Tiny‐Crack‐Net: A multiscale feature fusion network with attention mechanisms for segmentation of tiny cracks

Automatic crack classification plays an essential role in road maintenance. Using many features for the classification is inefficient for implementing embedded systems with low computational resources makes it difficult. Therefore, this work proposes a new data dimensionality reduction (DDR) for crack classification algorithms (DDR4CC). DDR4CC reduces the required information about the cracks to only four features. Using these features, the images can be classified into longitudinal, transverse, and alligator cracks or healthy pavement. DDR4CC is compared with eight DDR methods, and the reduced set of features is analyzed using five different classification algorithms. Besides, five different datasets, generated by a combination of several public datasets, are used. We are proposing a simple DDR method with high interpretability of the data, obtaining very fast computation and high accuracy. Experiments show that DDR4CC enhances the results of the classification algorithms, providing almost perfect classifiers with a minimum computation time.

Section: Ddr For Crack Classification Algorithms (Ddr4cc)mentioning

confidence: 99%

“…Approaches based on crack detection (Chu et al., 2022) usually have two different objectives. The first objective is to obtain the regions of interest where defects are located.…”

Section: Introductionmentioning

confidence: 99%

Efficient data dimensionality reduction method for improving road crack classification algorithms

Rodríguez-Lozano

Gámez-Granados

Palomares

et al. 2023

“…In this study, α box , α cls , and α obj are the coefficients, which are set as 0.05, 0.3, 0.7, respectively. II To further improve the feature extraction capability, the ASFF architectures were inserted into the YOLOv5s model to improve the detection accuracy by spatially filtering conflicting information by addressing inconsistency, thereby improving scale-invariant properties (Chu et al, 2022;Liu et al, 2019). As shown in Figure 2c, the features of all other levels in images can be resized into the same shape and spatial fusion according to the learned weight map.…”

Section: Improved Yolov5s Networkmentioning

confidence: 99%

“…These unfavorable and low‐quality factors inevitably bring huge challenges to vision‐based detection algorithms. Because of the interference of visibility and imaging perspective, small‐scale target detection problems often exist in civil infrastructures defect detection, further hindering the detection accuracy of vision‐based algorithms (Chu et al., 2022). Thus, it is necessary to improve the model feature extraction effect and study the damage identification method for small‐scale defects.…”

Section: Introductionmentioning

confidence: 99%

A robust real‐time method for identifying hydraulic tunnel structural defects using deep learning and computer vision

Bao

et al. 2022

Robots with cameras provide a non-contact information acquisition solution for hydraulic tunnels, while manual damage-related information extraction is timeconsuming and costs labor. This study proposes a robust real-time framework for identifying hydraulic tunnel underwater structural damage using deep learning and computer vision. First, a high-performance detector is built via the You Only Look Once v5s and adaptively spatial feature fusion module. A series of comparative experiments are used to explore the setting of the sparsification and pruning ratios to trade off a balance between accuracy and efficiency. Model sparsity ratio of 0.01 with a pruning ratio of 0.3 can be combined to change the weight distribution in the batch normalization layer and reduce network redundant parameters for slimming. Then, model fine-tuning with knowledge distillation is utilized to recover the accuracy degradation caused by pruning. A hydraulic tunnel is utilized as the case study, and three defects including rust, exfoliation, and calcification precipitate are utilized as research items. The performance of the models was evaluated based on detection accuracy, robustness, and efficiency. Five extreme attributes of underwater scenes, including oblique angles, high brightness, uneven illumination, low visibility, and obstacle interference, were considered to test model generalization and efficacy. Experimental results show it achieves good detection performance in complicated underwater scenes, achieving 0.814 precision, 0.980 recall, 0.889 F1_score, and 0.894 Mean Average Precision (mAP)@0.5 in the test set. Moreover, the proposed method achieved a 50 Frames Per Second (FPS) detection speed when detecting video with 1080p, indicating its real-time detection capability. INTRODUCTIONThe linings of hydraulic tunnels inevitably suffer from defects, like cracks, rust, exfoliation, and calcification precipitates, under the coupling effect of external pressure and high-speed water flow erosion (Zhou et al., 2022). Timely and reliable identification of these structural defects is critical for characterizing the health of © 2022 Computer-Aided Civil and Infrastructure Engineering.hydraulic tunnels. The periodic visual inspection by professionals has been a practical solution in civil infrastructures (Wang et al., 2021). However, it suffers from problems like high risk, labor cost, and low efficiency, especially for hydraulic tunnels in water-filled states. Moreover, emptying the water in tunnels may cause stress redistribution, even resulting in the local collapse of linings.

“…Deep learning algorithms (i.e., classification algorithm, target detection algorithm, and semantic segmentation algorithm) have been gradually applied in the detection of structural surface crack recently (Sajedi & Liang, 2020; Yu et al., 2021; Zou et al., 2018). Among them, the semantic segmentation algorithm can not only classify and locate the crack, but also accurately segment its specific morphological features (Chu et al., 2022). However, most of the semantic segmentation algorithms used in the existing research on tunnel lining crack identification are based on convolution neural network (CNN) (Benamara et al., 2021; LeCun et al., 1998; Peng et al., 2021).…”

Section: Introductionmentioning

confidence: 99%

Hybrid semantic segmentation for tunnel lining cracks based on Swin Transformer and convolutional neural network

Zhou

Zhang

Gong

2023

In the field of tunnel lining crack identification, the semantic segmentation algorithms based on convolution neural network (CNN) are extensively used. Owing to the inherent locality of CNN, these algorithms cannot make full use of context semantic information, resulting in difficulty in capturing the global features of crack. Transformer‐based networks can capture global semantic information, but this method also has the deficiencies of strong data dependence and easy loss of local features. In this paper, a hybrid semantic segmentation algorithm for tunnel lining crack, named SCDeepLab, is proposed by fusing Swin Transformer and CNN in the encoding and decoding framework of DeepLabv3+ to address the above issues. In SCDeepLab, a joint backbone network is introduced with CNN‐based Inverse Residual Block and Swin Transformer Block. The former is used to extract the local detailed information of the crack to generate the shallow feature layer, whereas the latter is used to extract the global semantic information to obtain the deep feature layer. In addition, Efficient Channel Attention enhanced Feature Fusion Module is proposed to fuse the shallow and deep features to combine the advantages of the two types of features. Furthermore, the strategy of transfer learning is adopted to solve the data dependency of Swin Transformer. The results show that the mean intersection over union (mIoU) and mean pixel accuracy (mPA) of SCDeepLab on the data sets constructed in this paper are 77.41% and 84.42%, respectively, which have higher segmentation accuracy than previous CNN‐based and transformer‐based semantic segmentation algorithms.