A worker posture coding scheme to link automatic and manual coding

Hainan Chen, Xiaowei Luo*, Zhenhua Zhu

*Corresponding author for this work

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

6 Citations (Scopus)

Abstract

Real-time and automatic monitoring of worker behaviors and activities have great potential to improve construction job site operation. Traditional behavior monitoring of construction workers relies on human interpretation to determine workers' semantic conditions (e.g., tasks performing, safety status). Although advanced sensing technologies provide more accurate quantitative data on worker behavior, how to effectively link the data to a worker's semantic condition in a form that is understandable for humans remains a challenge. This paper proposed a novel posture coding scheme based on the worker's body part relative position (BPRP) information. The proposed coding scheme compresses the quantitative 3D skeleton data into qualitative posture descriptions but keeps the body part relative space information. Afterward, an indoor motion test is conducted to validate the reliability of the proposed BPBR coding scheme. The test results showed that by employing the BPRP coding scheme, the manual and automatic posture coding could achieve consistent results. Therefore, the manual posture coding results can be transformed into human skeleton figures and then further processed by the quantitative algorithms. Correspondingly, the computer-captured human skeleton data can be easily connected to the manual observation results by interpreting the BPRB codes
Original languageEnglish
Article number103630
JournalAutomation in Construction
Volume125
Online published19 Feb 2021
DOIs
Publication statusPublished - May 2021

Research Keywords

  • Coding consistency
  • Human body posture
  • Manual and automatic body posture coding connection

Fingerprint

Dive into the research topics of 'A worker posture coding scheme to link automatic and manual coding'. Together they form a unique fingerprint.

Cite this