Automatic MEP Component Detection with Deep Learning

John Kufuor, Dibya Mohanty, Enrique Valero Rodriguez, Frédéric Bosché

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Scan-to-BIM systems convert image and point cloud data into accurate 3D models of buildings. Research on Scan-to-BIM has largely focused on the automated identification of structural components. However, design and maintenance projects require information on a range of other assets including mechanical, electrical, and plumbing (MEP) components. This paper presents a deep learning solution that locates and labels MEP components in 360deg images and phone images, specifically sockets, switches and radiators. The classification and location data generated by this solution could add useful context to BIM models. The system developed for this project uses transfer learning to retrain a Faster Region-based Convolutional Neural Network (Faster R-CNN) for the MEP use case. The performance of the neural network across image formats is investigated. A dataset of 249 360deg images and 326 phone images was built to train the deep learning model. The Faster R-CNN achieved high precision and comparatively low recall across all image formats.
Original languageEnglish
Title of host publicationPattern Recognition
Subtitle of host publicationICPR International Workshops and Challenges
ISBN (Electronic)978-3-030-68787-8
ISBN (Print)978-3-030-68786-1
Publication statusPublished - 21 Feb 2021
EventPattern Recognition and Automation in Construction & the Built Environment - Milan, Italy
Duration: 10 Jan 2021 → …

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


WorkshopPattern Recognition and Automation in Construction & the Built Environment
Abbreviated titlePRAConBE
Period10/01/21 → …
Internet address


  • Scan-to-BIM
  • MEP
  • Radiators
  • Sockets
  • Switches
  • Convolutional Neural Network
  • Deep learning

Fingerprint Dive into the research topics of 'Automatic MEP Component Detection with Deep Learning'. Together they form a unique fingerprint.

Cite this