Video-based Identification Model for the Off-duty of Construction Guardian
In order to prevent potential safety hazards caused by guardians leaving their posts during construction operations and petrochemical production operations,the retrained YOLOv8 model was used to intelligently analyze the surveillance video of the construction site and identify the construction operation guardians'off-duty behavior.A set of construction operation guardian off-duty identification and judgment process was designed:accessed the construction video,used YOLOv8 to identify personnel,vests,construction markers,construction area markers,etc.,determined whether the construction operation is in progress,determined the construction area and the area where the guardian should be,the guardian was detected and judged whether the guardian is in the construction area.Once the guardian leaves the post,an alarm message was immediately sent to automatically determine whether there is a guardian in the construction area.In this process,the variable area recognition method was introduced into YOLOv8,which only recognizes targets within the area,avoiding the interference of targets outside the area on the judgment.The construction operation guardian off-duty identification model can accurately and efficiently identify the off-duty of guardians during construction.The identification success rate reached 90%,which meets the needs of construction safety supervision.
万方数据
维普
