A night pavement crack detection method based on image‐to‐image translation

Abstract

Deep learning provides an efficient automated method for pavement condition surveys, but the datasets used for this model are usually images taken in good lighting conditions. If images are taken at night, this model cannot work effectively. This paper proposes a method for normalizing pavement images at night, which includes three main steps. First, the image feature point detection and matching method is used to process images taken during the day and night. Then, paired images of pavement during the day and night are obtained. Second, with the help of the image-to-image translation model, those paired images are used for training, and the best model for converting night images into day images is selected. Third, a convolutional neural network (CNN) based on VGGNet is constructed, and pavement images taken during the day are used for training.After that, six types of images are used and tested separately, namely, those taken during the day and the night, converted by the proposed method and converted by traditional methods. As evaluated by various evaluation indices and visualization methods, the detection performance of the CNN model can be significantly improved by using the proposed method of converted night-to-day images.

A Survey of Defect Detection Applications Based on Generative Adversarial Networks

“…For 400 road images, the average recognition rate reached 81.540% and 79.228%, respectively. Literature [110], [111], [112], [113], [114] has also made improvements and upgrades for GAN used in construction defect detection. Table 5 shows the research application and performance of GAN-based defect detection in the field of construction road quality inspection.…”

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

“…As an alternative, the computer‐aided method, which demonstrates powerful feature extraction (Aljarrah et al., 2022; Martins et al., 2020), analysis (Hassanpour et al., 2019), processing capabilities (Rafiei & Adeli, 2016; Rafiei et al., 2017), and condition adaptability (Yishun Li et al., 2021), has been successfully implied to automated classification of pavement conditions (Hsieh et al., 2021), monitoring of road degradation (Van Hauwermeiren et al., 2022), cracks detection (J. Chen & He, 2022; Chao Liu & Xu, 2022; Xie et al., 2022), and multiple pavement distresses detection (Zhang et al., 2022), in which the capability of generating virtually realistic road morphology for enriching pavement surface texture datasets was proved. Unlike previous methods of regenerating artificially predetermined features, an unsupervised learning method, generative adversarial networks (GANs; Goodfellow et al., 2020), was implemented in this study.…”

Real‐time automatic crack detection method based on drone

“…Liu et al, 2020;Meng et al, 2022). C. Liu and Xu (2022) used an image feature point matching technology and image-to-image conversion technology to standardize the pavement crack image at night. The detection of pavement crack images at night can be realized without re-labeling the dataset.…”

“…Besides, some researchers utilize multi‐stage crack detection and achieve high‐precision crack edge detection by combining the classification algorithm with the segmentation algorithm (J. Liu et al., 2020; Meng et al., 2022). C. Liu and Xu (2022) used an image feature point matching technology and image‐to‐image conversion technology to standardize the pavement crack image at night. The detection of pavement crack images at night can be realized without re‐labeling the dataset.…”