摘要
塔吊作业具有持续时间长、范围大以及与工人空间交互复杂的特点,导致塔吊事故频发且往往损失重大。为了提升塔吊运行相关危险场景检测的及时性和有效性,该文基于信息化手段,提出了塔吊运行过程中危险场景的自动检测方法。该方法基于建筑信息模型(BIM)、传感器和计算机视觉等技术,通过获取塔吊运行状态与工人作业状态,分析两者之间的空间交互关系,在判别规则的基础上实现了危险场景自动检测。案例应用表明,该自动检测方法是可行的和有效的。
[Objective] Tower crane operations are characterized by long durations,extensive moving scopes,heavy loads,and complex spatial interactions with workers.These factors often contribute to construction accidents.Furthermore,construction workers standing under crane hooks and their lifting objects during the lifting process pose high safety risks,often encountering accidents such as collisions and object falling.Information technology plays a crucial role in enhancing tower crane monitoring and reducing workers' safety risks.Although existing studies on tower crane monitoring have made considerable advancements,they primarily focus on the operating state of cranes and overlook safety issues arising from interactions between cranes and workers.This study aims to employ the schedule information extracted from building information modeling(BIM) and computer vision and sensing technologies to propose an automatic hazard detection method for detecting dangerous scenarios during the lifting process in tower cranes.[Methods] This study develops an automatic detection framework for identifying hazardous scenarios involving spatial interaction between tower cranes and workers.This framework comprises four components.(1) Equipment installation and network environment establishment:cameras are installed at elevated positions to monitor the spatial locations of workers under the operating plane of a tower crane in real time.Furthermore,various sensors and cameras are fixed beneath the crane's trolley and cab to collect data regarding its operating status.A local area network is set up on the site to facilitate instantaneous data transmission.(2) Collection of tower crane operating data:the exact spatial location of the crane's hook is calculated using arm tracking and spatial trigonometric relations to determine its operating status.(3) Collection of workers' operational status data:advanced image recognition techniques are used to identify workers' positions,which are then converted into three-dimensional spatial coordinates through coordinate transformation.(4) Spatial relationship analysis and identification:precise spatial mapping of the tower crane's operating status and workers' positions is obtained using a unified BIM,followed by automatic detection according to predefined hazard assessment rules.[Results] The effectiveness and feasibility of the proposed method are validated by implanting it during a one-month real construction project.The analysis of data collected for 15 days reveals that the number of hazardous scenarios fluctuates considerably,peaking 523 times and plunging 35 times.These fluctuations correlate strongly with the number of workers on site,verifying the reliability of the proposed method and highlighting the need for intelligent hazardous scenario detection.Moreover,the results show that construction workers generally lack adequate awareness of the safety implications of tower crane trajectories.[Conclusions] This study successfully integrates BIM,sensing,and computer vision technologies to develop an automatic hazard detection method that focuses on the spatial interaction between tower cranes and workers.The proposed method enhances the timeliness and accuracy of hazard detection and provides innovative perspectives and technical support for construction site safety management.However,this study has certain limitations,such as data interferences caused by minor vibrations during tower crane operations to be further mitigated using noise reduction techniques in future research.
作者
张知田
王园园
罗柱邦
郭子扬
郭红领
ZHANG Zhitian;WANG Yuanyuan;LUO Zhubang;GUO Ziyang;GUO Hongling(Department of Construction Management,Tsinghua University,Beijing 100084,China;Construction Engineering Branch,The Third Highway Engineering Co.,Ltd.,Beijing 100020,China)
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2024年第2期198-204,共7页
Journal of Tsinghua University(Science and Technology)
基金
国家自然科学基金面上项目(52278310)
清华大学“水木学者”计划。
关键词
施工安全
塔吊
工人
危险场景
自动检测
construction safety
tower crane
worker
hazardous scene
automatic detection
