HKU's AI Tool eCheckGo Revolutionizes Aging Building Inspections in Seconds

The University of Hong Kong (HKU) has developed an AI-powered system to tackle the growing challenge of maintaining aging buildings in crowded cities. Named eCheckGo, the tool uses advanced technology to speed up inspections and cut costs significantly. Its creators aim to replace slow, manual checks with faster and more reliable automated assessments. The eCheckGo system combines a Large Defect Model (LdM) with traditional AI to analyse visual and textual data. By processing dozens of images in seconds, it reduces labour costs by eight times compared to other automated solutions. The model interprets defects more accurately by training on large datasets that pair images with descriptive text prompts.

A mobile app allows inspectors to capture images directly on-site or pull from public sources like Google Street View. The system then generates interactive 3D point clouds, helping users visualise damage and assess its severity. This approach replaces time-consuming manual inspections with a streamlined digital process. In a pilot study, eCheckGo assessed 9,172 buildings in Kowloon in under four hours. The test confirmed its speed, reliability, and potential for large-scale use. The research team now plans to expand its capabilities to detect more types of structural distress and produce automated reports.

The system’s success in Kowloon demonstrates its ability to handle real-world demands efficiently. By cutting inspection times and costs, eCheckGo could become a key tool for maintaining urban infrastructure. Future updates may further broaden its use in building safety assessments.