Fondo Europeo de Desarrollo Regional. Unión Europea. Europa impulsa nuestro crecimiento
Arctica
Tecopysa Cotesa Enclave Fundación Tecopy
Cotesa logra la detección de tejados de amianto a través de la inteligencia artificial

Our technological business unit, Centro de Observación y Teledetección Espacial S.A.U. – COTESA, has just launched an innovative proposal using remote sensing and Artificial Intelligence for the location and control of asbestos roofs, in collaboration with the Department of Cartography and Urban Information of the City Council of Madrid.

For the development of this project, the Análisis Geoespacial y Observación de laTierra (AGEO) area has generated an algorithm that allows the location of asbestos located on the roofs of buildings through the use of very high spatial resolution (VHR) satellite images. . It is a solution that makes it easier for municipalities to comply with the new regulations on waste and contaminated soil (Law 7/2022 of April 8)., This new law determines, in its fourteenth additional provision that “the municipalities will prepare a census of installations and sites with asbestos, including a calendar that plans their removal” and that “public facilities or sites with the greatest risk must be managed before 2028”.

The algorithm designed by Cotesa for locating asbestos is based on artificial intelligence, developed within an R and Python environment; It has as inputs training areas for each class to classify and satellite images of very high and high spatial resolution.
For its development, training areas have been generated and radiometric analysis of roofs has been carried out. With adequate training of the algorithm and the necessary subsequent photo-interpretation, in which the data is crossed with those of buildings prior to 2002, it is possible to detect those roofs and roofs that are likely to have asbestos in their composition.

Since there is no real inventory of asbestos, one of the main shortcomings in this case is the lack of sufficient training data to feed the models. However, this tool has already been validated with a field review in an area of more than 20 square kilometers, obtaining a detection accuracy of over 95%.

Therefore, this tool developed by Cotesa and its functionalities can become essential allies for all those administrations that wish to further develop within this field.