Intelligent recognition of subsurface utilities and voids: A ground penetrating radar dataset for deep learning applicationsMendeley Data
Ground Penetrating Radar (GPR), has emerged as a powerful non-invasive geophysical technique for detecting subsurface utilities, voids, and other subsurface anomalies. However, despite its widespread use in geophysical investigations, and construction management, there is lack of available datasets...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-04-01
|
| Series: | Data in Brief |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2352340925000708 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Ground Penetrating Radar (GPR), has emerged as a powerful non-invasive geophysical technique for detecting subsurface utilities, voids, and other subsurface anomalies. However, despite its widespread use in geophysical investigations, and construction management, there is lack of available datasets containing B-scan images of the subsurface features publicly that could be used to train deep learning models for automated anomaly detection. This data article aims at contributing to fill up this gap by creating a dataset specifically designed for automatic detection of subsurface utilities, and voids using deep learning. The dataset consists of 2,239 Radargram images in JPEG format obtained from GPR surveys conducted in urban environments to identify utilities such as pipes, cables, and underground voids. The importance of this dataset lies in: (1) contribute to fill the gap of lack of GPR data, (2) the universality of the data, (3) its potential to enhance the accuracy and efficiency to detect subsurface anomaly through the application of deep learning models, (4) GPR surveys are highly effective but still expensive, and its processing is time-consuming. By providing this labelled dataset for deep learning model training, this can facilitate the development of automated systems, capable of detecting subsurface anomalies effectively, which could reduce manual errors. |
|---|---|
| ISSN: | 2352-3409 |