of the target building outlined by a white dashed rectangle, providing context within its surrounding environment. B. The roof of the building, using satellite imagery to analyze roofing materials. C-D. Front (C) and side (D) perspectives of the building’s façade, captured by Google Street View for classifying wall materials. These images, retrieved using precise geographic coordinates from OpenStreetMap, ensure accurate alignment and play a crucial role in comprehensively analyzing the building materials.

The research addresses a critical challenge: the significant contribution of the construction sector to global carbon emissions. “The construction sector is a major driver of global carbon emissions, with buildings alone responsible for nearly one-third of worldwide energy-related CO2 emissions,” the study notes. Existing methods for assessing building materials often lack the necessary scope, scalability, and accuracy to assess diverse urban landscapes comprehensively.

The new framework accurately classifies building materials using Google Street View imagery, satellite data, and OpenStreetMap geospatial information. The researchers trained convolutional neural networks (CNNs) to identify roof and façade materials with exceptional detail. The models were first trained using extensive datasets from Odense, Denmark, before being validated in major Danish cities such as Copenhagen, Aarhus, and Aalborg.

The implications for cities like New York, with its complex tapestry of old and new buildings, are significant. Assessing the material composition and environmental impact of such a vast and varied building stock is a resource-intensive undertaking. This technology could potentially provide a more efficient and cost-effective means of gathering this data.

“Conventional databases struggle to provide comprehensive material intensity assessments, especially across diverse urban landscapes,” the study highlights. The AI tool could help overcome these limitations, offering a more granular understanding of the materials used in different neighborhoods and building types.

A key innovation of the study is its use of advanced visualization techniques, including Gradient-weighted Class Activation Mapping (Grad-CAM), which offers a window into how the AI models interpret imagery. “This technique enhances model transparency and reliability by revealing which parts of an image most influence classification decisions,” according to the study. This transparency is crucial for building trust in the technology and ensuring its responsible application.

Potential impacts and future considerations

If this technology is accurate and scalable, the potential impact on cities like New York could be substantial:

• Improved building energy efficiency: Data on building materials can inform retrofitting programs to improve energy efficiency and reduce carbon emissions.
• Informed policy decisions: City planners could use the information to develop targeted policies that incentivize using sustainable materials and promote green building practices.
• Enhanced urban sustainability planning: The technology could support the development of more comprehensive urban sustainability plans that consider the environmental impact of the city’s built environment.

However, potential challenges and considerations need to be addressed:

• Data accuracy and bias: The accuracy of the AI model is dependent on the quality and representativeness of the data used to train it. It is important to ensure that the data reflects the diversity of building materials and construction practices found in cities like New York to avoid biased results.
• Privacy concerns: The use of Google Street View imagery and other publicly available data raises potential privacy concerns that need to be carefully considered and addressed.
• Scalability and implementation: The technology must be scalable and adaptable to the unique challenges of assessing building materials in large, complex urban environments.
• Transparency and public engagement: Open communication with the public about how the technology is being used and its potential impacts is essential for building trust and ensuring its responsible implementation.

Next steps

The researchers are likely working to refine the AI model, expand its dataset, and validate its accuracy in a wider range of urban settings. Future research could explore integrating this technology with other urban data sources to create a more comprehensive picture of the city’s built environment.

Overall, this research holds promise for improving how cities assess building materials and promote sustainability. As the technology continues to evolve, it is important to carefully consider its potential benefits and challenges to ensure its responsible and equitable application.