The field of 3D city modelling has attracted a growing interest for representing the city digital twin, providing interactive visualizations of building infrastructures integrated with a wide range of data typically produced in a Smart City environment. This paper presents a method for producing a 3D city model with photorealistic rooftop textures extracted from aerial images, as well as the integration of the 3D city model into an open-source Smart City framework. The proposed solution provides a smart visualization of 3D city entities integrated with a large variety of Smart City data (coming, for instance, from IoT Devices which generate time-series data, heatmaps, geometries and shapes related to traffic flows, bus routes, cycling paths etc.). The proposed method for rooftop detection and alignment follows a deep learning approach based on U-Net architecture, and it has been validated against a manually created ground-truth of 50 buildings scattered uniformly on the covered area. The solution is implemented in the open-source Snap4City Smart City platform.
Rendering 3D City for Smart City Digital Twin / Adreani, L; Colombo, C; Fanfani, M; Nesi, P; Pantaleo, G; Pisanu, R. - ELETTRONICO. - (2022), pp. 183-185. (Intervento presentato al convegno 2022 IEEE International Conference on Smart Computing (SMARTCOMP)) [10.1109/SMARTCOMP55677.2022.00046].
Rendering 3D City for Smart City Digital Twin
Adreani, L;Colombo, C;Fanfani, M;Nesi, P;Pantaleo, G;Pisanu, R
2022
Abstract
The field of 3D city modelling has attracted a growing interest for representing the city digital twin, providing interactive visualizations of building infrastructures integrated with a wide range of data typically produced in a Smart City environment. This paper presents a method for producing a 3D city model with photorealistic rooftop textures extracted from aerial images, as well as the integration of the 3D city model into an open-source Smart City framework. The proposed solution provides a smart visualization of 3D city entities integrated with a large variety of Smart City data (coming, for instance, from IoT Devices which generate time-series data, heatmaps, geometries and shapes related to traffic flows, bus routes, cycling paths etc.). The proposed method for rooftop detection and alignment follows a deep learning approach based on U-Net architecture, and it has been validated against a manually created ground-truth of 50 buildings scattered uniformly on the covered area. The solution is implemented in the open-source Snap4City Smart City platform.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.