Correction Method Of Construction Workers' Unsafe Behavior Based On BBS

At present,the safety situation of construction projects at home and abroad is severe,and the unsafe behaviors of workers causing safety accidents are widespread.Traditional methods such as pre-existing safety training cannot effectively prevent the occurrence of unsafe behavior of workers,making the new method of unsafe behavior of construction workers correct in real time become the field of research that foreign scholars are concerned about.Through literature review,compared to safety atmosphere and cognitive science,behavioral safety directly concerns observable behaviors and timely corrects unsafe behaviors,it has been successfully applied and can be widely combined with field devices.It's the research perspective of construction workers' unsafe behavior correction.Furthermore,the current correction methods are reviewed,and the defects of the traditional methods are found,so the new methods are still to be researched and developed.Based on the behavior-based safety(BBS)and the unsafe behavior of construction workers,this paper designs an unsafe behavior correction model to correct the unsafe behavior of construction workers in three steps: scene analysis---behavior identification---behavior intervention.The first step is the scene analysis of construction workers' unsafe behavior,referring to the the analysis of the scenes when unsafe behavior occurs in the construction safety accident investigation report.The main methods involved are text characteristic mining and text association rule mining,which are carried out in two stages.The first stage: based on the analysis of unsafe behavior scenarios,using the method of text characteristic lexicon mining,mining scene elements,excavating the accident occurrence site of the 1430 construction safety accident investigation report,the unsafe behavior of the person causing the accident,the three types of characteristic words of the accident category and forming the characteristic lexicon,forming the initial construction worker list of unsafe behaviors,a total of 200.The second stage: using the association rule mining method,mining the association rules of the characteristic lexicon,mining the strong association rules between the characteristic lexicon,that is,what unsafe behaviors are high in each part and what kind of accidents will be triggered,and then the occasional behaviors are eliminated.The incidents in the list of unsafe behaviors form a list of unsafe acts of construction workers with strong association rules,a total of 27.The second step is the behavior identification of construction workers' unsafe behavior,refering to the list of the strong association rules construction workers' unsafe behaviors formed in the previous step.The process of realtime identification and judgment of the workers' behavior on site is based on computer vision technology,which are carried out in three stages.The first stage: designing construction workers' unsafe behavior recognition methods,which are divided into parts identification and behavior recognition.For the former,by dividing the dangerous areas of each part---the arrangement and setting of the cameras---the manual selection of the identification area in the video implementation,it is achieved by identifying visual representations of behavior and forming a list of construction workers' unsafe behavior statements.The second stage: using the current popular deep learning method,using the Python language in the TensorFlow framework,designing the behavior recognition algorithm,and identifying the unsafe behavior video containing a worker near the edge and staying on the edge.The third stage: using the Gensim algorithm to calculate the similarity between the recognition result statement of the previous stage output and the a list of construction workers' unsafe behavior statements,and then determine if the behavior is safe,and propose each safe suggeations according to the standard and safety regulations.Security recommendations for unsafe behavior,and finally design an information storage solution for video and recognition results.The third step is the behavior intervention of construction workers' unsafe behavior,refering to the process of promptly preventing unsafe behavior according to the behavior identification and judgment results in the previous step.The main methods involved are system designing and questionnaire survey,which are carried out in two stages.The first stage: designing the informationization intervention framework for construction workers' unsafe behavior,the core of which is the information security alarm system for unsafe behavior,and then analyze the functional requirements and performance requirements of the system.The second stage: because the limitations of hardware and software and professional aspects,this article does not study the specific development of the system,but to explor the technical acceptance survey in the early stage of the system development.Based on the TAM2 model,the technical acceptance model of this paper is constructed,and the variables are hypothesized.Based on this,the questionnaire is designed for investigation.According to the survey data,the structural equation model is used to analyze the correlation between variables,and the hypothesis is deleted to obtain the final model.The most significant factors affecting perceived usefulness and use intention of the informational alarm system are technology value and manager support.Finally,this paper puts forward the implementation points of the construction workers' unsafe behavior correction process from three aspects: scene analysis,behavior identification,behavior intervention.It also expounds the main conclusions,main innovations,research deficiencies and prospects of this paper.The significance of this paper is significant,which can effectively reduce the occurrence of building safety accidents,provide reference for the development of relevant correction tools,expand the application of behavior safety theory in construction safety management,and provide new ideas for the safe operation management of construction site workers.