Segmentation Detection Method for Complex Road Cracks Collected by UAV Based on HC-Unet++

Abstract

Road cracks are one of the external manifestations of safety hazards in transportation. At present, the detection and segmentation of road cracks is still an intensively researched issue. With the development of image segmentation technology of the convolutional neural network, the identification of road cracks has also ushered in new opportunities. However, the traditional road crack segmentation method has these three problems: 1. It is susceptible to the influence of complex background noise information. 2. Road cracks usually appear in irregular shapes, which increases the difficulty of model segmentation. 3. The cracks appear discontinuous in the segmentation results. Aiming at these problems, a network segmentation model of HC-Unet++ road crack detection is proposed in this paper. In this network model, a deep parallel feature fusion module is first proposed, one which can effectively detect various irregular shape cracks. Secondly, the SEnet attention mechanism is used to eliminate complex backgrounds to correctly extract crack information. Finally, the Blurpool pooling operation is used to replace the original maximum pooling in order to solve the crack discontinuity of the segmentation results. Through the comparison with some advanced network models, it is found that the HC-Unet++ network model is more precise for the segmentation of road cracks. The experimental results show that the method proposed in this paper has achieved 76.32% mIOU, 82.39% mPA, 85.51% mPrecision, 70.26% dice and Hd95 of 5.05 on the self-made 1040 road crack dataset. Compared with the advanced network model, the HC-Unet++ network model has stronger generalization ability and higher segmentation accuracy, which is more suitable for the segmentation detection of road cracks. Therefore, the HC-Unet++ network model proposed in this paper plays an important role in road maintenance and traffic safety.

An integration–competition network for bridge crack segmentation under complex scenes

“…To prove the advancement of CCSNet, this section compares CCSNet with seven SOTA professional crack segmentation networks, including the tunnel crack segmentation network CrackSegNet (Ren et al., 2020), lightweight crack segmentation network CrackFormer (Miao et al., 2021), bridge crack segmentation network BC‐DUnet (Liu et al., 2022), building crack segmentation network DCSNet (Pang et al., 2022), concrete crack segmentation network CrackW‐Net (C. Han et al., 2022), road crack segmentation network HC‐UNet++ (Cao et al., 2023), and ARD‐Unet (Yu et al., 2023). As can be seen from the results in Table 7, to make each network achieve the best training parameters, the Adam optimizer and SGD optimizer are used to do two groups of training.…”

“…CrackW-Net is designed by skiplevel round-trip sampling block structure. Cao et al (2023) proposed an HC-Unet++ and designed a deep parallel feature fusion module. HC-Unet++ could identify multiple irregularly shaped cracks.…”

“…Cao et al. (2023) proposed an HC‐Unet++ and designed a deep parallel feature fusion module. HC‐Unet++ could identify multiple irregularly shaped cracks.…”

Gradient Structure Information-Guided Attention Generative Adversarial Networks for Remote Sensing Image Generation

“…In the field of aerospace remote sensing, attention mechanisms also play an important role in many directions [48][49][50]. LMNet [51] proposes the Residual Transformer 3D-spatial Attention Module (RT3DsAM), which can learn feature representations from global information and filter important information.…”

Assessment of asphalt pavement aging condition based on GF-2 high-resolution remote sensing image

“…[13][14][15][16][17] Researchers have used drone aerial images to accurately identify and classify pavement defects, such as potholes and cracks. [18][19][20][21] They can also extract and calculate parameters, such as length, width, and area of the defects, 22,23 with recognition accuracy reaching centimeter 24 or millimeter 25 levels. However, the pavement research using drone images is limited to a small area.…”