Automated Building Structural Parameters Extraction for Seismic Risk Assessment in Villa El Salvador Area

Seismic risk assessment is essential for reducing earthquake-related damage in urban areas. Accurate recognition of building features is a key factor in evaluating seismic vulnerability, yet traditional manual inspection methods are inefficient and prone to error. This study proposes a deep learning-based framework using convolutional neural networks for automated instance segmentation of building features to support seismic risk estimation in Lima, Peru, a seismically active region with diverse architectural styles. Leveraging state-of-the-art models like Mask R-CNN, the framework identifies and segments structural components such as walls, windows, and columns from building imagery. By integrating geospatial data and remote sensing technologies, the proposed approach enhances seismic risk evaluations through automated, scalable, and precise feature recognition. Despite challenges such as occlusions, varying lighting conditions, and architectural diversity, the model aims to adapt through specialized training tailored to Lima’s urban landscape, contributing to more efficient disaster preparedness and response in earthquake-prone regions.