Browsing by Author "Yavariabdi, Amir"

Now showing 1 - 2 of 2

Citation - WoS: 3
Citation - Scopus: 3
Adjacent-Net: Deep Learning Classification of Adjacent Buildings for Assessing Pounding Effects Using Building Facade Images in Earthquake-Prone Regions
(Elsevier Science inc, 2025) Ekici, M. Yusa; Yavariabdi, Amir; Dogan, Gamze; Arslan, M. Hakan
In earthquake-prone areas, it is extremely important to carry out risk analyses of existing buildings and to take proactive measures in advance of potential earthquakes. Despite the availability of Rapid Seismic Assessment Methods (RSAMs), prioritising the seismic risk of buildings is a significant challenge due to the large number of residential buildings in the building stock. In RSAMs, many factors are taken into consideration to determine the earthquake risk priority. While specific construction conditions determine the risk parameters for the considered structures, one of them is the possible pounding effects (collision) of adjacent buildings. The fact that RSAMs have many evaluation parameters makes it difficult in site survey for technical experts to make decisions in some cases. Therefore, it is very important to perform these operations with software support. Based on this motivation, this study aims to perform pre-earthquake risk analysis of residential reinforced concrete buildings by assisting expert engineers (or facilitating the decision-making process in the absence of technical expertise) and to estimate the adjacent building parameter using building facade images for risk prioritisation. To achieve these objectives, a novel deep learning Convolutional Neural Network (CNN) model, named Adjacent-Net, is designed and developed to classify building facade images into adjacent or non-adjacent categories. The performance of Adjacent-Net is compared with various state-of-the-art CNN models such as DarkNet-53, EfficientNet, Inception ResNetV2, NasNet Large, ResNet-101, ShuffleNet, SqueezeNet, VGG-19, and Xception. For evaluation purposes, a dataset comprising 6170 building facade images is collected, and the results indicate that Adjacent-Net can accurately extract building adjacency parameters from images with an accuracy rate of approximately 98 %. This underscores the potential of intelligent systems in detecting collision scenarios, assessing the seismic risk of structures, and determining critical geometric parameters of buildings.
Vision-Based Analysis of Soft Story, Short Columns, and Vertical Geometry in RC Structures
(Elsevier Science Inc, 2025) Yavariabdi, Amir; Asik, M. Fatih; Dogan, Gamze; Arslan, M. Hakan
In countries located in active seismic zones, it is very important to carry out a risk analysis of the existing building stock and to priorities buildings according to their risk status in order to implement effective and improvable measures. In this context, the earthquake risks of hundreds of thousands of buildings in cities, especially residential buildings, need to be assessed, especially in countries that experience major earthquakes, such as T & uuml;rkiye. Due to the uneconomical and time-consuming nature of detailed analyses for such a large building stock, Rapid Seismic Assessment Methods (RSAM) are very useful for prioritising at-risk structures. However, rapid assessment methods can also require the use of many technical experts, which can lead to differences in interpretation based on their knowledge, expertise and experience in the relevant field. It is very important that the assessment is as standardized as possible to be able to predict the presence of significant structural irregularities for a large building stock and to make decisions on seismic risk quickly and measurably. To achieve this, this paper proposes a two-step deep learning-based framework to automatically extract structural features such as soft stories, short columns, and standard floor windows from building facade images and to estimate the total building height. In the first step, the You Only Look Once version 5 (YOLOv5) object detection model is used to identify key architectural elements associated with seismic vulnerability. In the second step, the detected architectural elements along with the facade image are analyzed to estimate building height. The framework is trained and evaluated using a dataset of 4500 facade images collected from Google Street View (GSV). The results demonstrate the method's potential for large-scale, standardized, and rapid seismic risk assessment in urban environments.