LIMAP: Computer Vision for Site Deterioration Detection

The Challenge

Heritage conservation in Malta requires regular condition assessments of historic buildings and structures. Traditionally, architects and conservators manually photograph sites, visually identify areas of deterioration such as cracks, erosion, and staining, and then painstakingly map these onto technical drawings. This process is time-consuming, subjective, and difficult to standardize across different assessors.

Our Approach

Neural AI developed LIMAP, a custom computer vision solution trained on real deterioration cases from Maltese heritage sites. The system works with standard photographs — no specialized equipment is required — and automatically detects multiple types of surface decay including cracks, biological growth, salt crystallization, and material loss.

Key technical capabilities include:

• AI model trained on real Maltese heritage deterioration data for high accuracy in local conditions
• Automated detection of surface decay categories from standard digital photographs
• Direct overlay onto AutoCAD drawings preserving architectural scale and precision
• Generation of visual condition reports for architects, conservators, and project stakeholders
• Cloud-based, scalable architecture enabling processing of large site surveys efficiently