Iterative application of generative adversarial networks for improved buried pipe detection from images obtained by ground‐penetrating radar

Abstract

Ground‐penetrating radar (GPR) is widely used to determine the location of buried pipes without excavation, and machine learning has been researched to automatically identify the location of buried pipes from the reflected wave images obtained by GPR. In object detection using machine learning, the accuracy of detection is affected by the quantity and quality of training data, so it is important to expand the training data to improve accuracy. This is especially true in the case of buried pipes that are located underground and whose existence cannot be easily confirmed. Therefore, this study developed a method for increasing training data using you only look once v5 (YOLOv5) and StyleGAN2‐ADA to automate the annotation process. Of particular importance is developing a framework for generating images by generative adversarial networks with an emphasis on images that are challenging to detect buried pipes in YOLOv5 and add them to a training dataset to repeat training recursively, which has greatly improved the detection accuracy. Specifically, F‐values of 0.915, 0.916, and 0.924 were achieved by automatically generating training images step by step from only 500, 1000, and 2000 training images, respectively. These values exceed the F‐value of 0.900, which is obtained from training by manually annotating 15,000 images, a much larger number. In addition, we applied the method to a road in Shizuoka Prefecture, Japan, and confirmed that the method can detect the location of buried pipes with high accuracy on a real road. This method can contribute to labor‐saving training data expansion, which is time‐consuming and costly in practice, and as a result, the method contributes to improving detection accuracy.

A response‐compatible ground motion generation method using physics‐guided neural networks

“…On the contrary, when real data are scarce due to difficulties or high costs in acquisition, using generative adversarial networks (GANs) to produce high‐quality synthetic data as a supplement to real data is a highly effective means. GANs have been employed in vessel trajectory data augmentation (Zhang et al., 2022), generation for ground‐penetrating radar images of buried pipes (Chun et al., 2023) and scanning electron microscope images of nanofibers (Ieracitano et al., 2022), and electroencephalogram data augmentation (Xu et al., 2022). In the area of civil engineering, Rafiei & Adeli (2017b) proposed a neural earthquake early warning system to forecast earthquake magnitude and location.…”

Deep learning‐based automatic classification of three‐level surface information in bridge inspection

“…Wu et al (2019) pruned and utilized VGG16 to classify corrosion and crack defects in bridge structures, achieving accuracy rates of 93.6% and 98.5%, respectively. Novel networks like Faster Region-CNN (Faster RCNN) and You Only Look Once (YOLO) series (Chun et al, 2023; have emerged for object detection, while Mask RCNN and U-Net (Yamaguchi & Mizutani, 2023) are employed for seg-mentation. Recent researchers proposed EfficientNet and incorporated attention mechanisms (Chen & He, 2022;Y.…”

“…(2019) pruned and utilized VGG16 to classify corrosion and crack defects in bridge structures, achieving accuracy rates of 93.6% and 98.5%, respectively. Novel networks like Faster Region‐CNN (Faster RCNN) and You Only Look Once (YOLO) series (Chun et al., 2023; Z. Zhou et al., 2022) have emerged for object detection, while Mask RCNN and U‐Net (Yamaguchi & Mizutani, 2023) are employed for segmentation. Recent researchers proposed EfficientNet and incorporated attention mechanisms (Chen & He, 2022; Y. Pan & Zhang, 2022; L. Zhang et al., 2023), separable convolution (Zhu et al., 2023; Zou et al., 2022), deformable convolution (Lei et al., 2023), atrous convolution (Siriborvornratanakul, 2023), and other strategies (Zheng et al., 2022) to further enhance model performance.…”

Intelligent recognition of defects in high‐speed railway slab track with limited dataset

“…Guo et al (2019) proposed a method for generating images of railway intrusion targets based on improved conditional deep convolutional generative adversarial networks. Chun et al have made remarkable contributions to the field of damage detection (Chun & Hayashi, 2021;Chun et al, 2023). They completed multiple non-destructive testing and evaluation of internal damage in reinforced concrete based on the random forest algorithm (Chun et al, 2020) and completed the recording of bridge damage areas and automatic generation of image captions based on deep learning technology Yamane et al, 2023).…”