Rapid Visual Screening of Buildings for Potential Seismic Hazards: Automated Deep-Learning Classification Approach

Abstract

Rapid visual screening (RVS) is a way to assess unsafe structures and reduce urban earthquake vulnerability; which is labor-intensive, time-consuming, and costly through conventional methods. Given that screening relies on visual clues, this study has utilized deep-learning convolutional neural networks including Residual Network (ResNet), Visual Geometry Group (VGG), Inception, and Densely Connected Convolutional Networks (DenseNet) to extract and classify necessary visual features to perform, enhance, and provide a novel standard computer vision–based method, complete with image databases and a computer program, to automate this procedure. This approach addresses the limitations of previous research in direct screening applications. The algorithm achieved 61%, 64%, 80%, 85%, and 96% accuracy in extracting various features, which, due to task independence, do not compromise its overall robustness. Furthermore, wielding hierarchical classification resulted in a noteworthy 18% accuracy improvement in building type classification. The proposed program produces final scores in 53 s, making it 17 times faster than conventional methods. Combined with its compatibility with portable computers, this will significantly enhance the speed and efficiency of onsite procedures, drastically reducing labor costs and making large-scale projects more feasible. Lastly, the severely damaged buildings classifier is exploitable for postearthquake fast screening and identification. This study introduces a novel automated deep learning–based approach for rapid visual screening of buildings to assess seismic hazards. Traditional methods of RVS are labor-intensive, time-consuming, and costly, often requiring extensive human labor and expertise. Our proposed program significantly accelerates this process, producing final scores in just 53 s, 17 times faster than conventional methods, with up to 95% accuracy. This increase in speed, combined with compatibility with even old portable computers, allows for efficient onsite procedures, drastically reducing labor costs and making large-scale projects more feasible. The automated system uses convolutional neural networks to extract and classify visual features necessary for rapid visual screening from street view images, enhancing the accuracy and reliability of the screening process. By minimizing human intervention, the program reduces subjective errors and ensures consistent results. This method not only speeds up the initial evaluation of buildings but also provides a robust tool for postearthquake fast screening and identification of severely damaged structures. The practical application of this technology promises to improve urban earthquake preparedness and resilience, facilitating quicker and more cost-effective identification of buildings that require retrofitting.