HKU's eCheckGo AI System Cuts Building Inspection Time From Days to Hours
The University of Hong Kong has developed an AI system that analyzes dozens of building images in seconds, performing at least 100 times faster than manual inspections and eight times more cost-effectively than other automated solutions. The system, called eCheckGo, uses a proprietary Large Defect Model trained on internet-scale datasets to identify structural problems like cracks and concrete spalling.
The technology won a Gold Medal with Congratulations of the Jury at the 51st International Exhibition of Inventions in Geneva.
Why Speed Matters for Hong Kong's Aging Building Stock
Hong Kong faces a pressing maintenance problem. The government estimated 8,700 privately owned buildings were 50 years or older at the end of 2020. By 2030, that number is expected to reach nearly 14,000.
Traditional building inspections take several days to complete. Delays in identifying structural defects create public safety risks and drive up repair costs when problems worsen.
How eCheckGo Works
Users capture images inside and outside buildings using a mobile app or pull from Google Street View. The AI system automatically detects defects and integrates findings into a 3D datapoint cloud model. The interactive visualization lets inspectors zoom in to view exact geometry and dimensions of issues.
Professor Junjie Chen from the Department of Real Estate and Construction said the system's strength lies in consolidating building condition data in one place. "You can easily understand the overall condition of a building, pinpoint where defects are located, and assess their scale, geometry, and dimensions," he said.
City-Scale Testing Shows Practical Potential
The research team tested eCheckGo on 9,172 buildings in Kowloon using Google Street View images. The system generated a color-coded defect map in four hours, grading buildings from 0 (healthy) to 10 (dangerous). Results were verified against professional building surveyor inspections.
Government bodies and industry groups have begun discussions about adopting the system. The team is currently working to expand eCheckGo's capabilities to detect water leakage and dampness, and to generate automated reports in professional formats.
Learn more about AI for Real Estate & Construction or explore AI Agents & Automation tools transforming how professionals work.
Your membership also unlocks:
