Research On Construction Quality Supervision Technology Of Sleeve Grouting Based On Image Recognition

In prefabricated buildings,the quality of the connection nodes is related to the overall stability and seismic performance.The sleeve grouting connection method has become the most commonly used node connection method in China.With the continuous promotion of prefabricated buildings,the quality of the sleeve grouting connection gradually Get people’s attention.However,at present,the supervision of the construction process of the assembled concrete structure sleeve grouting connection site is insufficient and there is no effective and convenient supervision method,so that the quality problems at the construction site are endless,and it is difficult to accurately meet the design requirements.Based on this,this paper proposes a method of using image recognition technology to supervise the operation of workers during the grouting construction process to achieve grouting quality control.The main research work is as follows:First,the factors affecting the quality of the fabricated building sleeve grouting are studied.Through literature reading and field investigation,the common quality problems of the current assembly building sleeve grouting are found,and on this basis,the index system of the factors affecting the quality of the sleeve grouting is determined through the methods of literature search,standard research and expert interviews,and the level of application The analysis method determines the weight of each influencing factor,and finds out that the key influencing factors are personnel factors and process factors.Secondly,a technical solution for sleeve grouting supervision based on image recognition was designed.Based on the key influencing factors,the main points of sleeve grouting quality control were proposed,the limitations of the existing quality supervision technology were analyzed,the sleeve grouting quality supervisiontechnology based on image recognition was proposed,and the front-end video surveillance mid-end data transmission terminal platform model processing and design were designed.The intelligent supervision scheme of control,and based on the key control process of grouting and the analysis model of project progress,a warning model for the supervision of workers’ grouting process is established.Finally,based on the 3D residual network,grouting action training was carried out.The 185 grouting constructions collected at the scene were edited into 309 videos containing five construction actions,and the construction video clips were converted into frame-by-frame picture data using FFMPEG,a total of about 60,000 pictures,forming a mix,pull out A grouting action data set of five types of actions: a mixer,lifting a mixing barrel,inserting a cork,and pulling out a grouting gun,and using 3D residual network to train the data set to form a sleeve-based grouting quality supervision process based on image recognition.Engineering examples were applied.The results show that the trained recognition program can more accurately identify the pre-trained grouting action,and the established early warning model for workers’ grouting process supervision can also accurately determine the standardization of grouting construction operations.The sleeve grouting based on image recognition technology The quality supervision technology is practical.