In the rolling hills and historic towns of Italy, a silent threat looms: unauthorized construction, creeping into fragile territories and endangering the integrity of cherished hamlets and vernacular heritage. Enter a new ally in the fight to preserve these cultural treasures—artificial intelligence. Researchers, led by M. Seccaroni from the Department of Architecture, Built Environment, and Construction Engineering at Politecnico di Milano, have developed an innovative method to detect unregistered buildings, combining cutting-edge AI with traditional cadastral data and high-resolution imagery.
The team’s approach leverages the Segment Anything Model (SAM), a foundation model designed for image segmentation. By prompting SAM with cadastral centroids, they extract building footprints and compare them to official records, identifying discrepancies and undocumented structures. “This method allows us to automate the detection of unauthorized buildings, making the process more efficient and scalable,” Seccaroni explains. The research, published in ‘The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences’ (translated to English as ‘The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences’), demonstrates the potential of AI to revolutionize heritage documentation and territorial monitoring.
The implications for the energy sector are significant. Unauthorized buildings can disrupt planned energy infrastructure projects, leading to costly delays and legal battles. By identifying these structures early, energy companies can mitigate risks and streamline their planning processes. Moreover, the method’s scalability means it can be deployed across vast territories, providing a comprehensive overview of potential obstacles.
The research also highlights the importance of open-source tools in driving innovation. By making their workflow reproducible and accessible, Seccaroni and his team are empowering other researchers and practitioners to build upon their work. “We believe that open-source tools are crucial for advancing the field and ensuring that these technologies are used for the benefit of all,” Seccaroni states.
As the energy sector continues to evolve, the need for accurate and efficient territorial monitoring will only grow. This research offers a glimpse into a future where AI-driven tools play a pivotal role in preserving heritage sites, ensuring regulatory compliance, and facilitating smoother project execution. By embracing these technologies, the energy sector can not only protect valuable cultural assets but also enhance its operational efficiency and effectiveness.