摘要
套筒灌浆质量影响着装配式建筑整体性能,故其现场管理至关重要。针对目前套筒灌浆中存在的灌浆过程、灌浆质量难以控制等问题,本文集成计算机视觉领域的目标识别与姿态检测算法,提出了基于工人操作工序的套筒灌浆过程监测方法。本方法利用改进YOLOv4算法识别灌浆施工工具与工人;利用OpenPose算法检测工人姿态,计算工人动作比(Worker Action Ratio,WAR);依据施工工具与WAR的判断结果判定工人所处工序,完成套筒灌浆的过程监测。实现了套筒灌浆过程的无人化监测。实验结果表明,该监测方法能够正确识别施工过程,完成灌浆过程监测,在13.64 fps速度时,改进YOLOv4算法的平均精度可达96.03%,基本满足套筒灌浆实时监测的要求。
Site management of sleeve grouting is important because its quality affects the overall performance of the prefabricated building. Aiming at difficult control of grouting quality and grouting process, the research integrates the target recognition and posture detection algorithms in the field of computer vision, and proposes a process monitoring method of sleeve grouting based on the worker’s operation procedures. This method uses the improved YOLOv4 algorithm is used to identify grouting construction tools and workers, and the OpenPose algorithm is used to detect the workers’ posture and calculate the Worker Action Ratio(WAR). According to the judgment of construction tools and WAR, the working procedure is determined to complete the process monitoring of sleeve grouting. Unnamed monitoring of sleeve grouting process is realized. The experimental results show that the monitoring method can correctly identify the construction process and complete the monitoring of grouting process. At the speed of 13.64 FPS, the average accuracy of the improved YOLOv4 algorithm can reach 96.03%. Basically meet the requirements of sleeve grouting real-time monitoring.
作者
丁静怡
王凌霄
袁竞峰
薛竣
DING Jingyi;WANG Lingxiao;YUAN Jingfeng;XUE Jun(School of Civil Engineering,Southeast University,Nanjing 211100,China;HuaXin Construction Engineering Group Co Ltd,Nantong 226600,China)
出处
《土木工程与管理学报》
2023年第1期136-144,共9页
Journal of Civil Engineering and Management
基金
海安市2019年度市级建筑产业现代化科技支撑项目(2020-10-22)。
