Font Size: a A A

Research And Application Of Construction Status Supervision On Construction Sites Based On Deep Learning

Posted on:2021-03-07Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q HaoFull Text:PDF
GTID:2432330605979828Subject:Engineering
Abstract/Summary:PDF Full Text Request
In recent years,with the continuous exploration and research on the concept of deep learning and the continuous maturity of image recognition technology based on deep learning.researchers have applied the recognition technology to various fields.On the other hand,with the acceleration of urban construction,the phenomenon of insufficient strength in the process of supervision is becoming more and more prominent:the construction period of the projects under construction is long,there are more and more new construction site projects.resulting in the increasing number of projects under construction at the same time.In the construction process,because of procedures and other reasons,the project is also gradually increased,the construction of such construction site supervision is particularly important.This topic through the use of image recognition technology to determine the construction site construction state,to solve the above problemsIt is understood that in the current urban planning supervision,the construction status of illegal projects is supervised mainly by watching videos online and patrolling offline.The two supervision methods cost a lot of manpower and material resources and work efficiency is low.In order to solve this problem,the VGG model of convolutional neural network is adopted in this project.Through digital image processing technology,the image collected on the construction site is enhanced for sample data and image smooth pretreatment,so that it can meet the training model standards and greatly improve the accuracy of recognition.Target detection algorithm and migration learning method are used to improve the efficiency of helmet wear recognition.The model is trained and invoked by deep learning neural network to complete image recognition module and contrast algorithm module.This method can effectively improve the work efficiency of construction status supervisors on construction sites,and at the same time effectively improve the accuracy of identification,which is more intelligent and efficient than the traditional way of supervision by watching videos.This paper provides a reference method for safety helmet wearing monitoring on construction site.
Keywords/Search Tags:Deep Learning, Convolutional Neural Network, Helmet identification
PDF Full Text Request
Related items