Abstract
Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928
Key Words
modified U-net; random elastic deformation; Self-Attention-Self-Adaption neuron; semantic segmentation; steel crack identification
Address
(1) Key Lab of Smart Prevention and Mitigation of Civil Engineering Disasters of the Ministry of Industry and Information Technology, Harbin Institute of Technology, Harbin, 150090, China;
(2) Key Lab of Structures Dynamics Behavior and Control of the Ministry of Education, Harbin Institute of Technology, Harbin, 150090, China;
(3) School of Civil Engineering, Harbin Institute of Technology, Harbin, 150090, China.
Abstract
The cable component of cable-stayed bridges is gradually impacted by weather conditions, vehicle loads, and material corrosion. The stayed cable is a critical load-carrying part that closely affects the operational stability of a cable-stayed bridge. Damaged cables might lead to the bridge collapse due to their tension capacity reduction. Thus, it is necessary to develop structural health monitoring (SHM) techniques that accurately identify damaged cables. In this work, a combinational identification method of three efficient techniques, including statistical analysis, clustering, and neural network models, is proposed to detect the damaged cable in a cable-stayed bridge. The measured dataset from the bridge was initially preprocessed to remove the outlier channels. Then, the theory and application of each technique for damage detection were introduced. In general, the statistical approach extracts the parameters representing the damage within time series, and the clustering approach identifies the outliers from the data signals as damaged members, while the deep learning approach uses the nonlinear data dependencies in SHM for the training model. The performance of these approaches in classifying the damaged cable was assessed, and the combinational identification method was obtained using the voting ensemble. Finally, the combination method was compared with an existing outlier detection algorithm, support vector machines (SVM). The results demonstrate that the proposed method is robust and provides higher accuracy for the damaged cable detection in the cable-stayed bridge.
Key Words
anomaly detection; clustering; deep learning; LSTM; time series
Address
(1) Hyesook Son, Chanyoung Yoon, Yejin Kim, Yun Jang:
Department of Computer Engineering and Convergence Engineering for Intelligent Drone, Sejong University, 209 Neungdong-ro, Gwangjin-gu , Seoul, Republic of Korea;
(2) Linh Viet Tran, Seung-Eock Kim, Dong Joo Kim:
Department of Civil and Environmental Engineering, Sejong University, 209, Neungdong-ro, Gwangjin-gu, Seoul, Republic of Korea;
(3) Jongwoong Park:
School of Civil and Environmental Engineering, Urban Design and Studies, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul, Republic of Korea.
Abstract
Crack detection plays an important role in the maintenance and protection of steel box girder of bridges. However, since the cracks only occupy an extremely small region of the high-resolution images captured from actual conditions, the existing methods cannot deal with this kind of image effectively. To solve this problem, this paper proposed a novel three-stage method based on deep learning technology and morphology operations. The training set and test set used in this paper are composed of 360 images (4928 × 3264 pixels) in steel girder box. The first stage of the proposed model converted highresolution images into sub-images by using patch-based method and located the region of cracks by CBAM ResNet-50 model. The Recall reaches 0.95 on the test set. The second stage of our method uses the Attention U-Net model to get the accurate geometric edges of cracks based on results in the first stage. The IoU of the segmentation model implemented in this stage attains 0.48. In the third stage of the model, we remove the wrong-predicted isolated points in the predicted results through dilate operation and outlier elimination algorithm. The IoU of test set ascends to 0.70 after this stage. Ablation experiments are conducted to optimize the parameters and further promote the accuracy of the proposed method. The result shows that: (1) the best patch size of sub-images is 1024 × 1024. (2) the CBAM ResNet-50 and the Attention U-Net achieved the best results in the first and the second stage, respectively. (3) Pre-training the model of the first two stages can improve the IoU by 2.9%. In general, our method is of great significance for crack detection.
Address
(1) Shiqiao Meng, Ying Zhou, Qingzhao Kong:
State Key Laboratory of Disaster Reduction in Civil Engineering, Tongji University, Shanghai 200092, China;
(2) Zhiyuan Gao:
College of Civil Engineering, Tongji University, Shanghai 200092, China;
(3) Bin He:
College of Electronic and Information Engineering, Tongji University, Shanghai 200092, China.
Abstract
Cables are critical components of cable-stayed bridges. A structural health monitoring system provides real-time cable tension recording for cable health monitoring. However, the measurement data involve multiple sources of variability, i.e., varying environmental and operational factors, which increase the complexity of cable condition monitoring. In this study, a one-class classification method is developed for cable condition assessment using Bayesian factor analysis (FA). The singlepeaked vehicle-induced cable tension is assumed to be relevant to vehicle positions and weights. The Bayesian FA is adopted to establish the correlation model between cable tensions and vehicles. Vehicle weights are assumed to be latent variables and the influences of different transverse positions are quantified by coefficient parameters. The Bayesian theorem is employed to estimate the parameters and variables automatically, and the damage index is defined on the basis of the well-trained model. The proposed method is applied to one cable-stayed bridge for cable damage detection. Significant deviations of the damage indices of Cable SJS11 were observed, indicating a damaged condition in 2011. This study develops a novel method to evaluate the health condition of individual cable using the FA in the Bayesian framework. Only vehicle-induced cable tensions are used and there is no need to monitor the vehicles. The entire process, including the data pre-processing, model training and damage index calculation of one cable, takes only 35 s, which is highly efficient.
Abstract
The deployment of advanced structural health monitoring (SHM) systems in large-scale civil structures collects large amounts of data. Note that these data may contain multiple types of anomalies (e.g., missing, minor, outlier, etc.) caused by harsh environment, sensor faults, transfer omission and other factors. These anomalies seriously affect the evaluation of structural performance. Therefore, the effective analysis and mining of SHM data is an extremely important task. Inspired by the deep learning paradigm, this study develops a novel generative adversarial network (GAN) and convolutional neural network (CNN)-based data anomaly detection approach for SHM. The framework of the proposed approach includes three modules : (a) A three-channel input is established based on fast Fourier transform (FFT) and Gramian angular field (GAF) method; (b) A GANomaly is introduced and trained to extract features from normal samples alone for class-imbalanced problems; (c) Based on the output of GANomaly, a CNN is employed to distinguish the types of anomalies. In addition, a dataset-oriented method (i.e., multistage sampling) is adopted to obtain the optimal sampling ratios between all different samples. The proposed approach is tested with acceleration data from an SHM system of a long-span bridge. The results show that the proposed approach has a higher accuracy in detecting the multi-pattern anomalies of SHM data.
Key Words
convolutional neural network; data anomaly detection; generative adversarial network; Gramian angular field; long-span bridge; structural health monitoring
Address
(1) Gaoyang Liu, Yanbo Niu, Weijian Zhao, Yuanfeng Duan, Jiangpeng Shu:
College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, P.R. China;
(2) Gaoyang Liu:
Center for Balance Architecture, Zhejiang University, Hangzhou 310058, P.R. China;
(3) Yanbo Niu:
The Architectural Design & Research Institute of Zhejiang University Co. Ltd., Hangzhou 310058, P.R. China.
Abstract
The raw data collected by structural health monitoring (SHM) systems may suffer multiple patterns of anomalies, which pose a significant barrier for an automatic and accurate structural condition assessment. Therefore, the detection and classification of these anomalies is an essential pre-processing step for SHM systems. However, the heterogeneous data patterns scarce anomalous samples and severe class imbalance make data anomaly detection difficult. In this regard, this study proposes a convolutional neural network-based data anomaly detection method. The time and frequency domains data are transferred as images and used as the input of the neural network for training. ResNet18 is adopted as the feature extractor to avoid training with massive labelled data. In addition, the focal loss function is adopted to soften the class imbalance-induced classification bias. The effectiveness of the proposed method is validated using acceleration data collected in a long-span cable-stayed bridge. The proposed approach detects and classifies data anomalies with high accuracy.
Key Words
class imbalance; data anomaly detection; focal loss; limited labelled data; transfer learning
Address
Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, China.
Abstract
Various monitoring systems have been implemented in civil infrastructure to ensure structural safety and integrity. In long-term monitoring, these systems generate a large amount of data, where anomalies are not unusual and can pose unique challenges for structural health monitoring applications, such as system identification and damage detection. Therefore, developing efficient techniques is quite essential to recognize the anomalies in monitoring data. In this study, several machine learning techniques are explored and implemented to detect and classify various types of data anomalies. A field dataset, which consists of one month long acceleration data obtained from a long-span cable-stayed bridge in China, is employed to examine the machine learning techniques for automated data anomaly detection. These techniques include the statistic-based pattern recognition network, spectrogram-based convolutional neural network, image-based time history convolutional neural network, image-based time-frequency hybrid convolution neural network (GoogLeNet), and proposed ensemble neural network model. The ensemble model deliberately combines different machine learning models to enhance anomaly classification performance. The results show that all these techniques can successfully detect and classify six types of data anomalies (i.e., missing, minor, outlier, square, trend, drift). Moreover, both image-based time history convolutional neural network and GoogLeNet are further investigated for the capability of autonomous online anomaly classification and found to effectively classify anomalies with decent performance. As seen in comparison with accuracy, the proposed ensemble neural network model outperforms the other three machine learning techniques. This study also evaluates the proposed ensemble neural network model to a blind test dataset. As found in the results, this ensemble model is effective for data anomaly detection and applicable for the signal characteristics changing over time.
Key Words
data anomaly classification; ensemble neural network; GoogLeNet; machine learning
Address
(1) Jau-Yu Chou, Chia-Ming Chang:
Department of Civil Engineering, National Taiwan University, No.1 Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan;
(2) Yuguang Fu:
School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, 639798 Singapore;
(3) Shieh-Kung Huang:
Department of Civil Engineering, National Chung Hsing University, 145, Xingda Road, Taichung 40227, Taiwan.
Abstract
With the wider availability of sensor technology through easily affordable sensor devices, several Structural Health Monitoring (SHM) systems are deployed to monitor vital civil infrastructure. The continuous monitoring provides valuable information about the health of the structure that can help provide a decision support system for retrofits and other structural modifications. However, when the sensors are exposed to harsh environmental conditions, the data measured by the SHM systems tend to be affected by multiple anomalies caused by faulty or broken sensors. Given a deluge of high-dimensional data collected continuously over time, research into using machine learning methods to detect anomalies are a topic of great interest to the SHM community. This paper contributes to this effort by proposing a relatively new time series representation named "Shapelet Transform" in combination with a Random Forest classifier to autonomously identify anomalies in SHM data. The shapelet transform is a unique time series representation based solely on the shape of the time series data. Considering the individual characteristics unique to every anomaly, the application of this transform yields a new shape-based feature representation that can be combined with any standard machine learning algorithm to detect anomalous data with no manual intervention. For the present study, the anomaly detection framework consists of three steps: identifying unique shapes from anomalous data, using these shapes to transform the SHM data into a local-shape space and training machine learning algorithms on this transformed data to identify anomalies. The efficacy of this method is demonstrated by the identification of anomalies in acceleration data from an SHM system installed on a long-span bridge in China. The results show that multiple data anomalies in SHM data can be automatically detected with high accuracy using the proposed method.
Key Words
anomaly detection; long-span bridge; machine learning; shapelet transform; structural health monitoring; time series shapelets
Address
Nathaz Modeling Laboratory, Department of Civil & Environmental Engineering & Earth Sciences, University of Notre Dame, Notre Dame, Indiana 46556, USA.
Abstract
Stay cables play an essential role in cable-stayed bridges. Severe vibrations and/or harsh environment may result in cable failures. Therefore, an efficient structural health monitoring (SHM) solution for cable damage detection is necessary. This study proposes a data-driven method for immediately detecting cable damage from measured cable forces by recognizing pattern transition from the intact condition when damage occurs. In the proposed method, pattern recognition for cable damage detection is realized by time series classification (TSC) using a deep learning (DL) model, namely, the long short term memory fully convolutional network (LSTM-FCN). First, a TSC classifier is trained and validated using the cable forces (or cable force ratios) collected from intact stay cables, setting the segmented data series as input and the cable (or cable pair) ID as class labels. Subsequently, the classifier is tested using the data collected under possible damaged conditions. Finally, the cable or cable pair corresponding to the least classification accuracy is recommended as the most probable damaged cable or cable pair. A case study using measured cable forces from an in-service cable-stayed bridge shows that the cable with damage can be correctly identified using the proposed DL-TSC method. Compared with existing cable damage detection methods in the literature, the DL-TSC method requires minor data preprocessing and feature engineering and thus enables fast and convenient early detection in real applications.
Key Words
bridge cable; damage detection; deep learning; time series classification
Address
(1) Zhiming Zhang:
School for Engineering of Matter, Transport and Energy, Arizona State University, Tempe, AZ, USA;
(2) Jin Yan:
Palo Alto Research Center, Palo Alto, CA, USA;
(3) Liangding Li:
Department of Computer Science, University of Central Florida, Orlando, FL, USA;
(4) Hong Pan:
Department of Civil and Environmental Engineering, North Dakota State University, Fargo, ND, USA;
(5) Chuanzhi Dong:
Department of Civil, Environmental, and Construction Engineering, University of Central Florida, Orlando, FL, USA.
Abstract
Data anomalies seriously threaten the reliability of the bridge structural health monitoring system and may trigger system misjudgment. To overcome the above problem, an efficient and accurate data anomaly detection method is desiderated. Traditional anomaly detection methods extract various abnormal features as the key indicators to identify data anomalies. Then set thresholds artificially for various features to identify specific anomalies, which is the artificial experience method. However, limited by the poor generalization ability among sensors, this method often leads to high labor costs. Another approach to anomaly detection is a data-driven approach based on machine learning methods. Among these, the bidirectional long-short memory neural network (BiLSTM), as an effective classification method, excels at finding complex relationships in multivariate time series data. However, training unprocessed original signals often leads to low computation efficiency and poor convergence, for lacking appropriate feature selection. Therefore, this article combines the advantages of the two methods by proposing a deep learning method with manual experience statistical features fed into it. Experimental comparative studies illustrate that the BiLSTM model with appropriate feature input has an accuracy rate of over 87-94%. Meanwhile, this paper provides basic principles of data cleaning and discusses the typical features of various anomalies. Furthermore, the optimization strategies of the feature space selection based on artificial experience are also highlighted.
Key Words
BiLSTM; data anomaly detection; feature extraction; long-span bridge; structural health monitoring
Address
School of Civil Engineering, Key Laboratory of C&PC Structures of the Ministry of Education, Southeast University, Nanjing 210096, China.
Abstract
The effectiveness of system identification, damage detection, condition assessment and other structural analyses relies heavily on the accuracy and reliability of the measured data in structural health monitoring (SHM) systems. However, data anomalies often occur in SHM systems, leading to inaccurate and untrustworthy analysis results. Therefore, anomalies in the raw data should be detected and cleansed before further analysis. Previous studies on data anomaly detection mainly focused on just single type of data anomaly for denoising or removing outliers, meanwhile, the existing methods of detecting multiple data anomalies are usually time consuming. For these reasons, recognising multiple anomaly patterns for real-time alarm and analysis in field monitoring remains a challenge. Aiming to achieve an efficient and accurate detection for multi-type data anomalies for field SHM, this study proposes a pattern-recognition-based data anomaly detection method that mainly consists of three steps: the feature extraction from the long time-series data samples, the training of a pattern recognition neural network (PRNN) using the features and finally the detection of data anomalies. The feature extraction step remarkably reduces the time cost of the network training, making the detection process very fast. The performance of the proposed method is verified on the basis of the SHM data of two practical long-span bridges. Results indicate that the proposed method recognises multiple data anomalies with very high accuracy and low calculation cost, demonstrating its applicability in field monitoring.
Key Words
data anomaly detection; feature extraction; pattern recognition neural network; structural health monitoring
Address
School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology, 1037 Luoyu-road, Wuhan, China.
Abstract
In recent years, the industry and research communities have focused on developing autonomous crack inspection approaches, which mainly include image acquisition and crack detection. In these approaches, mobile devices such as cameras, drones or smartphones are utilized as sensing platforms to acquire structural images, and the deep learning (DL)-based methods are being developed as important crack detection approaches. However, the process of image acquisition and collection is timeconsuming, which delays the inspection. Also, the present mobile devices such as smartphones can be not only a sensing platform but also a computing platform that can be embedded with deep neural networks (DNNs) to conduct on-site crack detection. Due to the limited computing resources of mobile devices, the size of the DNNs should be reduced to improve the computational efficiency. In this study, an architecture called pruned crack recognition network (PCR-Net) was developed for the detection of structural cracks. A dataset containing 11000 images was established based on the raw images from bridge inspections. A pruning method was introduced to reduce the size of the base architecture for the optimization of the model size. Comparative studies were conducted with image processing techniques (IPTs) and other DNNs for the evaluation of the performance of the proposed PCR-Net. Furthermore, a modularly designed framework that integrated the PCR-Net was developed to realize a DL-based crack detection application for smartphones. Finally, on-site crack detection experiments were carried out to validate the performance of the developed system of smartphone-based detection of structural cracks.
Key Words
deep learning; edge computing; fully convolutional networks; structural crack detection; structural health monitoring
Address
Department of Civil Engineering, Zhejiang University, Hangzhou 310058, China.
Abstract
Identifying fine cracks in steel bridge facilities is a challenging task of structural health monitoring (SHM). This study proposed an end-to-end crack image segmentation framework based on a one-step Convolutional Neural Network (CNN) for pixel-level object recognition with high accuracy. To particularly address the challenges arising from small object detection in complex background, efforts were made in loss function selection aiming at sample imbalance and module modification in order to improve the generalization ability on complicated images. Specifically, loss functions were compared among alternatives including the Binary Cross Entropy (BCE), Focal, Tversky and Dice loss, with the last three specialized for biased sample distribution. Structural modifications with dilated convolution, Spatial Pyramid Pooling (SPP) and Feature Pyramid Network (FPN) were also performed to form a new backbone termed CrackDet. Models of various loss functions and feature extraction modules were trained on crack images and tested on full-scale images collected on steel box girders. The CNN model incorporated the classic U-Net as its backbone, and Dice loss as its loss function achieved the highest mean Intersection-over-Union (mIoU) of 0.7571 on full-scale pictures. In contrast, the best performance on cropped crack images was achieved by integrating CrackDet with Dice loss at a mIoU of 0.7670.
Key Words
CNN; crack detection; data imbalance; feature extraction; loss function
Address
Department of Civil Engineering, Monash University, VIC 3800, Australia.
Abstract
This article develops a long-term condition assessment method for stay cables in cable stayed bridges using the monitored cable tension forces under operational condition. Based on the concept of influence surface, the matched cable tension ratio of two cables located at the same side (either in the upstream side or downstream side) is theoretically proven to be related to the condition of stay cables and independent of the positions of vehicles on the bridge. A sensor grouping scheme is designed to ensure that reliable damage detection result can be obtained even when sensor fault occurs in the neighbor of the damaged cable. Cable forces measured from an in-service cable-stayed bridge in China are used to demonstrate the accuracy and effectiveness of the proposed method. Damage detection results show that the proposed approach is sensitive to the rupture of wire damage in a specific cable and is robust to environmental effects, measurement noise, sensor fault and different traffic patterns. Using the damage sensitive feature in the proposed approach, the metrics such as accuracy, precision, recall and F1 score, which are used to evaluate the performance of damage detection, are 97.97%, 95.08%, 100% and 97.48%, respectively. These results indicate that the proposed approach can reliably detect the damage in stay cables. In addition, the proposed approach is efficient and promising with applications to the field monitoring of cables in cable-stayed bridges.
Address
(1) Zhen Peng, Jun Li:
Guangzhou University-Curtin University Joint Research Centre for Structural Monitoring and Protection against Multi-Dynamic Hazards, School of Civil Engineering, Guangzhou University, Guangzhou 510006, China;
(2) Zhen Peng, Jun Li, Hong Hao:
Centre for Infrastructural Monitoring and Protection, School of Civil and Mechanical Engineering, Curtin University, WA 6102, Australia.
Abstract
Data-driven structural health monitoring (SHM) of civil infrastructure can be used to continuously assess the state of a structure, allowing preemptive safety measures to be carried out. Long-term monitoring of large-scale civil infrastructure often involves data-collection using a network of numerous sensors of various types. Malfunctioning sensors in the network are common, which can disrupt the condition assessment and even lead to false-negative indications of damage. The overwhelming size of the data collected renders manual approaches to ensure data quality intractable. The task of detecting and classifying an anomaly in the raw data is non-trivial. We propose an approach to automate this task, improving upon the previously developed technique of image-based pre-processing on one-dimensional (1D) data by enriching the features of the neural network input data with multiple channels. In particular, feature engineering is employed to convert the measured time histories into a 3-channel image comprised of (i) the time history, (ii) the spectrogram, and (iii) the probability density function representation of the signal. To demonstrate this approach, a CNN model is designed and trained on a dataset consisting of acceleration records of sensors installed on a long-span bridge, with the goal of fault detection and classification. The effect of imbalance in anomaly patterns observed is studied to better account for unseen test cases. The proposed framework achieves high overall accuracy and recall even when tested on an unseen dataset that is much larger than the samples used for training, offering a viable solution for implementation on full-scale structures where limited labeled-training data is available.
Key Words
convolutional neural network (CNN); data anomaly detection; sensor-fault identification; structural health monitoring
Address
Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, USA.
Abstract
Crack detection is essential for inspection of existing structures and crack segmentation based on deep learning is asignificant solution. However, datasets are usually one of the key issues. When building a new dataset for deep learning, laborious and time-consuming annotation of a large number of crack images is an obstacle. The aim of this study is to develop an approach that can automatically select a small portion of the most informative crack images from a large pool in order to annotate them, not to label all crack images. An active learning method with difficulty learning mechanism for crack segmentation tasks is proposed. Experiments are carried out on a crack image dataset of a steel box girder, which contains 500 images of 320
Key Words
acquisition function; active learning; crack detection; probability attention module; semantic segmentation
Address
(1) Jiangpeng Shu, Jun Li, Jiawei Zhang, Weijian Zhao, Yuanfeng Duan, Zhicheng Zhang:
College of Civil Engineering and Architecture, Zhejiang University, 310058 Hangzhou, China;
(2) Jun Li:
Center for Balance Architecture, Zhejiang University, 310058 Hangzhou, China.
Abstract
Structural health monitoring (SHM) plays a vital role in the maintenance and operation of constructions. In recent years, autonomous inspection has received considerable attention because conventional monitoring methods are inefficient and expensive to some extent. To develop autonomous inspection, a potential approach of crack identification is needed to locate defects. Therefore, this study exploits two deep learning-based segmentation models, DeepLabv3+ and Mask R-CNN, for crack segmentation because these two segmentation models can outperform other similar models on public datasets. Additionally, impacts of label quality on model performance are explored to obtain an empirical guideline on the preparation of image datasets. The influence of image cropping and label refining are also investigated, and different strategies are applied to the dataset, resulting in six alternated datasets. By conducting experiments with these datasets, the highest mean Intersection-over-Union (mIoU), 75%, is achieved by Mask R-CNN. The rise in the percentage of annotations by image cropping improves model performance while the label refining has opposite effects on the two models. As the label refining results in fewer error annotations of cracks, this modification enhances the performance of DeepLabv3+. Instead, the performance of Mask R-CNN decreases because fragmented annotations may mistake an instance as multiple instances. To sum up, both DeepLabv3+ and Mask R-CNN are capable of crack identification, and an empirical guideline on the data preparation is presented to strengthen identification successfulness via image cropping and label refining.
Key Words
crack recognition; deep learning; image segmentation; label quality
Address
(1) Shun-Hsiang Hsu:
NCREE-NTUCE Joint Artificial Intelligence Research Center, No. 200, Sec. 3, Xinhai Rd., Da'an Dist., Taipei City 106219, Taiwan (R.O.C.);
(2) Ting-Wei Chang, Chia-Ming Chang:
Department of Civil Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd., Da'an Dist., Taipei City 106319, Taiwan (R.O.C.).
Abstract
Manual inspection of steel box girders on long span bridges is time-consuming and labor-intensive. The quality of inspection relies on the subjective judgements of the inspectors. This study proposes an automated approach to detect and segment cracks in high-resolution images. An end-to-end cascaded framework is proposed to first detect the existence of cracks using a deep convolutional neural network (CNN) and then segment the crack using a modified U-Net encoder-decoder architecture. A Naive Bayes data fusion scheme is proposed to reduce the false positives and false negatives effectively. To generate the binary crack mask, first, the original images are divided into 448 × 448 overlapping image patches where these image patches are classified as cracks versus non-cracks using a deep CNN. Next, a modified U-Net is trained from scratch using only the crack patches for segmentation. A customized loss function that consists of binary cross entropy loss and the Dice loss is introduced to enhance the segmentation performance. Additionally, a Naive Bayes fusion strategy is employed to integrate the crack score maps from different overlapping crack patches and to decide whether a pixel is crack or not. Comprehensive experiments have demonstrated that the proposed approach achieves an 81.71% mean intersection over union (mIoU) score across 5 different training/test splits, which is 7.29% higher than the baseline reference implemented with the original U-Net.
Key Words
Bayesian data fusion; crack detection; deep learning; semantic segmentation; structural health monitoring
Address
(1) Wen Tang, Mohammad R. Jahanshahi:
Lyles School of Civil Engineering, Purdue University, West Lafayette, IN 47906, USA;
(2) Mohammad R. Jahanshahi:
Elmore Family School of Electrical and Computer Engineering (Courtesy), Purdue University, West Lafayette, IN 47907, USA;
(3) Rih-Teng Wu :
Department of Civil Engineering, National Taiwan University, Taipei, Taiwan.
Abstract
Structural health monitoring (SHM) plays an important role in ensuring the safety and functionality of critical civil infrastructure. In recent years, numerous researchers have conducted studies to develop computer vision and machine learning techniques for SHM purposes, offering the potential to reduce the laborious nature and improve the effectiveness of field inspections. However, high-quality vision data from various types of damaged structures is relatively difficult to obtain, because of the rare occurrence of damaged structures. The lack of data is particularly acute for fatigue crack in steel bridge girder. As a result, the lack of data for training purposes is one of the main issues that hinders wider application of these powerful techniques for SHM. To address this problem, the use of synthetic data is proposed in this article to augment real-world datasets used for training neural networks that can identify fatigue cracks in steel structures. First, random textures representing the surface of steel structures with fatigue cracks are created and mapped onto a 3D graphics model. Subsequently, this model is used to generate synthetic images for various lighting conditions and camera angles. A fully convolutional network is then trained for two cases: (1) using only real-word data, and (2) using both synthetic and real-word data. By employing synthetic data augmentation in the training process, the crack identification performance of the neural network for the test dataset is seen to improve from 35% to 40% and 49% to 62% for intersection over union (IoU) and precision, respectively, demonstrating the efficacy of the proposed approach.
Key Words
fully convolutional networks; semantic segmentation; steel fatigue crack; synthetic data
Address
(1) Guanghao Zhai, Shuo Wang, Shaik Althaf V. Shajihan, Billie F. Spencer Jr.:
Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, USA;
(2) Yasutaka Narazaki:
Zhejiang University – University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, Zhejiang 314400, China.
Abstract
Structural Health Monitoring (SHM) of critical infrastructure comprises a major pillar of maintenance management, shielding public safety and economic sustainability. Although SHM is usually associated with data-driven metrics and thresholds, expert judgement is essential, especially in cases where erroneous predictions can bear casualties or substantial economic loss. Considering that visual inspections are time consuming and potentially subjective, artificial-intelligence tools may be leveraged in order to minimize the inspection effort and provide objective outcomes. In this context, timely detection of sensor malfunctioning is crucial in preventing inaccurate assessment and false alarms. The present work introduces a sensor-fault detection and interpretation framework, based on the well-established support-vector machine scheme for anomaly detection, combined with a coalitional game-theory approach. The proposed framework is implemented in two datasets, provided along the 1st International Project Competition for Structural Health Monitoring (IPC-SHM 2020), comprising acceleration and cable-load measurements from two real cable-stayed bridges. The results demonstrate good predictive performance and highlight the potential for seamless adaption of the algorithm to intrinsically different data domains. For the first time, the term "decision trajectories", originating from the field of cognitive sciences, is introduced and applied in the context of SHM. This provides an intuitive and comprehensive illustration of the impact of individual features, along with an elaboration on feature dependencies that drive individual model predictions. Overall, the proposed framework provides an easyto-train, application-agnostic and interpretable anomaly detector, which can be integrated into the preprocessing part of various SHM and condition-monitoring applications, offering a first screening of the sensor health prior to further analysis.
Key Words
decision trajectories; decision trajectory assurance criterion; DTAC; interpretable AI; one class classifiers; sensor fault detection; SHAP; SHM
Address
(1) Panagiotis Martakis, Yves Reuland, Eleni Chatzi:
Department of Civil, Environmental and Geomatic Engineering, ETH Zurich, Stefano-Franscini-Platz 5, 8093 Zurich, Switzerland;
(2) Artur Movsessian, David Garcıa Cava:
School of Engineering, Institute for Infrastructure and Environment, University of Edinburgh, Alexander Graham Bell Building, Thomas Bayes Road, Edinburgh EH9 3FG, UK;
(3) Sai G.S. Pai:
Intellithink Industrial IoT Labs, Bengaluru, India (previously with Future Cities Laboratory, Singapore ETH Centre, Singapore);
(4) Said Quqa:
Department of Civil, Chemical, Environmental, and Materials Engineering, University of Bologna, Viale del Risorgimento 2, 40136 Bologna, Italy;
(5) Dmitri Tcherniak:
Bruel & Kjaer Sound and Vibration Measurements, Skodsborgvej 307, Naerum 2850, Denmark.
