Vol.:(0123456789) 1 3 Architecture, Structures and Construction https://doi.org/10.1007/s44150-022-00060-x ORIGINAL PAPER Two‑stage method based on the you only look once framework and image segmentation for crack detection in concrete structures Mayank Mishra 1  · Vipul Jain 1  · Saurabh Kumar Singh 2  · Damodar Maity 2 Received: 29 September 2021 / Accepted: 13 June 2022 © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022 Abstract Detecting the presence of cracks and identifying their severity are crucial tasks for determining the structural health of a concrete building. In this study, we develop a two-stage automated method based on the You Only Look Once (YOLOv5) deep learning framework for the identifcation, localization, and quantifcation of cracks in the concrete structures. In the frst stage, cracks are identifed and localized using bounding boxes, while in the second stage, the length of cracks and, therefore, the damage severity are determined. The developed deep learning model is trained using 4500 annotated images from a total of 40000 images of size 227 × 227 pixel, which are obtained from an open-source dataset collected at various campus buildings of Middle East Technical University (METU). The concept of transfer learning (i.e., pre-trained weights) is used for the training, which drastically reduces the training time. The detection and localization accuracy of this model is measured in terms of the average precision, average recall, and F1-score. The YOLOv5 model achieves the mean average precision (mAP_0.5) of 95.02%. A ResNet model is also developed just to capture the supremacy of the YOLOv5 model. The proposed method can help in identifying structural anomalies through real-time monitoring that must be urgently repaired and thus can be used in high-quality civil infrastructure monitoring systems. Keywords Deep learning · Structural health monitoring · Convolutional neural networks · Crack detection · Damage detection · Computer vision Introduction Crack formation and crack propagation through concrete can lead to signifcant defects in civil structures. Cracks have a substantial efect on the durability of reinforced con- crete because they provide pathways for aggressive agents to reach the steel reinforcement and trigger corrosion [1]. This corrosion phenomenon thus reduces the load-carrying capacity of reinforced concrete members [2]. The develop- ment of cracks in concrete reduces the efective surface area of load-bearing concrete structures and frequently leads to structural failure over time [3]. Therefore, crack identif- cation is critical when conducting structural maintenance and inspection and determining structural health [4]. When cracks are identifed, methods such as thermal scanning and laser scanning can be used to obtain information regarding the extent of cracking. Conventional approaches for detecting surface cracks on structures rely on the expertise and experiences of profes- sional inspectors. Although such approaches are efcient, their outcomes difer according to the expertise of the pro- fessional performing them. To overcome the drawbacks of the aforementioned methods, various image processing techniques (IPTs) have been developed to detect cracks on concrete surfaces, such as those of tunnels [5], pipes [6], masonry structures [7], pavements [8], bridges [9], rails [10], concrete walls [11–14], and beams [15]. IPTs are primarily used to detect cracks from images and measure the width * Mayank Mishra mayank@iitbbs.ac.in; mayank_mishra@outlook.in Vipul Jain vj10@iitbbs.ac.in Saurabh Kumar Singh saurabhksingh@iitkgp.ac.in Damodar Maity dmaity@civil.iitkgp.ac.in 1 School of Infrastructure, Indian Institute of Technology Bhubaneswar, Argul, Khordha, Odisha 752050, India 2 Department of Civil Engineering, Indian Institute of Technology Kharagpur, Kharagpur, West Bengal 721302, India