Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

 
 
Session Overview
Session
Data Sensing & Acquisition 1
Time:
Monday, 26/July/2021:
11:00am - 12:00pm

Session Chair: Blanca Tejedor Herrán
Session Chair: Alessandro Carbonari

Subtopic: Computer Vision


Show help for 'Increase or decrease the abstract text size'
Presentations
11:00am - 11:20am

Joint Detection And Activity Recognition Of Construction Workers Using Convolutional Neural Networks*

Ghazaleh Torabi, Amin Hammad, Nizar Bouguila

Concordia University, Canada

Manually gathering information about activities on construction sites for project management purposes is labor-intensive and time-consuming. As a result, several works leveraged the already installed surveillance cameras to automate this process. However, the recent learning-based methods discretize continuous activities by assigning a single label to multiple consecutive frames. They do not fully leverage the contextual cues in the scene, and are not optimized end-to-end. A variation of the YOWO network, called YOWO53, is proposed in this paper to address these limitations. YOWO53 shows better classification and detection results over YOWO and allows using smaller input frames with real-time speed.



11:20am - 11:40am

Efficient Vertical Object Detection in Large High-Quality Point Clouds of Construction Sites*

Miguel Vega1, Alexander Braun1, Heiko Bauer2, Florian Noichl1, André Borrmann1

1Chair of Computational Modeling and Simulation, Technical University of Munich, Germany; 2FARO EUROPE GmbH & Co.KG., Korntal-Münchingen, Germany

Even when adherence to the project schedule is a critical performance metric, still 53% of construction projects exhibit schedule delays. To contribute to efficient construction progress monitoring, a method is proposed to detect temporary objects in scans of construction sites. The proposed workflow includes: image processing, computer vision, and deep learning techniques. The method was tested on three real scans and with three object categories (cranes, scaffolds, and formwork). It achieved average rates above 88% for precision and recall and outstanding computational performance (1s to process 10^5points). These metrics demonstrate the method’s capability to segment point clouds of construction sites.



11:40am - 12:00pm

Towards real-time Scan-versus-BIM :methods applications and challenges

Marcus Wallbaum1,2, Ranjith K. Soman1

1Centre for Systems Engineering and Innovation, Imperial College London, United Kingdom; 2The Alan Turing Institute, United Kingdom

There has been much work on Scan-vs-BIM, and it has implications on construction productivity, quality, and safety. However, these methods need to be extended /altered to do real-time Scan-vs-BIM. This paper presents the extensions for existing methods such as registration, point matching, object detection, and pose estimation to cater to real-time Scan-vs-BIM. Further, we describe the applications for such methods in construction and de-scribes the challenges for their implementation. We con-clude by implication of this paper on researchers working on augmented reality and construction robotics.



 
Contact and Legal Notice · Contact Address:
Privacy Statement · Conference: EC³ 2021
Conference Software - ConfTool Pro 2.6.141+TC
© 2001 - 2021 by Dr. H. Weinreich, Hamburg, Germany