2017
DOI: 10.1109/tits.2016.2568758
|View full text |Cite
|
Sign up to set email alerts
|

Deep Multitask Learning for Railway Track Inspection

Abstract: Railroad tracks need to be periodically inspected and monitored to ensure safe transportation. Automated track inspection using computer vision and pattern recognition methods have recently shown the potential to improve safety by allowing for more frequent inspections while reducing human errors. Achieving full automation is still very challenging due to the number of different possible failure modes as well as the broad range of image variations that can potentially trigger false alarms. Also, the number of … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
191
0
5

Year Published

2018
2018
2022
2022

Publication Types

Select...
7
2

Relationship

0
9

Authors

Journals

citations
Cited by 346 publications
(196 citation statements)
references
References 37 publications
(40 reference statements)
0
191
0
5
Order By: Relevance
“…Gibert et al. () designed a deep CNN, which was trained under a novel multitask learning framework, to detect defects on railway ties and fasteners. The aforementioned CNN‐based approaches mainly focus on detecting one type of defect.…”
Section: Literature Reviewmentioning
confidence: 99%
“…Gibert et al. () designed a deep CNN, which was trained under a novel multitask learning framework, to detect defects on railway ties and fasteners. The aforementioned CNN‐based approaches mainly focus on detecting one type of defect.…”
Section: Literature Reviewmentioning
confidence: 99%
“…Gibert et al. () and Cha et al. () introduced a customized convolutional neural network framework for identifying rail track cracks and slab cracks, respectively.…”
Section: Literature Reviewmentioning
confidence: 99%
“…In this study, the method developed by Gibert et al. () was employed here to generate such a mask. Finally, Figure f shows the results of the extracted cracks.…”
Section: Applications Of the Proposed Modelmentioning
confidence: 99%
“…Aytekin achieved real-time railway fastener inspection using a high-speed laser range finder camera and pixel and histogram similarity analysis [10]. As Deep Convolutional Neural Network (DCNN) [11] prevails in object recognition, Gibert et al [12] applied DCNN in railroad track detection. This multi-task learning system combined a 10-class track material classification detector (e.g., wood, concrete, and metal fasteners etc.)…”
Section: I! Introductionmentioning
confidence: 99%