Figure 4 - uploaded by Leo Pauly
Content may be subject to copyright.
ReLU activation function 

ReLU activation function 

Source publication
Conference Paper
Full-text available
Pavement crack detection using computer vision techniques has been studied widely over the past several years. However, these techniques have faced several limitations when applied to real world situations due to for example changes of lightning conditions or variation in textures. But the recent advancements in the field of artificial neural netwo...

Context in source publication

Context 1
... classification tasks. Hence, we have used max pooling in our network. Fig. 2 illustrates the max pooling operation. The input matrix of size 4x4 is divided into 4 submatrices. Then the max value in each of the submatrix is taken and the output matrix is created using these values. [16] introduced a very effective activation function called ReLU. Fig. 4 shows the ReLU activation function. One of its main advantages is that it has zero output when the input is negative and has the same value as input when the input is positive. And hence the gradients of the activation functions are always either 1 or 0 and this helps to avoid the problem of vanishing gradients [17] in deeper neural ...

Citations

... Compared with traditional digital image processing methods, CNNs are characterized by their high level of automation and strong feature extraction capability, as CNNs do not rely on manually designed feature operators. In terms of crack recognition applications, some studies localize cracks in images by classification [20,21] or object detection [22,23] methods. However, these methods cannot obtain detailed information about the cracks, making them less optimal. ...
Article
Full-text available
In recent years, convolutional neural-network-based crack segmentation methods have performed excellently. However, existing crack segmentation methods still suffer from background noise interference, such as dirt patches and pitting, as well as the imprecise segmentation of fine-grained spatial structures. This is mainly due to the fact that convolutional neural networks dilute low-level spatial information in the process of extracting deep semantic features, and the network cannot obtain accurate context awareness because of the limitation of the actual receptive field size. To address these problems, an encoder–decoder crack segmentation network based on multi-scale contextual information enhancement is proposed. First, a new architecture of skip connection is proposed, enabling the network to obtain refined crack segmentation results; then, a contextual feature enhancement module is designed to make the network more effective at distinguishing between cracks and background noise; finally, the deformable convolution is introduced into the encoder network to further enhance its ability to extract the diverse morphological features of cracks by adaptively adjusting the sampling area and the receptive field size. Experiments show that the proposed method is effective in crack segmentation and outperforms mainstream segmentation networks such as DeepLab V3+ and UNet++.
... Cracks and other auxiliary damages such as voids, spalling, and cement corrosion have been identified from images using CNN classifiers that are less impacted by noise introduced by illumination, shadow, and projection [24]. In [25], they used deeper networks to detect pavement cracks, demonstrating the promise of deep learning. One of the difficulties in the literature is determining which holes are actual cracks and which are simply sealed over. ...
Article
Full-text available
Automatic road crack detection is an important transportation maintenance responsibility for ensuring driving comfort and safety. Manual inspection is considered to be a risky method because it is time consuming, costly, and dangerous for the inspectors. Automated road crack detecting techniques have been extensively researched and developed in order to overcome this issue. Despite the difficulties, most of the proposed methodologies and solutions involve machine vision and machine learning, which have lately acquired traction largely due to the increasingly more affordable processing power. Nonetheless, it remains a difficult task due to the inhomogeneity of crack intensity and the intricacy of the background. In this paper, a convolutional neural network-based method for crack detection is proposed. The method is inspired from recent advancements in applying machine learning to computer vision. The primary goal of this work is to employ convolutional neural networks to detect the road crack. Data in the form of images has been used as input, preprocessing and threshold segmentation is applied to the input data. The processed output is feed to CNN for feature extraction and classification. The training accuracy was found to be 96.20 %, the validation accuracy to be 96.50 %, and the testing accuracy to be 94.5 %.
... Shatnawi (8) used image-processing techniques and ANNs to detect pavement distresses in secondary roads, based on images acquired by drones. Pauly et al. (9) investigated how deep learning-based methods may affect several factors in pavement distress detection. Image-based mobile mapping has a major impact on conventional transportation surveying and mapping, such as modelling and estimation of road boundaries in road safety assistance (10)(11)(12). ...
Article
Full-text available
Rutting leads hydroplaning, accidents, poor riding quality, and significant maintenance costs. This study assists the development of statistical and Artificial pavement rutting models. The proposed methodology is reliable, time-saving, cost-saving, and comfortable. The suggested technique to anticipate rutting considers traffic volumes, pavement, and geometrical parameters such as lane and shoulder widths. This research modeled 33 main highways' ruts. Most of these roads have serious de-stressing problems with rutted pavement. The developed rutting prediction models demonstrated a medium to high correlation between rut depth and independent variables including annual average daily traffic, truck fleet percentage, pavement thickness, and number of lanes. The correlation coefficients such as R ² were found to be moderate for most of the developed models. The linear models of rutting prediction were statistically significant, with R2 values averaging around 66%, whereas the logistic regression model was the best developed rutting model, with an R2 value of 67%, when all variables, including traffic, pavement, and geometry, were considered. Nonlinear models with an R2 value of 57% were used to get similar findings. The artificial neural network (ANN) has been used in this study to model rut depth with same independent variables and gave higher results with R2 value of 82%. The findings showed that an ANN outperformed regression modeling in predicting the depth of a rut.
... The pretrained model results outperformed CNN models that were trained from scratch. CNN models trained from scratch, such as those of [43], only achieved an accuracy of around 90%. This was expected, as pretrained models have weights generated from millions of images, while models trained from scratch have only thousands. ...
Article
Full-text available
Infrastructure, such as buildings, bridges, pavement, etc., needs to be examined periodically to maintain its reliability and structural health. Visual signs of cracks and depressions indicate stress and wear and tear over time, leading to failure/collapse if these cracks are located at critical locations, such as in load-bearing joints. Manual inspection is carried out by experienced inspectors who require long inspection times and rely on their empirical and subjective knowledge. This lengthy process results in delays that further compromise the infrastructure’s structural integrity. To address this limitation, this study proposes a deep learning (DL)-based autonomous crack detection method using the convolutional neural network (CNN) technique. To improve the CNN classification performance for enhanced pixel segmentation, 40,000 RGB images were processed before training a pretrained VGG16 architecture to create different CNN models. The chosen methods (grayscale, thresholding, and edge detection) have been used in image processing (IP) for crack detection, but not in DL. The study found that the grayscale models (F1 score for 10 epochs: 99.331%, 20 epochs: 99.549%) had a similar performance to the RGB models (F1 score for 10 epochs: 99.432%, 20 epochs: 99.533%), with the performance increasing at a greater rate with more training (grayscale: +2 TP, +11 TN images; RGB: +2 TP, +4 TN images). The thresholding and edge-detection models had reduced performance compared to the RGB models (20-epoch F1 score to RGB: thresholding −0.723%, edge detection −0.402%). This suggests that DL crack detection does not rely on colour. Hence, the model has implications for the automated crack detection of concrete infrastructures and the enhanced reliability of the gathered information.
... Deep Learning is a technique in Artificial Intelligence (AI) that uses artificial neural networks and is widely used to classify images [54]. Traditionally, crack detection algorithms use non-AI techniques such as Local Binary Patterns and shape-based algorithms [54]. ...
... Deep Learning is a technique in Artificial Intelligence (AI) that uses artificial neural networks and is widely used to classify images [54]. Traditionally, crack detection algorithms use non-AI techniques such as Local Binary Patterns and shape-based algorithms [54]. Since 2012, deep learning algorithms have gained popularity and outperformed traditional detection methods [22]. ...
Preprint
Full-text available
Inspection of cracks is an important process for properly monitoring and maintaining a building. However, manual crack inspection is time-consuming, inconsistent, and dangerous (e.g., in tall buildings). Due to the development of open-source AI technologies, the increase in available Unmanned Aerial Vehicles (UAVs) and the availability of smartphone cameras, it has become possible to automate the building crack inspection process. This study presents the development of an easy-to-use, free and open-source Automated Building Exterior Crack Inspection Software (ABECIS) for construction and facility managers, using state-of-the-art segmentation algorithms to identify concrete cracks and generate a quantitative and qualitative report. ABECIS was tested using images collected from a UAV and smartphone cameras in real-world conditions and a controlled laboratory environment. From the raw output of the algorithm, the median Intersection over Unions for the test experiments is (1) 0.686 for indoor crack detection experiment in a controlled lab environment using a commercial drone, (2) 0.186 for indoor crack detection at a construction site using a smartphone and (3) 0.958 for outdoor crack detection on university campus using a commercial drone. These IoU results can be improved significantly to over 0.8 when a human operator selectively removes the false positives. In general, ABECIS performs best for outdoor drone images, and combining the algorithm predictions with human verification/intervention offers very accurate crack detection results. The software is available publicly and can be downloaded for out-of-the-box use at: https://github.com/SMART-NYUAD/ABECIS
... Zhang et al. [15] developed a deep CNN-based method to classify image patches with or without cracks in road images captured by smartphone, which was reported to work better than the conventional methods. Pauly et al. [16] also reported that deeper networks can achieve a superior accuracy in pavement crack detection. The patch-based classification methods generally obtain a lower accuracy than the pixel-wise segmentation methods, so the latter are now the mainstream methods, most of which use a fully convolutional network (FCN) [17] as the backbone structure. ...
Article
Detecting the terrain surface cracks caused by earthquakes, which are termed coseismic ruptures, has important significance for discovering concealed faults, monitoring their movements, and forecasting possible follow-on earthquakes. On May 22, 2021, Maduo County in Qinghai province, China, suffered an earthquake with a magnitude of 7.4, which created densely distributed cracks. In this study, we designed an automatic crack detection framework based on remote sensing technology. With the use of low-altitude unmanned aerial vehicles (UAVs), we obtained very high-resolution aerial images of the area affected by the earthquake, which were further processed by photogrammetric software to produce digital orthophoto maps (DOMs). We then designed a novel terrain surface crack detection neural network, which differs from the previous methods that focus on detecting cracks in man-made object surfaces such as flat roads. We investigated the spatial property of the sinuous linear cracks and handled this by introducing adaptive deformable convolutions with a context-channel-space boosted mechanism. The feature extraction stage, feature optimization stage, and upsampling stage were embedded with the deformable convolutions to form a compact and powerful crack detector, named Crack-CADNet (the Context-chAnnel-space boosted Deformable convolutional neural network for crack detection). The postprocessing included filtering out the nontectonic cracks, aided by annotations from experts, and grouping and vectorizing the generated binary segmentation map as crack polygons, which were evaluated at the instance level. In addition to the first in-depth investigation of detecting earthquake cracks with aerial remote sensing and a deep learning based process, the crack detection network we propose outperformed the recent convolutional neural network (CNN)-based methods designed for general semantic segmentation and crack detection. Source code and the Maduo earthquake crack dataset will be available at http://gpcv.whu.edu.cn/data/.
... The input image size of 256 × 256 and the trained network recorded an accuracy of 0.9822. Similarly, [64] Pa et al. used almost the same architecture as Cha with an input image size of 99 × 99 without batch normalisation, and the accuracy of testing got 0.9020. It can be seen that simple CNN models can be used to classify binary classification problems. ...
Article
Full-text available
Cracks are an acute distress in an asphalt pavement, which must be detected and quantified to diagnose the pavement’s health. Hence, many researchers have developed methods to detect cracks based on three main techniques: image processing, machine learning (ML), and deep learning (DL). Among these three techniques, DL has been recognised as an excellent method for crack detection because it assures high accuracy with an adequate analysis time. However, choosing an appropriate DL algorithm to identify cracks in an asphalt pavement is challenging for both transportation agencies and researchers. This study has identified the bigger picture of DL methods for crack identification in asphalt pavement. The authors evaluated several DL-based crack identification algorithms from the literature, such as crack classification, crack object detection, pixel-level crack segmentation, generative adversarial networks (GANs) for crack segmentation, and crack identification using unsupervised learning. Moreover, 26 DL-based crack detection models (25 supervised learning models and one unsupervised learning model) were analysed on the same dataset to test the performance of each model using consistent assessment metrics. The testing results suggest that ResNet and DenseNet are the best options for crack classification, while Faster R-CNN should be used for crack object detection and pix2pix is suggested for crack segmentation. It is also recommended that semi-supervised and unsupervised learning be further studied to efficiently detect cracks in an asphalt pavement.
... The first crack detection attempts are based on the combination of image processing & ML regression tasks [12,[15][16][17]. Other researchers address the topic performing crack detection using DL crack classification [7,8,14,[18][19][20] or semantic segmentation of cracks [6,[21][22][23]. ...
... Sometimes, it is also seen as an additional augmentation approach [24]. The value of its hyperparameter ρ, representing the probability of disconnecting neurons at each training step, is usually set to 0.5 for crack detection [8,18,21,45]. Another very popular approach for the same problem is so called transfer learning. ...
Article
Deep Learning (DL) semantic image segmentation is a technique used in several fields of research. The present paper analyses semantic crack segmentation as a case study to review the up to date research on semantic segmentation in the presence of fine structures and the effectiveness of established approaches to address the inherent class imbalance issue. The established UNet architecture is tested against networks consisting exclusively of stacked convolution without pooling layers (straight networks), with regard to the resolution of their segmentation results. Dice and Focal losses are also compared against each other to evaluate their effectiveness on highly imbalanced data. With the same aim, dropout and data augmentation approaches are tested, as additional regularizing mechanisms, to address the uneven distribution of the dataset. The experiments show that the good selection of the loss function has more impact in handling the class imbalance and boosting the detection performance than all the other regularizers with regards to segmentation resolution. Moreover, UNet, the architecture considered as reference, clearly outperforms the networks with no pooling layers both in performance and training time. The authors argue that UNet architectures, compared to the networks with no pooling layers, achieve high detection performance at a very low cost in terms of training time. Therefore, the authors consider such architecture as the state of the art for semantic segmentation of cracks. On the other hand, once computational cost is not an issue anymore thanks to constant improvements of technology, the application of networks without pooling layers might become attractive again because of their simplicity of and high performance.
... [39], [52], [56] S Approaches using a standard or slightly modified U-Net ...
... One of the earliest approaches for classification using CNNs is by Zhang et al. [51] where a shallow CNN using 4 convolutional and 2 fully connected layers is trained and tested on small, 99×99 pixel patches which have been annotated as to whether they contain a crack or not. This was then improved upon in [52] where it was found that classification networks for cracks with more layers (e.g. deeper networks) improve the classification performance. ...
... To illustrate those metrics, crack semantic segmentation recall states how much of the actual crack pixels have been correctly predicted by the method and precision as to how many of the predicted pixels are actually on a crack. A large number of works use those metrics to report their results for classification [18], [51], [52], [54], detection [120], [128] or segmentation [39], [57]- [59]. It is common to use P R, RE, and F 1 on boolean data, meaning the results are either true (1) or false (0). ...
Preprint
Surface cracks are a very common indicator of potential structural faults. Their early detection and monitoring is an important factor in structural health monitoring. Left untreated, they can grow in size over time and require expensive repairs or maintenance. With recent advances in computer vision and deep learning algorithms, the automatic detection and segmentation of cracks for this monitoring process have become a major topic of interest. This review aims to give researchers an overview of the published work within the field of crack analysis algorithms that make use of deep learning. It outlines the various tasks that are solved through applying computer vision algorithms to surface cracks in a structural health monitoring setting and also provides in-depth reviews of recent fully, semi and unsupervised approaches that perform crack classification, detection, segmentation and quantification. Additionally, this review also highlights popular datasets used for cracks and the metrics that are used to evaluate the performance of those algorithms. Finally, potential research gaps are outlined and further research directions are provided.
... After this first application, researchers started investigating different aspects of the problem. Pauly et al. [25] studied the effect of depth (the number of deep layers) on the crack detection capability of the network. They concluded that the deeper the networks are, the more information can be learnt by the architecture. ...
Article
Full-text available
The application of deep architectures inspired by the fields of artificial intelligence and computer vision has made a significant impact on the task of crack detection. As the number of studies being published in this field is growing fast, it is important to categorize the studies at deeper levels. In this paper, a comprehensive literature review of deep learning-based crack detection studies and the contributions they have made to the field is presented. The studies are categorised according to the computer vision aspect and at deeper levels to facilitate exploring the studies that utilised similar approaches to address the crack detection problem. Moreover, the authors perform a comparison between the studies which use the same publicly available data sets, in order to find the most promising crack detection approaches. Critical future directions for research are proposed, based on these reviewed studies as well as on trends and developments in areas similar to the crack detection area.