On March 16th, Milena Gabanelli's Dataroom made a news report about the cooperation between Arpa Lombardia and Dipartimento di Elettronica, Informazione e Bioingegneria within the Savager experimental project, developed by the regional agency to enhance its environmental monitoring activity with the introduction of Geospatial Intelligence technologies and Earth observation from satellites, aircrafts and drones.
The Dipartimento di Elettronica, Informazione e Bioingegneria has been involved in the project to experiment the automatic identification of potential situations of non-compliance to the environmental and waste legislation, based on satellite imagery and very high resolution aerial ortho-photos, using Artificial Intelligence technologies.
More in detail, DEIB researchers, coordinated by Prof. Piero Fraternali, have trained the ResNet-50 classification model, one of the most consolidated deep learning algorithm, to understand if a remote sensing image shows potentially critical sites.
The algorithm is able to identify potentially dangerous sites using remote sensing images with a precision close to 90%. The ongoing researches are aimed at identifying specific typologies of waste and evaluating the environmental risk level in the considered sites.
The Dipartimento di Elettronica, Informazione e Bioingegneria has been involved in the project to experiment the automatic identification of potential situations of non-compliance to the environmental and waste legislation, based on satellite imagery and very high resolution aerial ortho-photos, using Artificial Intelligence technologies.
More in detail, DEIB researchers, coordinated by Prof. Piero Fraternali, have trained the ResNet-50 classification model, one of the most consolidated deep learning algorithm, to understand if a remote sensing image shows potentially critical sites.
The algorithm is able to identify potentially dangerous sites using remote sensing images with a precision close to 90%. The ongoing researches are aimed at identifying specific typologies of waste and evaluating the environmental risk level in the considered sites.