Artificial intelligence (AI) is delivering major advances in the construction engineering sector in this era of building information modelling, applying data collection techniques based on urban image analysis. In this study, building heights were calculated from street-view imagery based on a semantic segmentation machine learning model. The model has a fully convolutional architecture and is based on the HRNet encoder and ResNexts depth separable convolutions, achieving fast runtime and state-of-the-art results on standard semantic segmentation tasks. Average building heights on a pilot German street were satisfactorily estimated with a maximum error of 3 m. Further research alternatives are discussed, as well as the difficulties of obtaining valuable training data to apply these models in countries with no training datasets and different urban conditions. This line of research contributes to the characterisation of buildings and the estimation of attributes essential for the assessment of seismic risk using automatically processed street-view imagery.

Comprehensive and accurate Geospatial Exposure Databases (GED) are fundamental for seismic risk assessment, but their creation and maintenance is a major challenge in regions with limited or non-existent cadastral data. GEDs store the characterization of seismic exposure attributes at the individual building scale spatially linked to geometric footprints, usually developed after costly manual and field procedures. This manuscript presents a novel methodology for the automated obtention of seismic behaviour modifiers from individual building footprints, implemented in the accessible open-source Python package, . The methodology employs a semi-automatic approach for footprint digitalisation. Fine-tuned instance segmentation models are used to extract initial building geometries from high-resolution aerial imagery, which can then be refined with user guidance. From these digitalised footprints, the tool automatically computes key behaviour modifiers, including building direction, relative position within an urban block (e.g., corner, confined), and plan irregularity. A key contribution is the translation of irregularity parameters from multiple international seismic codes, including Eurocode 8, ASCE-7, Mexican NTC-23, and GNDT II Livello, into robust geometric algorithms. This allows for the objective and automatic calculation of indices related to slenderness, eccentricity, and setbacks, that before depended on subjective expert assessment. The methodology was trained and validated using detailed manually annotated field inventories from three pilot study areas in San José, Santo Domingo, and Guatemala City, selected for their representative construction systems and broad diversity of attributes within the three countries under study. In these pilot areas, the error of the automated analysis was found to be comparable to the variability among human surveyors, confirming its objectivity and consistency. The resulting workflow enables the rapid and automatic, large-scale generation of detailed exposure inventories, providing a critical resource for seismic vulnerability studies, urban planning, and disaster risk management, particularly in data-scarce environments.