标题：A NOVEL VISION BASED WARNING SYSTEM FORPROACTIVE PREVENTION OF ACCIDENTS INDUCED BYFALLING OBJECTS BASED ON BUILDING ANDENVIRONMENTAL ENGINEERING REQUIREMENTS
摘要：As one of the most dangerous industries, how to ensure the personal safety of on-site workers and reduce the casualty rate in the environmental construction industry remains a serious issue. In order to reduce accidents caused by falling objects at construction sites, this paper proposes a vision-based warning system on the basis of target tracking and monocular vision measurements. Different from traditional methods of accident prevention, such as optim ization in the construction environment and equipment, the proposed system prevents accidents from the perspective of the workers monitoring. W hile target tracking tracks field workers and extracts the coordinates of their position on an image, the monocular vision measurement is used to convert the position coordinates into the world coordinates on those images. Tests on a construction site verified that the system exhibits good performance in terms of tracking and system operation. The results show that the tracking success rate fluctuates between 90% and 95%, and the operational accuracy is about 70%. The main contribution to the body of knowledge is to provide an intelligent management concept for the safety management of accidents by falling objects (AFO) and to put forward a systematic framework for realizing the intelligent management of AFO. The novel work can effectively protect the safety of construction workers, reduce accidents, and u ltimately effectively maintain the building & engineering environment.
