Automatic detection of moisture damages in asphalt pavements from GPR data with deep CNN and IRS method

• Published in: Automation in construction Vol. 113; p. 103119
• Main Authors: Zhang, Jun, Yang, Xing, Li, Weiguang, Zhang, Shaobo, Jia, Yunyi
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.05.2020; Elsevier BV
• Subjects: Algorithms; Artificial neural networks; Asphalt pavement; Asphalt pavements; Bridge inspection; Convolutional neural network (CNN); Damage detection; Damage localization; Deep learning; Feature extraction; Ground penetrating radar; Ground penetrating radar (GPR); Human error; Localization method; Machine learning; Moisture; Moisture damage; Non-destructive evaluation (NDE); Random sampling; Deep learning; Ground penetrating radar (GPR); Non-destructive evaluation (NDE); Moisture damage; Asphalt pavement; Convolutional neural network (CNN); Bridge inspection
• ISSN: 0926-5805; 1872-7891
• DOI: 10.1016/j.autcon.2020.103119

Accurate detection and localization of moisture damage in asphalt pavements using Ground Penetrating Radars (GPR) has been attracting more and more interest in research. Existing approaches rely heavily on human efforts and expert experience and are thus both time and cost consuming and are also subject to accuracy issues caused by stochastic human errors. To address this issue, this paper presents an automated moisture damage detection and localization method by leveraging the state-of-the-art deep learning approach and newly proposed incremental random sampling (IRS) approach. First, 2.3 GHz Ground coupled GPR system was used to survey moisture damages on 16 asphalt pavement bridges to create three moisture damage datasets with different resolutions including 2135 moisture damages and 474 steel joints. On this basis, we propose mixed deep convolutional neural networks (CNN) including ResNet50 network, for feature extraction, and YOLO v2 network, for recognition, to detect and localize moisture damages. In addition, to prepare the input for the deep learning models, an IRS algorithm is proposed to generate suitable GPR images from GPR data to feed the CNN. Comprehensive experimental testing, analysis, and comparison of the proposed approaches are conducted. Experimental results demonstrated the promising performance and superiority of the proposed approaches in detecting and localizing moisture damages in asphalt pavements. •Moisture damage in bridge deck asphalt pavement is successfully detected and visualized in GPR image.•Moisture damage dataset is constructed from 16 asphalt pavement bridges.•Proposed CNN model is applied to detect moisture damage with F1 score (91.97%), Recall (94.53%) and Precision (91.00%).•A novel IRS algorithm for selecting GPR image with suitable plot scale for deep learning is proposed.•Experimental results demonstrate promising performance in detection and localization of moisture damages with IRS and CNN.