The AIDETECT project, carried out in partnership with the company Techonis, has as its main goal the development of a remote inspection methodology for the automatic identification and mapping of anomalies in concrete bridges, using Deep Learning algorithms and georeferenced three-dimensional models from multiple types of sensors.

By integrating digital technologies and artificial intelligence, the AIDETECT project aims to establish itself as an innovative and robust approach to structural inspection, promoting more objective, reproducible, and adaptable methodologies to meet current infrastructure requirements. In this way, it seeks to increase the accuracy, safety, and speed of bridge inspections, making the process less dependent on specialized equipment and ensuring increasingly objective results.

The project methodology can be summarized in three fundamental stages:

1. Automatic identification of cracks, exposed reinforcement, and efflorescence, using models such as Convolutional Neural Networks and Vision Transformers;
2. 3D reconstruction of models through multiple digital technologies such as photogrammetry (supported by UAVs) and terrestrial and aerial Laser Scanning;
3. Three-dimensional mapping of the identified anomalies using projection techniques such as Ray Casting.