Tracasa Global is a leading company in the classification of point clouds obtained from aerial LiDAR sensors using Artificial Intelligence. Historically, the task of classifying point clouds to label various categories of elements (buildings, roads, trees, etc.) has been performed manually or semi-automatically. However, due to the vast amount of data, this methodology has become costly, inefficient, and, in some cases, unfeasible. Leveraging its expertise in artificial intelligence, Tracasa Global approaches this task with a supervised automatic focus to achieve a more competitive classification in terms of accuracy, efficiency, time, and cost.
In 2017, Tracasa Global conducted the first data capture in Europe using the SLP 100 (Single Photon LiDAR) sensor. The project involved capturing data from over 13,000 km² covering the entire province of Navarra (Spain) with an average density of more than 14 points per m². This massive amount of information (580 billion points) was impossible to process and classify using “traditional” methods, leading to the development of the first Machine Learning models. This project was executed for the Cartography Service of the Government of Navarra.
In 2020, Tracasa Global captured data from the metropolitan area of Pamplona (approximately 50 km²) using the CityMapper 2 hybrid sensor equipped with a LiDAR sensor, 4 oblique cameras, and a nadir camera. The resulting point cloud had a density of 50 points per m² and was classified using Machine Learning techniques. This work was funded by the Cartography Service of the Government of Navarra.
This allowed Tracasa Global’s Innovation and Production teams to continue advancing in this field of knowledge. In 2022, Deep Learning techniques were introduced into the classification models for the pilot project of classifying LiDAR data from the third coverage of the PNOA (National Aerial Orthophotography Plan) for the National Geographic Institute of Spain. This project was carried out in the community of Castilla y León and aimed to fine-tune the tools for classifying the third LiDAR coverage of the PNOA with a density of 5 points per m². It was a more ambitious project that sought to classify ground, low vegetation, medium/high vegetation, buildings, and a series of minor classes such as power towers, vehicles, power lines, and bridges. Additionally, this project confirmed that this technology allows for the massive classification of point clouds with AI with very good precision results.
In 2024, Tracasa Global was awarded a contract by the Ministry of Transport and Sustainable Mobility to classify LiDAR point clouds from the third LiDAR coverage of the PNOA, covering more than 20,000 km² of Aragón, Comunidad Valenciana, and Región de Murcia. This project has a more demanding legend, proposing the classification of 12 classes: ground, low vegetation, medium/high vegetation, buildings, bridges, railways, roads, vehicles, cars, transmission lines, power towers, and wind turbines.
This contract allows Tracasa Global to showcase the knowledge and specialization of two areas of the company: Territorial and Spatial Engineering and R&D+i, which in recent years has consolidated as a leading and internationally recognized team, with excellent results in various competitions organized in the field of artificial intelligence and land management.
