Lightweight pixel-wise segmentation for efficient concrete crack detection using hierarchical convolutional neural network

Abstract

The aging of concrete structures is a threat to public safety; therefore, maintenance and repair of these structures have been highly emphasized. However, regular inspections to detect concrete cracks that rely on operators lack objectivity and consume a lot of time. To overcome this limitation, high-resolution image processing and deep learning have been adopted. Nevertheless, cracks on structure surfaces are still challenging to detect owing to the variety of shapes of cracks and the dependence of recognition performance on image conditions. Herein, we propose a new concrete crack detection method that applies the semantic segmentation technique using 1196 concrete crack images and labeled images produced in this study. A new segmentation algorithm is developed using a hierarchical convolutional neural network to improve speed, and a multi-loss update method is proposed to improve accuracy. The performance of the proposed network is evaluated in terms of accuracy and speed. The results show that the proposed network produces a 2.165% increase in the intersection over union of crack, 65.90% decrease in the average inference time, and 99.90% decrease in the number of parameters compared with the best accuracy results using existing segmentation networks. It is expected that the application of this improved crack detection method will result in faster and more accurate crack detection and, consequently, improved safety, thereby making it suitable for application in structure safety inspections.

Bridge crack detection based on improved single shot multi-box detector

“…The bridge crack dataset used for model training and testing was introduced, and the effectiveness of each method was verified separately. The proposed network was then compared with FCN [ 33 ], SSD, U-Net [ 34 ], CrackDFANet [ 35 ], LDCC-Net [ 36 ], FPHBN [ 37 ], and (ABCNet) Network in reference [ 38 ]. Finally, conclusions were drawn by analyzing the experimental results.…”

“…In the model training process, thresholds of different IOU should be set to measure the detection accuracy of the model. Experimental results in [ 36 ] show that it is appropriate to set the threshold value of IOU as 0.5 in the bridge crack detection task. The accuracy of model detection is usually described by a precision–recall curve (PR curve).…”

MSCNet: A Framework With a Texture Enhancement Mechanism and Feature Aggregation for Crack Detection

“…Considering the limited computing capacity of the bridge crack detection platform, six crack detection networks, namely FPHBN [27], DeepCrack [31], CliqueNet [26], SSENets [32], LDCC-Net [33], and DFANet [25], were selected for comparison. All the networks were trained and tested on the same experiment setup and dataset.…”

“…Li et al [32] designed a network composed of skipsqueeze-excitation and the atrous spatial pyramid pooling to detect cracks. Kim et al [33] applied the semantic segmentation technique to develop a hierarchical convolutional neural network to improve the accuracy and rate of crack detection.…”

Bibliometric Analysis and Review of Deep Learning-Based Crack Detection Literature Published between 2010 and 2022

“…In comparison with other segmentation networks, the encoder-decoders network architecture (U-net) requires a small amount of data to train and provide crack location information precisely. Various authors have also improved state-of-the-art DL models and named their architectures as APLCNet [98], RetinaNet [80], [172], SCHNet [94], Ci-Net [103], Cls-GAP Net [96], Ef-ficientNet [168], ProposedNet [109], CMDNet [161], TernausNet [185], Spiral Net [166], LDCC-Net [189], RRCE-Net [170], RAO-UNet [181] and so on. From the above analysis, it can be inferred that crack segmentation can be performed by implementing a simple lightweight encoder-decoder network that can be trained with less data and resources.…”