Loading...

Monitoring Risks of Tower Crane Operations Using Computer Vision and Deep Learning Techniques

Pazari, Parham | 2023

19 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56759 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Alvanchi, Amin
  7. Abstract:
  8. The utilization of tower cranes at construction sites presents numerous inherent risks. These cranes are commonly employed for lifting heavy loads, which carry the potential hazard of accidental falls. Simultaneously, workers may inadvertently overlook overhead dangers while focusing on their tasks. To mitigate these risks, laws in many countries explicitly prohibit individuals from occupying the vicinity directly beneath suspended loads, known as the fall zone. Such measures are vital to safeguard against the peril of heavy loads plummeting onto people. However, existing studies have not offered a comprehensive and efficient approach to identify crane load fall zone. To address this gap, this study proposes a novel methodology that leverages computer vision and deep learning techniques for robust identification of tower crane load fall zones, as well as accurate determination of worker locations relative to these zones. The proposed system comprises four integral components. Firstly, a stereo camera setup is employed to extract precise depth information from the scene. Subsequently, the system detects crane loads based on their movement patterns and elevation. Thirdly, workers are detected using the YOLOv7 deep learning object detection algorithm. Lastly, the system employs depth data to compare the locations of the load fall zone and workers in the world coordinate system. Consequently, the system classifies each worker's position as either within the red zone (directly beneath the load), the yellow zone (two meters away from the red zone), or the green/safe zone (outside these two zones). Remarkably, the proposed method achieves an acceptable speed of 8 frames per second, while exhibiting over 90% accuracy in accurately determining the zone of workers. This renders the system highly reliable for promptly alerting the presence of individuals in the load fall zone and providing valuable information to enhance safety management
  9. Keywords:
  10. Computer Vision ; Deep Learning ; Construction Safety ; Tower Crane ; Load Falling ; Automatic Identification

 Digital Object List

 Bookmark

No TOC