Design And Implementation Of Construction Worker Safety Protection System

The construction industry is known as one of the pillar industries in the period of China’s socialist modernization.The sustainable development of the construction industry and safty production have a bearing on the vital interests of all citizens.In recent years,there have been major casualties in China’s construction projects.The number of accidental casualties has remained high.How to prevent and reduce construction accidents has attracted more and more attention.Due to the diversity and complexity of the construction process at the construction site,the high number of high-altitude operations,the overall quality of the construction workers,and the diversification of the construction market,coupled with the uncertainties that cannot be avoided during construction and production of the site,and operations.Due to the changing status of the process and the construction environment,potential hidden dangers are increasing.If it is handled improperly,it can easily cause traffic accidents,it will also cause people’s property damage,and even cause social turmoil.According to relevant information,accidents such as heavy objects fell,workers falling from high altitude even happening everyday.Casualties have been high for many years and are called 4 major injuries.In addition,the dust on the building construction site and the noisy environment have led to a high prevalence of cardiovascular,respiratory and other diseases among construction workers.Therefore,if we want to fundamentally reduce the occurrence of casualties in construction projects and protect the health of workers,we must take appropriate measures and methods.By analyzing the collected data,combined with workers’ accidental postures such as consciously or unconsciously taking off helmets,picking up heavy objects,falling and wrestling,falling from high altitudes,etc.,these construction accidents are judged and analyzed using RBF neural network in machine learning algorithms.RBF is a three-layer feedforward neural network,including input layer,hidden layer,and output layer.The network center and network width of the learning hidden layer are unsupervised learning processes;learning the connection weights from the hidden layer to the output layer is a supervised learning process.The supervised learning representative training data has marker information,that is,the results of training examples.In our model,there are ADL,forward falling,backward falling,side falling,hit and falling,and irregular falling.The experimental results prove the effectiveness of the algorithm’s improvement strategy.It has greatly improved the classification accuracy and the computational efficiency.It breaks through the bottleneck that construction workers cannot monitor the personal safety of workers and provides real-time monitoring and protection of technological innovation points for workers.It provides an additional layer of protection for the safety of construction workers and can make contributions to the safety of construction sites.When the workers are monitored for accidents,they can be fed back into the system for the first time.After investigation,it has been found that in many cases,trips,falls,wrestling,falling from high places,etc.cannot be discovered in many cases,even if discovered by others.Many cases miss the best treatment time for workers.At present,there are more and more anti-fall devices for the elderly persons on the market.Most of the technical realization is to set the threshold of acceleration and angular velocity collected by sensors,which will give an alarm when exceeding the threshold.Simple threshold setting can not meet the complexity of workers’ action on the construction site.So RBF neural network is used in the analysis of the algorithm in this paper.The author applies RBF neural network to worker accident gesture recognition.The experimental results prove the availability and accuracy of the algorithm in the field of accident gesture recognition.Then the paper analyzes the original algorithm based on the actual conditions of the construction site,points out the shortcomings and proposes an algorithm improvement strategy.In the testing phase,we collected 500 sampling data,24 hours × 7 days running modules,and safety accident attitude data to verify the effectiveness of the method in this system.In addition,compared with other algorithms such as Two-Stream CNN fall detection method and multi-weight neural network fall detection method,the experimental results show that RBF neural network has more advantageous in human fall applications.