Regeneration of pavement surface textures using M‐sigmoid‐normalized generative adversarial networks

Lu, Jiale; Pan, B.; Ren, Wei; Liu, Quan; Liu, Pengfei; Oeser, Markus

doi:10.1111/mice.12987

Cited by 10 publications

(6 citation statements)

References 84 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The conditional forecasted crack‐generative adversarial network (CFC‐GAN) model (Sekar & Perumal, 2022) exploits this framework to generate crack images under varying lighting conditions and seasons. To address the high costs associated with collecting pavement textures, GAN‐based models have been effectively applied for texture regeneration (Lu et al., 2023). The data generation method based on GAN can also be utilized to relieve the issues of decreased recognition accuracy arising from data insufficiency and imbalanced data categories (Gao et al., 2021; Maeda et al., 2021).…”

Section: Related Workmentioning

confidence: 99%

A controllable generative model for generating pavement crack images in complex scenes

Zhang,

Qian,

Zhou

et al. 2024

Computer aided Civil Eng

Self Cite

View full text Add to dashboard Cite

Existing crack recognition methods based on deep learning often face difficulties when detecting cracks in complex scenes such as brake marks, water marks, and shadows. The inadequate amount of available data can be primarily attributed to this factor. To address this issue, a controllable generative model of pavement cracks is proposed that can generate crack images in complex scenes by leveraging background images and crack mask images. The proposed model, the crack diffusion model (CDM), is based on the diffusion model network, which enables better control over the position and morphology of cracks by adjusting the conditional input of cracks. Experiments show that CDM has several advantages, including high definition, controllability, and sensitivity to narrow cracks. Utilizing CDM to create a synthetic crack data set in complex scenes resulted in substantial improvements of crack detection and segmentation. The method proposed in this study can effectively alleviate the effort required for data acquisition and labeling, especially in complex scenes.

show abstract

Section: Related Workmentioning

confidence: 99%

A controllable generative model for generating pavement crack images in complex scenes

Zhang,

Qian,

Zhou

et al. 2024

Computer aided Civil Eng

Self Cite

View full text Add to dashboard Cite

show abstract

“…Hoang et al [11] developed an advanced image processing method and selected an extreme randomization tree (ERT) and a deep neural network (DNN) to analyze the features extracted from the above-mentioned methods. He et al [12] utilized unmanned aerial vehicle (UAV) to capture images of pavement damage and then combined with gray level co-occurrence matrix (GLCM) algorithm and cloud model theory to develop the pavement damage identification and evaluation model.…”

Section: Introductionmentioning

confidence: 99%

Research on the Skidding Resistance and Attenuation Characteristics of Asphalt Pavement Based on Image Recognition-Analysis Strategy

Zhang,

Xi,

Zhao

et al. 2024

Coatings

View full text Add to dashboard Cite

To accurately evaluate the skidding resistance of asphalt pavements, a texture imaging device was developed to realize the standardized acquisition of pavement images. Based on the gray-level co-occurrence matrix and multifractal theory of texture structure, the influence of segregation degree and gradation type on the texture properties of asphalt pavement was studied. Meanwhile, a comprehensive evaluation index of skidding resistance was proposed for asphalt pavement. Furthermore, the attenuation characteristics of the anti-skidding performance for asphalt mixture were explored, and the corresponding attenuation model of asphalt pavement was established. The results show that the segregation degree and gradation type significantly affected the texture parameters and anti-skidding performance of asphalt mixture. Specially, with an increase in the segregation degree of coarse aggregate, the parameters of energy, entropy, and multifractal spectrum width gradually increased, whereas the inertial moment gradually decreased. The variation range of the multifractal spectrum difference initially increased and subsequently decreased. For the texture parameters such as energy, entropy, inertial moment, and multifractal spectrum width Δα, the values of the asphalt mixture with larger nominal maximum particle were higher than those of the mixture with smaller nominal maximum particle, whereas the multifractal spectrum difference value showed the opposite law. In addition, the texture parameters of energy, entropy, and multifractal spectrum width exhibited good linear correlation with the texture depth (TD) of asphalt mixtures with various segregation levels and gradation types. The index based on the texture parameters of energy, entropy, and multifractal spectrum width effectively evaluated the skidding resistance of asphalt pavements, which showed the same trend as the TD with the increase of the abrasion number. The achievement provides an effective solution for the evaluation of skidding resistance and attenuation characteristics of asphalt mixtures.

show abstract

“…As such, it is essential to conduct regular maintenance and inspections to identify and address any potential issues preemptively. Current conventional inspection methods typically involve manual visual inspections carried out by professional inspectors (Lu et al., 2023). However, such manual visual inspection methods require heavy costs in time and human labor (Yeum & Dyke, 2015).…”

Section: Introductionmentioning

confidence: 99%

A method of concrete damage detection and localization based on weakly supervised learning

Jiang,

Pang,

et al. 2023

Computer aided Civil Eng

View full text Add to dashboard Cite

Automatic inspection of concrete surface defects based on visual elements is crucial for the timely detection of security risks in infrastructure. Moreover, accurate determination of the geographical location of the detected defects is critical for subsequent maintenance and reinforcement tasks. This study employed convolutional neural network (CNN) training methods for detection and localization. This approach employs bounding boxes to confine damaged pixels and utilizes projection loss to foster similarity learning between pixels. In addition, geotags are automatically indexed through the vectorization of invariant features in a scene, which maintains high model accuracy and reduces training costs. The proposed method can detect and classify typical concrete defects in complex scenarios and accurately locate them without the use of external sensors. In addition, the proposed model can achieve pixel‐level defect detection and geographic location determination through the cost of bounding box annotation and automatic indexing. The proposed model was evaluated using 10,691 images of four typical concrete defects in various complex environments. The results demonstrate that the proposed method achieves a pixel‐level detection accuracy of 48.75% and a location accuracy of 83.69%, showing a better performance than other methods.

show abstract

Regeneration of pavement surface textures using M‐sigmoid‐normalized generative adversarial networks

Cited by 10 publications

References 84 publications

A controllable generative model for generating pavement crack images in complex scenes

A controllable generative model for generating pavement crack images in complex scenes

Research on the Skidding Resistance and Attenuation Characteristics of Asphalt Pavement Based on Image Recognition-Analysis Strategy

A method of concrete damage detection and localization based on weakly supervised learning

Contact Info

Product

Resources

About