Automated Data Acquisition System to Assess Construction Worker Performance

TLDR

The study introduces a real‑time, fully automated system that detects construction workers and tracks their movements to measure tool time and assess productivity. The system employs signal‑processing of digital video, audio, and thermal imagery, using hard‑hat detection, infrared cameras, and microphones to track workers, distinguish work status, and identify tool sounds. The system’s real‑time data enables measurement of tool time and productivity, providing project managers with insights to better plan and optimize labor and crew allocation.

Abstract

This paper presents a real time and fully automated system using signal processing techniques for data extraction from digital video images, audio files and thermal images of construction work activities. This data extraction system is developed to detect the construction workers and their movements within a given work area to measure tool time and to assess worker productivity. The worker tracking system is based on the characteristics of the hardhat extracted from digital videos to differentiate worker from others such as supervisors and engineers. Furthermore, the work status of the worker is tracked using Infrared cameras and unidirectional microphones. Thermal images extracted from infrared cameras recognize whether the worker is working or idle. The audio files from the microphones identify the sound wave patterns of the worker tools. This real time and automated information is expected to measure the tool time and productivity of a given work within a specific time period. The information extracted and analysed from the system will certainly aid the project managers to better plan to optimize labor and crews to achieve the expected productivity.