Share Email Print
cover

Proceedings Paper • new

Surface damage detection for concrete bridges using single-stage convolutional neural networks
Author(s): Chaobo Zhang; Chih-Chen Chang
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Detecting surface damages is vital for keeping concrete bridges structurally healthy and reliable. Currently, most of imagebased detection techniques are based on handcrafted low-level features which make them less applicable to actual images taken under varying environmental conditions. Recent rapid advancement in convolution neural network has enabled the development of deep learning-based visual inspection techniques for detecting multiple structural damages without needing manually-crafted features. However, most deep learning-based techniques are built on two-stage, proposal-driven detectors using less complex image data, which is not promising for practical applications and integration within intelligent autonomous inspection systems. In this study, a faster, simpler single-stage detector is proposed based on the real time object detection technique, You Only Look Once (YOLOv3) for detecting multiple surface damages of concrete bridges. A large field inspection images dataset of bridge damage is used for training and testing of YOLOv3. The original YOLOv3 is further improved by introducing a novel transfer learning method, batch renormalization and focal loss. The results show that the improved YOLOv3 has a detection accuracy of up to 80% and its performance is about 13% better than the original YOLOv3.

Paper Details

Date Published: 1 April 2019
PDF: 8 pages
Proc. SPIE 10972, Health Monitoring of Structural and Biological Systems XIII, 109722E (1 April 2019); doi: 10.1117/12.2513571
Show Author Affiliations
Chaobo Zhang, Hong Kong Univ. of Science and Technology (Hong Kong, China)
Chih-Chen Chang, Hong Kong Univ. of Science and Technology (Hong Kong, China)


Published in SPIE Proceedings Vol. 10972:
Health Monitoring of Structural and Biological Systems XIII
Paul Fromme, Editor(s)

© SPIE. Terms of Use
Back to Top