Machine-Learning-Enhanced Procedural Modeling for 4D Historical Cities Reconstruction

A novel framework integrating procedural modeling and machine learning within GIS enables the generation of detailed 3D models for historical cities, addressing gaps in incomplete datasets. This approach, demonstrated using footprints of Jerusalem's Old City (1840–1940), produces LOD 2.1 models directly from GIS files while ensuring interoperability through CityJSON format. The open-source Python module streamlines the process, eliminating reliance on specialized software and supporting dynamic updates with new data. This methodology significantly enhances accessibility, accuracy, and usability of historical city reconstructions for documentation and educational purposes.

Vaienti, B., Petitpierre, R., di Lenardo, I., & Kaplan, F. (2023). Machine-Learning-Enhanced Procedural Modeling for 4D Historical Cities Reconstruction. Remote Sensing, 15(13), 3352. https://doi.org/10.3390/rs15133352