Automatic Pixel-Level Crack Detection for Civil Infrastructure Using Unet++ and Deep Transfer Learning

Automatic concrete cracks detection and mapping of terrestrial laser scan data

The identification of pavement cracks is critical for ensuring road safety. Currently, manual crack detection is quite time-consuming. To address this issue, automated pavement crack-detecting technology is required. However, automatic pavement crack recognition remains challenging, owing to the intensity heterogeneity of cracks and the complexity of the backdrop, e.g., low contrast of damages and backdrop may have shadows with comparable intensity. Motivated by breakthroughs in deep learning, we present a new network architecture combining the feature pyramid with the attention mechanism (PSA-Net). In a feature pyramid, the network integrates spatial information and underlying features for crack detection. During the training process, it improves the accuracy of automatic road crack recognition by nested sample weighting to equalize the loss caused by simple and complex samples. To verify the effectiveness of the suggested technique, we used a dataset of real road cracks to test it with different crack detection methods.

Karayolu yüzeyindeki asfalt çatlakları en yaygın kaplama hasarlarından biridir. Bu çatlaklar zamanında tespit edilip bakıma alınmazsa büyüyerek trafik güvenliğini ve yoğunluğunu arttıracak boyutlara ulaşmaktadır. Bu nedenle, asfalt çatlaklarının tespitinin hızlı bir şekilde gerçekleştirilmesi çok önemlidir. Geleneksel manuel çatlak tespiti, son derece zaman alıcı, çok maliyetli ve çok çaba gerektirir. Bu nedenle, araştırmacılar odaklarını otomatik asfalt çatlaklarının tespitine yoğunlaştırdılar. Ancak, son yıllarda otomatik çatlak tespiti kapsamlı olarak araştırılmasına rağmen çatlakların farklı yoğunluğa sahip olması ve kaplama ortamının karmaşıklığı nedeniyle hala zor bir görevdir. Bu çalışmada, bu zorluğun üstesinden gelmek için konvolüsyon sinir ağı temelli bir yöntem önerildi. Önerilen yöntem, sınıflandırma ve segmentasyondaki başarısı ve hafifliği bilinen MobileNetv2 ’nin temelde kullandığı konvolüsyon ve ters artıklı blok yapılarını baz alarak geliştirildi. Yapılan deneysel testler sonucunda, önerilen yöntemin performansı literatürdeki diğer yöntemlere göre daha yüksek olduğu görülmektedir. Bu da, otomatik asfalt çatlak tespitinin daha başarılı olması anlamına gelmektedir.

Within the framework of intelligent bridge detection, a number of crack detection methods based on image processing techniques have been implemented. In this study, a combined novel approach with deep learning of a single shot multibox detector (SSD) and the eight neighborhood algorithm is proposed and applied to bridge crack image identification to provide an automatic method for crack detection. First, a large number of concrete crack images collected from the site were segmented and preprocessed for the establishment of a crack image dataset. Deep learning of the SSD algorithm was introduced on the training set to establish the detection model, where the model parameters were adjusted by the validation set. Sliding window technology was integrated to identify the cracks in the test set. The effects of the sliding window size and dataset size on the crack detection results were discussed. Moreover, the eight neighborhood algorithm was adopted for further crack detection correction. The results show that the configuration achieves good crack detection by the deep learning of the SSD algorithm with high precision and recall. The introduction of the eight neighborhood correction algorithm further improves the detection results by eliminating some misjudged results. Finally, the developed algorithm was placed into a portable device, with which cracks were effectively identified. The introduced method shows significantly better performance in crack detection, and the system installed on the portable device provides a way to broaden its application in the automatic crack detection of concrete bridges.

Automatic crack detection is a main task in a crack map generation of the existing concrete infrastructure inspection. This paper presents an automatic crack detection and classification method based on genetic algorithm (GA) to optimize the parameters of image processing techniques (IPTs). The crack detection results of concrete infrastructure surface images under various complex photometric conditions still remain noise pixels. Next, a deep convolution neural network (CNN) method is applied to classify crack candidates and non-crack candidates automatically. Moreover, the proposed method is compared with the state-of-the-art methods for crack detection. The experimental results validate the reasonable accuracy in practical application.

A novel approach of surface crack detection using super pixel segmentation

Automatic real-time crack detection using lightweight deep learning models

Automatic road crack detection can play an significant role in improving driving safety. However, a number of factors make this task challenging. Background complexity, as well as the fact that cracks are visually inhomogenous to one another also contribute to this difficulty. Furthermore, cracks in road surfaces can be easily confused with foreign objects, shadows and background textures leading to detection ambiguities. Timely detection of cracks is important for both drivers and road maintenance crews. Quaternion neural networks (QNNs) are a relatively new class of neural networks that employ quaternionvalued activations and parameters. They benefit from reduced costs in terms of hardware as they require fewer parameters. In this work, we explore the usability of QNNs for automatic road crack detection. To this end, we propose quaternionic versions of deep networks and evaluate their performance in datasets with images of road cracks. We show that the proposed models are light-weight in terms of parameter requirements while they are on par in terms of performance with real-valued networks for crack detection, with potential applications in resource constrained scenarios or in cases where few training data is available.

To evaluate the integrity of a building, many experts and engineers have evaluated the damage classification of a building based on superficial visual information through field surveys. On-site surveys are hazardous and require several years of experience and expertise. In this study, a system for detecting the presence or absence of cracks and calculating their lengths was developed using image processing technology. The accuracy of the system was examined using crack image data obtained from shear force experiments. For crack detection, a crack detection method was developed using canny edge, threshold, and HSV color detection. The detection of the presence of cracks was proposed to be coupled with image segmentation to improve detection accuracy. A method for calculating the crack length using image processing was also developed. In this study, we proposed a method to calculate cracks as straight lengths, and obtained results with 98.1% accuracy. However, for curved cracks, it was necessary to rotate or segment the image.

Automatic surface crack detection using segmentation-based deep-learning approach

Automatic crack detection in construction facilities is a challenging yet crucial task. However, existing deep learning (DL)-based semantic segmentation methods for this field are based on fully supervised learning models and pixel-level manual annotation, which are time-consuming and labor-intensive. To solve this problem, this paper proposes a novel crack semantic segmentation network using weakly supervised approach and mixed-label training strategy. Firstly, an image patch-level classifier of crack is trained to generate a coarse localization map for automatic pseudo-labeling of cracks combined with a thresholding-based method. Then, we integrated the pseudo-annotated with manual-annotated samples with a ratio of 4:1 to train the crack segmentation network with a mixed-label training strategy, in which the manual labels were assigned with a higher weight value. The experimental data on two public datasets demonstrate that our proposed method achieves a comparable accuracy with the fully supervised methods, reducing over 65% of the manual annotation workload.

Automatic edge detection of ply cracks in glass fiber composite laminates under quasi-static and fatigue loading using multi-scale Digital Image Correlation

Crack detection is a critical component of structural health monitoring for concrete infrastructure. Traditional inspection methods are often labor-intensive, subjective, and inefficient. The emergence of deep learning, particularly Convolutional Neural Networks, has introduced a revolutionary shift, enabling accurate, automated, and robust crack detection. This paper presents a comprehensive review of deep learning-based methods for concrete crack detection, synthesizing findings from recent literature. We examine three primary approaches: image classification for identifying crack presence, object detection for real-time localization using models like Faster R-CNN, YOLO and semantic segmentation for precise, pixel-level crack mapping with Fully Convolutional Networks. The pivotal role of transfer learning in overcoming data limitations is highlighted. Furthermore, we explore advanced applications such as the quantification of crack parameters like width and depth, and innovative methods for detecting internal defects. Key challenges including dataset scarcity, data imbalance, and model generalization under diverse environmental conditions are discussed. Finally, we outline promising future directions, including real-time integrated systems, unsupervised learning techniques, and the development of novel fused architectures for more efficient and comprehensive structural assessments.

Automated bridge surface crack detection and segmentation using computer vision-based deep learning model

Disks (e.g., compressor disk, turbine disk) are key components of aero-engine, which play an important role in flight safety. Over the past several decades, extensive studies have been carried out to study the modeling and detection methods of aero-engine disks with cracks. However, to the best of the authors’ knowledge, no comprehensive review has been reported yet to summarize the existing studies. In this paper, the dynamic models of disks with cracks are summarized firstly. Then, the crack detection methods are reviewed. Finally, the limitations and future trends are discussed, which can provide inspiration for future research.