Utilizing Computer Vision And Fuzzy Inference To Evaluate Level Of Collision Safety For Workers And Equipment In A Dynamic Environment

Complex construction environment and many dangerous factors are the important leading causes of fatalities and injuries in the construction site.The existing management method for detecting workers' unsafe behaviors and unsafe states of objects relies primarily on manual monitoring,which does not only consume large amounts of time and money but also cannot cover all workers in the entire construction site.Meanwhile,a common industry problem is that the identification of collision hazards largely depends on previous subjective coordination by safety supervisors or empirical analysis of past accidents,which can not objectively and quantitatively analyze the dynamic interaction process between the workers and the working environment.In addition,these data(such as site observation,construction log,etc.)cannot provide comprehensive and sufficient information to describe the risk factors of future accidents.Specifically,the workers' perception of being at risk of injury decreases when they are concentrated in a crowded and noisy environment.In this case,it is difficult for them to take essential measures to protect themselves in the face of danger.However,the rapid development of intelligent assessment and monitoring technology provides a new perspective and method for safety management of construction site.In view of the above issues,this study proposes a method of evaluating the collision safety level of construction workers based on computer vision and fuzzy inference.The specific contents and conclusions are as follows:(1)From the perspective of personnel,environment and equipment,the relationships between collision accidents and the proximity of equipment to workers and crowdedness of a target entity in a certain range were analyzed based on the Track Crossing Theory.(2)Using intelligent image recognition technology to identify and extract entity information in field environment.In evaluation experiments,the evaluation system framework of Faster R-CNN was built based on TP,FP and FN.The test experiment results demonstrate that the overall precision and recall for identifying the equipment are up to 95.53% and 92.67%,respectively.The statistical results also indicate that the said rates for the workers are approximately 94.09% and 91.49%,respectively.The network takes the average time to test an image is 0.25 s.Thus,the application potential of the faster R-CNN model for worker and equipment detection,which provides a strong support for the follow-up of worker's safety assessment,is verified.(3)In the safety assessment,the fuzzy inference system is used to assess the safety level of workers.The safety level,which indicates the workers' comprehensive collision risks of moving Human-Device in a particular environment,will be displayed numerically,breaking through the limitations of conventional qualitative evaluation.Meanwhile,by setting a safety level threshold,the onsite safety management personnel can take corresponding measures to avoid collision accidents when the worker's safety level is lower than the threshold.(4)Building an early warning framework based on mobile camera and promoting the use of cameras as a decision-making tool for construction safety management through image recognition technology.Using intelligent methods to replace the conventional manual monitoring,which lays a solid foundation for real-time early warning.The experiment is demonstrated that the method can be utilized to evaluate the workers' safety level and provide a new concept for the prevention of collision accident.