Acknowledgement
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science and ICT (No. 2021R1A2C2004308) and the Ministry of Education (No. 2018R1A6A1A08025348).
International conference on construction engineering and project management (국제학술발표논문집)
- 2022.06a
- /
- Pages.871-876
- /
- 2022
- /
- 2508-9048(eISSN)
Automated Analysis of Scaffold Joint Installation Status of UAV-Acquired Images
- Paik, Sunwoong (Department of Civil and Environmental Engineering, Yonsei University) ;
- Kim, Yohan (Department of Civil and Environmental Engineering, Yonsei University) ;
- Kim, Juhyeon (Department of Civil and Environmental Engineering, Yonsei University) ;
- Kim, Hyoungkwan (Department of Civil and Environmental Engineering, Yonsei University)
- Published : 2022.06.20
Abstract
In the construction industry, fatal accidents related to scaffolds frequently occur. To prevent such accidents, scaffolds should be carefully monitored for their safety status. However, manual observation of scaffolds is time-consuming and labor-intensive. This paper proposes a method that automatically analyzes the installation status of scaffold joints based on images acquired from a Unmanned Aerial Vehicle (UAV). Using a deep learning-based object detection algorithm (YOLOv5), scaffold joints and joint components are detected. Based on the detection result, a two-stage rule-based classifier is used to analyze the joint installation status. Experimental results show that joints can be classified as safe or unsafe with 98.2 % and 85.7 % F1-scores, respectively. These results indicate that the proposed method can effectively analyze the joint installation status in UAV-acquired scaffold images.