RUB arrow Bauwesen arrow Informatik arrow




Progress monitoring of finishing works based on videos and BIM

Unexpected delays and disruptions in the interior finishing phase of construction frequently lead to extensive re-scheduling. To avoid manual, thus labor-intensive and time-consuming on-site inspections, researchers have proposed several methods to automate the process of progress monitoring. However, existing approaches predominantly deal with outdoor sites, usually focus on accurate geometry reconstruction, but lack indoor progress estimation and delay prediction. In this paper we present a conceptual framework that can potentially predict finishing delays from as-performed video data and as-designed BIM data using computer vision algorithms and construction simulation. As a result, the criticality of each activity’s delay can be marked on the updated BIM to easily identify the objects and areas with the highest impact on the project schedule. Moreover, the updated schedule can also be used to make informed decisions about resources in the on-going project.

Background and Problem Statement

  • Interior finishing covers a big part
  • Unexpected delays have a big influence on project’s performance
  • Lack of frequent finishing progress inspection induce late problem recognition
  • Current practice includes manual progress monitoring involving a lot of time
  • Need for higher degree of automation in progress monitoring
Figure 1: Relation between existing approaches using 4D BIM or vision based state recognition and progress monitoring
  • BIM-based progress monitoring
    • Only few construction
    • activities
    • Mainly target building shell
    • Interior activities not regarded
  • Vision-based state recognition
    • Not 4D BIM related
    • Objects not identifiable
  • Formalized and generic framework    
    • link vision-based state recognition with 4D BIM progress monitoring


Figure 2: Integration of the visual state recognition within progress monitoring
  • Register image frames in BIM
    • Initial absolute registration
    • Relative motion estimation
  • Filter relevant information
    • Select relevant objects
    • Compose image subset
    • Reduction to regions of interest
  • Identification of finishing objects
    • Frequent re-occurance of objects
    • Identification needed and possible
  • Delay prediction
    • Degree of completeness
    • Single tasks / project delay
    • Construction schedule pdate


  • Higher degree of automation in progress monitoring achievable
    • BIM as a rich information source
    • Image registration in BIM is key
  • Delay prediction possible
    • Frequent inspections reveal deviations
    • Time and financial effort reduced

Project data

Title: Progress monitoring of finishing works based on videos and BIM

Type: International collaboration with Dr. Brilakis (University of Cambridge)

Period: Seit 08/2011

Researcher:Christopher Kropp, M.Sc., Dr.-Ing. Christian Koch

Chair of Computing in Engineering
Gebäude IC, Raum 6-65
Universitätsstraße 150
44780 Bochum



sponsored by