Image segmentation of underfloor scenes using a mask regions convolutional neural network with two-stage transfer learning

Atkinson, Gary A.; Zhang, Wenhao; Hansen, Mark F.; Holloway, Mathew L.; Napier, Ashley A.

Abstract

© 2020 Elsevier B.V. Enclosed spaces are common in built structures but pose a challenge to many forms of manual or robotic surveying and maintenance tasks. Part of this challenge is to train robot systems to understand their environment without human intervention. This paper presents a method to automatically classify features within a closed void using deep learning. Specifically, the paper considers a robot placed under floorboards for the purpose of autonomously surveying the underfloor void. The robot uses images captured using an RGB camera to identify regions such as floorboards, joists, air vents and pipework. The paper first presents a standard mask regions convolutional neural network approach, which gives modest performance. The method is then enhanced using a two-stage transfer learning approach with an existing dataset for interior scenes. The conclusion from this work is that, even with limited training data, it is possible to automatically detect many common features of such areas.

• Journal Article Type: Article
• Acceptance Date: Feb 1, 2020
• Online Publication Date: Feb 13, 2020
• Publication Date: May 1, 2020
• Deposit Date: May 6, 2020
• Publicly Available Date: Feb 14, 2021
• Journal: Automation in Construction
• Print ISSN: 0926-5805
• Electronic ISSN: 1872-7891
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 113
• Article Number: 103118
• DOI: https://doi.org/10.1016/j.autcon.2020.103118
• Keywords: Control and Systems Engineering; Civil and Structural Engineering; Building and Construction; Computer vision; Underfloor maintenance; Convolutional neural network
• Public URL: https://uwe-repository.worktribe.com/output/5963692
• Additional Information: This article is maintained by: Elsevier; Article Title: Image segmentation of underfloor scenes using a mask regions convolutional neural network with two-stage transfer learning; Journal Title: Automation in Construction; CrossRef DOI link to publisher maintained version: https://doi.org/10.1016/j.autcon.2020.103118; Content Type: article; Copyright: © 2020 Elsevier B.V. All rights reserved.