Diogo Lavado receives REN honorable mention for thesis on machine learning


During the REN Prize ceremony, Diogo Lavado's thesis, conducted as part of his master's degree in Computer Engineering, received an honorable mention.

The awarded thesis, titled "Detection of Power Line Supporting Towers Via Interpretable Semantic Segmentation of 3D Point Clouds," aims to enhance the inspection and maintenance of power transmission networks. The innovative proposal seeks to replace traditional physical and manual inspection methods with more efficient, technologically advanced, and cost-effective approaches.

"Technological advances, particularly the use of drones, enable remote operations. The obtained result is a three-dimensional representation—referred to as a 3D point cloud—of the transmission network area. This data is later analyzed by specialists to prevent forest fires and damage to the electrical structure. The machine learning model we developed, SCENE-Net, utilizes artificial intelligence to inspect the 3D point clouds and extract the coordinates of the towers supporting the relay," explains Diogo Lavado.

Under the supervision of Professors Cláudia Soares (Department of Computer Science) and Alessandra Micheletti (UniMi), Diogo Lavado completed his master's degree at NOVA FCT and is currently pursuing further studies at the Università degli Studi di Milano.

Established in 1995, the REN Prize, one of Portugal's most long-standing scientific awards, recognizes outstanding academic work in electrical systems and, more recently, natural gas. The revamped REN Prize focuses on innovation in energy systems and includes candidates from diverse areas, such as engineering, economics, science, and information systems. The ceremony took place on November 14 in Lisbon.