A methodology for qualitative assessment of cracks in structural concrete using a combination of image processing techniques and knowledge-based models

Early identification of concrete cracks and multi-class detection can help to avoid future deformation or collapse in concrete structures. Available traditional detection and methodologies require enormous effort and time. To overcome such difficulties, current vision-based deep learning models can effectively detect and classify various concrete cracks. This study introduces a novel multi-stage deep learning framework for crack detection and type classification. First, the recently developed YOLOV10 model is trained to detect possible defective regions in concrete images. After that, a modified vision transformer (ViT) model is trained to classify concrete images into three main types: normal, simple cracks, and multi-branched cracks. The evaluation process includes feeding concrete test images into the trained YOLOV10 model, identifying the possible defect regions, and finally delivering the detected regions into the trained ViT model, which decides the appropriate crack type of those detected regions. Experiments are conducted using the individual ViT model and the proposed multi-stage framework. To improve the generation ability, multi-source datasets of concrete structures are used. For the classification part, a concrete crack dataset consisting of 12,000 images of three classes is utilized, while for the detection part, a dataset composed of various materials from historical buildings containing 1116 concrete images with their corresponding bounding boxes, is utilized. Results prove that the proposed multi-stage model accurately classifies crack types with 90.67% precision, 90.03% recall, and 90.34% F1-score. The results also show that the proposed model outperforms the individual classification model by 10.9%, 19.99%, and 19.2% for precision, recall, and F1-score, respectively. The proposed multi-stage YOLOV10-ViT model can be integrated into the construction systems which are based on crack materials to obtain early warning of possible future deformation in concrete structures.

Stone monuments stand as enduring testaments to human history and cultural heritage, yet they are susceptible to deterioration over time. In this paper, we propose a comprehensive approach for the automated detection and classification of cracks in ancient monuments, integrating machine learning and advanced image processing techniques. Our method addresses the pressing need for efficient and objective assessment of structural integrity in these invaluable artifacts. The proposed algorithm begins with preprocessing steps, including image enhancement using adaptive histogram equalization to improve crack visibility. Subsequently, feature extraction techniques such as Grey Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP) are applied to capture essential characteristics of crack patterns. Central to our approach are the Back Propagation Neural Network (BPNN) and Improved Support Vector Machine (ISVM) classifiers, which are trained on the extracted features to detect and classify cracks with high accuracy. The BPNN learns complex relationships between input features and crack types, while the ISVM leverages a margin-based approach for robust classification. Through extensive experimentation on a diverse dataset of ancient monuments, we demonstrate the effectiveness of our approach in accurately identifying and categorizing cracks. The proposed method offers a scalable and objective solution for monitoring the structural health of ancient monuments, contributing to proactive conservation efforts and the preservation of cultural heritage.

Pavement cracking is a common road infrastructure issue which significantly affects road performance, safety and longevity. This article employed a Beamlet Transform algorithm to detect and classify different types of flexible asphalt concrete pavement cracks. Additionally, a dedicated crack segmentation network was employed for precise segmentation of pavement crack. This approach incorporates advancements that has improve precision in crack classification and segmentation. Based on the results of the beamlet transform, significant improvements in the gray scale representation of crack, enhanced crack detection, reduced noise in crack images and a more precise measurement of cracks length were achieved. Computations were performed to determine the length of linear cracks and the area of block cracks. A total of 1000 pavement images were used for training and testing the accuracy of asphalt pavement crack detection and classification models. The research results showed that block cracking, alligator cracking, transverse cracking, and longitudinal cracking can all be recognized with a remarkable accuracy. Alligator cracks and block cracks achieved detection rates more than 90%, while detection rates for the longitudinal and transverse cracks reached more than 95% accuracy.

The existing methods for pavement crack classification and identification solely offer information about the crack type, neglecting size and direction details, which are essential for guiding repair efforts and forming the engineer digital information data. In response to the challenges posed by insufficient crack information, prolonged training time and intricate parameter adjustment inherent in employing deep learning algorithms for pavement crack classification and recognition, we propose an integrated approach combining tensor voting with the random sample consensus for pavement crack classification and recognition. The method involves pre-processing road images using gray value transformation and the K-Means clustering algorithm. Subsequently, the tensor voting algorithm is applied to enhance the linear features, resulting in the generation of linear saliency maps of cracks along with crack junction information. Furthermore, a non-maximum suppression method and the RANSAC algorithm are employed to refine and fit the crack skeleton curves respectively, accomplishing the crack classification and recognition. The outcomes demonstrate that the proposed integrated approach in the crack skeleton segmentation algorithm yields an average F1-score of 0.7879, outperforming traditional non-maximum suppression methods. The accuracy of crack classification and recognition reaches 96%, outperforming other crack classification and recognition algorithms grounded in digital image processing methods. Compared with the neural networks employed for classification and recognition, the proposed algorithm is able to capture direction and size details of cracks, which can provide guidance for intelligent crack repair. This additional information can offer valuable guidance for intelligent crack repair processes.

Building cracks such as gaping cracks, separation, and horizontal cracks are a few types of cracks that possess a severe issue on reinforced concrete; hence, the earlier the detection the cheaper the repairs. Numerous studies about crack detection considered VGG16, and Faster R-CNN because the severity level is crucial to each construction company business owner. This paper aims to build a deep learning model using Yolov3 that can detect a crack in reinforce concrete structures and categorize a medium, severe, or very severe crack using an android application. An android application was developed instead of using an expensive Ultrasonic Pulse Velocity in the market to detect the severity of the crack on the concrete. The overall accuracy summary of the android application is 93.33%, while the kappa value is. 97. Therefore, the deep learning model and android application produced an accurate calculation in detecting the crack and determining its crack classification.

Railway concrete sleepers are key safety-critical components in ballasted railway tracks. Due to frequent high-intensity impact loadings from train-track interaction over irregularities together with hostile environmental conditions, complicated characteristics of various crack patterns can incur on railway concrete sleepers, which will decrease their durability and service life overtime. Early warning of those cracks can help railway engineers to plan and schedule for renewal and maintenance timely and effectively. This study thus explores the artificial intelligence application of YOLOv5OBB (YOLOv5 with Oriented Bounding Box output) in the identification and classification of cracks in railway sleepers into three distinct types: longitudinal, transverse, and inclined, based on their specific crack angles, which have not been investigated in the past. The identification of crack angles is the novelty of this study. Recognising the various types of cracks is critical, given their varying causes and degrees of severity. Current corrective maintenance methods pose considerable safety risks to workers and exhibit low efficiency, underscoring the need for a more autonomous and efficient solution. This study marks a significant stride towards revolutionising railway maintenance, evidenced by an impressive mAP (Mean Average Precision) of 0.72 for crack detection and a 92% accuracy rate for angle detection. These promising results substantiate our study's potential to pioneer advancements in railway infrastructure maintenance.

A probabilistic analysis of acoustic emission events and associated energy release during formation of shear and tensile cracks in cementitious materials under uniaxial compression

Formation of cracks is a major cause of damage in reinforced concrete (RC) structures which adversely affects the durability along-with appearance of structures. Hence, to understand the behavior of RC structural element under different stresses, it is necessary to obtain significant data describing the characteristics of fracture such as types of cracks, its propagation, failure pattern etc. which may help in control and prevention of failure in RC structures. Acoustic emission (AE) technique is one of the effective non-destructive techniques used for identification and characterization of cracks. The fracture mode of cracking in concrete normally changes from tensile mode to shear mode at impending failure which can be identified by AE data analysis. The aim of present work is to identify and characterize the fracture in plain as well as RC beams using parameter based AE analysis. The experimental work investigated failure of plain and RC beams with and without notch subjected to flexural loading using AE technique. The results show dissimilarity in failure patterns of these beams which is well identified and analyzed by parametric analysis of AE technique. Thus, it is concluded that, continuous AE monitoring is an effective tool for diagnosis and characterization of cracks in concrete.

The detection and repair of the cracks in the road pavement is a very time consuming task which should be performed periodically in order to maintain the safety and quality of the road network. There are various types of road pavement cracks and each type requires different management and repair method and also each type indicates a different problem in that section of the road. In this paper, an autonomous machine learning based visual inspection system for detection and classification of the road pavement cracks is proposed. The proposed framework uses deep neural networks in order to detect and classify longitudinal, alligator and asphalt cracks. A dataset of images from different road conditions and various pavement cracks is collected. The proposed framework increases the speed and scale of road pavement analysis and repair and can be used for smart road maintenance management in the smart cities. The experimental results show that the accuracy of the proposed framework is 95% for detection and classification of the cracks in the road pavements.

Concrete cracks pose significant threats to concrete structures, causing immediate strength loss and leading to gradual erosion that compromises structural integrity. Therefore, accurate and automatic detection and classification of concrete cracks, along with the evaluation of their effects on target structures, are critically important. This study focuses on the No. 3 Huaiyin pumping station, a large-scale hydraulic structure on the Eastern Route of the South-to-North Water Diversion Project in Jiangsu, China. First, relevant field test literature is reviewed, and the finite element method is applied to investigate the effects of an existing crack on the No. 2 supporting wall. Using thermomechanically coupled numerical simulations, the distribution of tensile stress in the supporting wall is reported in two cases: without a crack and with an existing crack. The findings indicate that the increase in tensile stress due to the existing crack is relatively small and can be considered negligible for the No. 2 supporting wall. Next, the pretrained YOLOX network for the detection and classification of three types of cracks is proposed and retrained using collected concrete crack datasets. The mean average precision of the retrained YOLOX network for all three types of cracks reaches 80%. Finally, the retrained YOLOX network is applied to detect and classify cracks at the No. 3 Huaiyin pumping station. This automatic detection and classification approach will enhance the high-quality management of the pumping station because it is labor-saving and easy to deploy.

Pattern recognition of concrete surface crack defects is very important in determining stability of structure like building, roads or bridges. Surface crack is one of the subjects in inspection, diagnosis, and maintenance as well as life prediction for the safety of the structures. Traditionally determining defects and cracks on concrete surfaces are done manually by inspection. Moreover, any internal defects on the concrete would require destructive testing for detection. The researchers created an automated surface crack detection for concrete using image processing techniques including Hough transform, LoG weighted, Dilation, Grayscale, Canny Edge Detection and Haar Wavelet Transform. An automatic surface crack detection robot is designed to capture the concrete surface by sectoring method. Surface crack classification was done with the use of Haar trained cascade object detector that uses both positive samples and negative samples which proved that it is possible to effectively identify the surface crack defects.

MultiScaleCrackNet: A parallel multiscale deep CNN architecture for concrete crack classification

Crack identification plays an essential role in the health diagnosis of various concrete structures. Among different intelligent algorithms, the convolutional neural networks (CNNs) has been demonstrated as a promising tool capable of efficiently identifying the existence and evolution of concrete cracks by adaptively recognizing crack features from a large amount of concrete surface images. However, the accuracy as well as the versatility of conventional CNNs in crack identification is largely limited, due to the influence of noise contained in the background of the concrete surface images. The noise originates from highly diverse sources, such as light spots, blurs, surface roughness/wear/stains. With the aim of enhancing the accuracy, noise immunity, and versatility of CNN-based crack identification methods, a framework of enhanced intelligent identification of concrete cracks is established in this study, based on a hybrid utilization of conventional CNNs with a multi-layered image preprocessing strategy (MLP), of which the key components are homomorphic filtering and the Otsu thresholding method. Relying on the comparison and fine-tuning of classic CNN structures, networks for detection of crack position and identification of crack type are built, trained, and tested, based on a dataset composed of a large number of concrete crack images. The effectiveness and efficiency of the proposed framework involving the MLP and the CNN in crack identification are examined by comparative studies, with and without the implementation of the MLP strategy. Crack identification accuracy subject to different sources and levels of noise influence is investigated.

Because of the great variety of possible types of cracks and causes for their formation, the problem of safeguarding against crack formation in welding operations is a highly complex one. A comprehensive survey of the cracking problems associated with the welding of different steel grades, is given by Baker [282] with special emphasis given to the different crack phenomena and the various causes of their formation. A survey of the different types of cracks found and a basic diagram of the temperature for possible crack formation in the welding of steel is also included in DIN 8524, Part 3. Because of their intricate appearance, hot cracks in austenitic welded joints are of particular importance and the literature available on this topic is quite substantial. Borland and Younger [283] compiled a survey of 162 publications up to the year ending 1959 regarding the hot cracking phenomena during the welding of austenitic chromium-nickel steels. In the documentation „Schweistechnik, Bibliographie zum Thema Heisrisse beim Schweisen“ [284] which covers the years from 1968 to 1978, 81 out of a total of205 papers deal with austenitic materials. Another survey and a classification of hot cracks as they occur in the welding of various steels is supplied by Hemsworth, Boniszewski and Eaton [285]. A comprehensive survey on the present state of knowledge of heat affected zone (HAZ) cracking in thick sections of austenitic stainless steels is given with 91 references by Thomas Jr. [449].

Problem statement. The relevant issue of modern Civil Engineering is the development of new methods for identifying defects in building structures, which would save human resources and reduce the dependence of survey results on subjective human factors. The purpose of the research is to develop artificial neural networks for the identification and classification of cracks in vertical elements of building structures (e. g., concrete and reinforced concrete walls). Methodology. The cloud tool Teachable Machine is used to develop a neural network with a pre-defined internal architecture. The libraries of TensorFlow software platform allow us to develop a convolutional neural network with a tunable architecture. The program code is written in Python. The training was performed using the cloud environment Colab. Scientific novelty. New models of artificial neural networks for the identification of defects in building structures are developed. The rational magnitudes of the training parameters and the topology of the convolutional neural network are determined allowing to achieve the highest accuracy and the lowest losses of the model. Practical value. The developed models of neural networks and the results obtained with their help ensures a high efficiency of the artificial intelligence to solve problems of the health monitoring of building structures. Unlike traditional approaches, the proposed models allow real-time automatic diagnostics by analyzing photographic images, images from computer, smartphone or quadcopter webcams. The latter makes it possible to inspect buildings without the physical presence of humans at the site, which is especially important for working in dangerous places, such as tall buildings, partially destroyed buildings, mined areas, etc. Conclusions. The proposed methods can be further extended for the monitoring and classification of a wide range of defects in building structures.