Abstract
This research investigates vision-based automated bridge component recognition, which is critical for automating visual inspection of bridges during initial response after earthquakes. Semantic segmentation algorithms with up to 45 convolutional layers are applied to recognize bridge components from images of complex scenes. One of the challenges in such scenarios is to get the recognition results consistent with high-level scene structure using limited amount of training data. To impose the high-level scene consistency, this research combines 10-class scene classification and 5-class bridge component classification. Three approaches are investigated to combine scene classification results into bridge component classification: (a) naïve configuration, (b) parallel configuration, and (c) sequential configuration of classifiers. The proposed approaches, sequential configuration in particular, are demonstrated to be effective in recognizing bridge components in complex scenes, showing less than 1% of accuracy loss from the naïve/parallel configuration for bridge images, and less than 1% false positives for the nonbridge images.
| Original language | English (US) |
|---|---|
| Journal | Computer-Aided Civil and Infrastructure Engineering |
| DOIs | |
| State | Accepted/In press - Jan 1 2019 |
Fingerprint
ASJC Scopus subject areas
- Civil and Structural Engineering
- Computer Science Applications
- Computer Graphics and Computer-Aided Design
- Computational Theory and Mathematics
Cite this
Vision-based automated bridge component recognition with high-level scene consistency. / Narazaki, Yasutaka; Hoskere, Vedhus; Hoang, Tu A.; Fujino, Yozo; Sakurai, Akito; Spencer, Billie F.
In: Computer-Aided Civil and Infrastructure Engineering, 01.01.2019.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Vision-based automated bridge component recognition with high-level scene consistency
AU - Narazaki, Yasutaka
AU - Hoskere, Vedhus
AU - Hoang, Tu A.
AU - Fujino, Yozo
AU - Sakurai, Akito
AU - Spencer, Billie F.
PY - 2019/1/1
Y1 - 2019/1/1
N2 - This research investigates vision-based automated bridge component recognition, which is critical for automating visual inspection of bridges during initial response after earthquakes. Semantic segmentation algorithms with up to 45 convolutional layers are applied to recognize bridge components from images of complex scenes. One of the challenges in such scenarios is to get the recognition results consistent with high-level scene structure using limited amount of training data. To impose the high-level scene consistency, this research combines 10-class scene classification and 5-class bridge component classification. Three approaches are investigated to combine scene classification results into bridge component classification: (a) naïve configuration, (b) parallel configuration, and (c) sequential configuration of classifiers. The proposed approaches, sequential configuration in particular, are demonstrated to be effective in recognizing bridge components in complex scenes, showing less than 1% of accuracy loss from the naïve/parallel configuration for bridge images, and less than 1% false positives for the nonbridge images.
AB - This research investigates vision-based automated bridge component recognition, which is critical for automating visual inspection of bridges during initial response after earthquakes. Semantic segmentation algorithms with up to 45 convolutional layers are applied to recognize bridge components from images of complex scenes. One of the challenges in such scenarios is to get the recognition results consistent with high-level scene structure using limited amount of training data. To impose the high-level scene consistency, this research combines 10-class scene classification and 5-class bridge component classification. Three approaches are investigated to combine scene classification results into bridge component classification: (a) naïve configuration, (b) parallel configuration, and (c) sequential configuration of classifiers. The proposed approaches, sequential configuration in particular, are demonstrated to be effective in recognizing bridge components in complex scenes, showing less than 1% of accuracy loss from the naïve/parallel configuration for bridge images, and less than 1% false positives for the nonbridge images.
UR - http://www.scopus.com/inward/record.url?scp=85074359524&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85074359524&partnerID=8YFLogxK
U2 - 10.1111/mice.12505
DO - 10.1111/mice.12505
M3 - Article
AN - SCOPUS:85074359524
JO - Computer-Aided Civil and Infrastructure Engineering
JF - Computer-Aided Civil and Infrastructure Engineering
SN - 1093-9687
ER -