Data augmentation is one of the effective solutions to improve the performance of machine learning models in general and deep learning in particular. Data augmentation techniques bring different effects to each model, but very few studies have considered this issue. This study investigated the effect of five distinct data augmentation strategies on a custom-built Convolutional Neural Network (CNN) and nine pre-trained CNN models for crack detection. All ten models were initially trained on a reference dataset of unaugmented images, followed by separate experiments using the augmented datasets. The results show that the pre-trained models, especially VGG-16, EfficientNet-B7, Xception, DenseNet-201, and EfficientNet-B0, consistently achieved greater than 98% in accuracy across all augmentation techniques. Meanwhile, the custom-built CNN was very sensitive to illumination changes and noise. Image rotation and cropping have minimal negative impact and sometimes improve performance. The findings demonstrate that combining data augmentation with state-of-the-art pre-trained models offers a powerful and efficient alternative to the reliance on large-scale datasets for accurate crack detection using CNNs.
3 February 2025 in Disaster in Civil Engineering and Architecture
Early damage detection in bridges is fundamental to their continued safety, and therefore of utmost importance to the bridge managers and policy makers. The traditional visual inspection which is the common practice for bridge inspection is inefficient, time-consuming, costly, risky, subjective and require the expertise of highly qualified inspectors. Consequently, the use of unmanned aerial systems (UASs) has gained significant attention in the area of bridge inspection. Most UAVs are quite expensive, ranging between $8000-$25000 for the best categories. It however requires the use of UAV equipped with high-cost sensors and longer battery duration, and adequate man power for real-time inspection. The expense of the UAV based inspection makes it unaffordable for many African countries. More recently, 3D models are increasingly deployed for image-based damage identification in bridges and other structures. The 3D models are constructed inform of a digital twin, and allow for computerized inspection using low-cost drones. This paper presents 3D Reconstruction approach for damage inspection and condition assessment of a bridge using images from low-cost unmanned aerial vehicle (UAV). The overall approach was illustrated in form of a case study on a precast concrete bridge. The 3D Reconstructed model of the bridge was virtually inspected to detect damages such as cracks, delamination, concrete deterioration, etc. The results showed that 3D reconstruction using low-cost UAV has great potential in its applications in bridge assessment.
7 November 2024 in African Journal of Advances in Science and Technology Research
Building collapse is arguably one of the most devastating disasters faced by many African countries. Nigeria has experienced a number of building collapses that have posed serious threat to human life and properties. As such, this alarming trend calls for innovative and sustainable solutions to safeguard the built environment. This paper examines how Structural Health Monitoring (SHM) and Building Information Modeling (BIM) can be integrated as a remedy for such disasters. The data from 175 respondents including construction professionals, policy makers and other stakeholders were collected through internet-based questionnaires. Descriptive statistics were used to analyse the data, and the results showed trends in people's perceptions and awareness of various technologies. To investigate the variations in perceptions among the different stakeholder groups, an ANOVA single factor test was also performed. According to the data, 45.14 percent of respondents said that building collapses in Nigeria occur frequently and that poor construction techniques are the main reason for these failures. Additionally, the fact that 75.43% and 69.71% of them had never utilised SHM and BIM, respectively, before suggests that they have little understanding of the two technologies. However, 85.14% of respondents were sufficiently certain that integrating these technologies would lower the number of building failures since it would be feasible to identify structural issues sooner and make better constructions. Therefore, capacity building programs, better collaboration between government and industry and supportive regulations should be put in place in order to promote SHM and BIM adoption.
Missing bolts compromise the structural integrity of bolted connections in steel bridges and industrial infrastructure. Manual visual inspection remains labour-intensive, subjective, and hazardous in hard-to-reach locations. This study presents a comparative benchmarking framework for unmanned aerial vehicles (UAVs) missing-bolt patch classification using convolutional neural network (CNN), focusing on balancing accuracy and computational efficiency. A UAV-acquired dataset of bolt-centric image patches was developed to evaluate four systematic experimental schemes: (i) a custom lightweight CNN trained from scratch, (ii) the lightweight CNN integrated with Squeeze-and-Excitation (SE) attention blocks across multiple positions, (iii) nine fine-tuned state-of-the-art (SOTA) pretrained CNN backbones, and (iv) SE-enhanced versions of these pretrained models. All architectures were evaluated under a standardised experimental protocol. Results show that the proposed lightweight CNN achieves classification performance comparable to heavyweight pretrained models while requiring significantly lower computational resources. Integrating SE blocks did not improve classification performance for this localised task and, in several configurations, reduced accuracy and training stability. Pretrained transfer learning models achieved high accuracy overall, but their computational complexity limits direct deployment on edge devices and UAV platforms. Grad-CAM visual explanations confirmed that the lightweight CNN consistently focuses on relevant bolt and hole regions. The findings demonstrate that a task-specific lightweight CNN offers a practical balance between inspection reliability and deployment efficiency for automated structural monitoring.