Abstract:Currently, the incidence of accidents involving falls from height at construction sites caused by workers climbing guardrails is still high. Traditional unsafe behavior management mainly relies on a safety patrol of construction-site supervisors, which consumes considerable laborpower and time. There is still a critical need for an automated safety management method to identify unsafe guardrail climbing behavior. This study proposes a worker behavior identification method based on visual data fusion of a worker's surrounding environment and posture data. Videos of seven participants' guardrail climbing behavior through multiangle and multidistance cameras were analyzed to verify this method. By analyzing the environment and posture of the participants, three methods based on environment, posture, and fusion data were used to detect the stage of guardrail climbing action of the workers and compare them with the ground truth labeled by safety experts. The precision and recall of worker guardrail climbing behavior based on the fusion method were 82% and 83% respectively, which is better performance than that obtained using a single method. The data fusion-based method avoids the misjudgment generated by a single detection method and can identify the guardrail climbing behavior more accurately. Guardrail climbing is a typical unsafe behavior that exposes workers to a high risk of falling from height. However, there is a lack of research on the interaction between workers and guardrail systems in the construction industry. This study provides a nonintrusive method for automating detection and management of guardrail climbing behavior on construction site. Using existing surveillance cameras, this method can be deployed at low cost with slight interference with workers. Based on the detection, appropriate interventions are expected to effectively reduce workers' unsafe behaviors during construction and improve safety on site. The detection of guardrail climbing, which is one of the variety of unsafe behaviors associated with falls from height, can enrich the intelligent construction safety management system effectively. Moreover, this study also provides reference and quantitative indicators (e.g., a guardrail climbing unsafe behavior database) for risk assessment and early warning of workers who are exposed to risk of fall from height.

Recognition of Unsafe Behavior of Electrical Workers based on Spatial-Temporal Visual Attention

Intelligent Video Surveillance for Checking Attendance of Traffic Controllers in Level Crossing

Key frame extraction based on the visual attention model for lane surveillance video

Deep Learning Based Abnormal Behavior Detection for Elderly Healthcare Using Consumer Network Cameras

Computer vision and long short-term memory: Learning to predict unsafe behaviour in construction

Unsafe behavior identification on construction sites by combining computer vision and knowledge graph–based reasoning

Unsupervised anomalous behavior detection using spatial-temporal interest points

Worker Abnormal Behavior Recognition Based on Spatio-Temporal Graph Convolution and Attention Model

Vision-Based Detection of Unsafe Worker Guardrail Climbing Based on Posture and Instance Segmentation Data Fusion

Skeleton Based Abnormal Behavior Recognition Using Spatio-Temporal Convolution and Attention-Based LSTM

Intelligent approach for analyzing surveillance videos in urban emergency management

Abnormal Object Representation Based on Surprise Model

Learning Attention Augmented Spatial-temporal Normality for Video Anomaly Detection

Student Dangerous Behavior Detection in School

Multi-Scale Temporal Relations and Segmented Channel Attention for Video Anomaly Detection

Sequential Attention Mechanism for Weakly Supervised Video Anomaly Detection

Real-time smart video surveillance to manage safety: A case study of a transport mega-project

Abnormal visual event detection based on multi‐instance learning and autoregressive integrated moving average model in edge‐based Smart City surveillance

Image-and-Skeleton-Based Parameterized Approach to Real-Time Identification of Construction Workers' Unsafe Behaviors

An Unsafe Behavior Detection Method Based on Improved YOLO Framework

An Intelligent Video Analysis Method for Abnormal Event Detection in Intelligent Transportation Systems