Intelligent Classroom System Design And Classroom Behavior Recognition Research

In recent years,the concept of smart education has been gradually known by people.As the carrier of the concept of smart education,smart classroom has also attracted the attention of the society.With the development of artificial intelligence,people’s functional requirements for smart classrooms are no longer limited to providing convenience for teachers and students’ teaching,but regard the classroom as a part of teaching.Intelligent algorithms enable the classroom to participate in the teaching process.Therefore,the combination of intelligent algorithm and smart classroom has become a key and difficult point in the development of smart classroom.According to the development trend and actual needs of smart classroom,this paper designs a smart classroom system based on behavior recognition algorithm.In addition to providing a good teaching environment for teachers and students,this system can also automatically check students’ attendance and assist teachers to improve the teaching quality by recognizing students’ class and examination status.Firstly,the paper analyzes the requirements of the smart classroom system.According to the functional requirements of the system proposed by the school,the system is divided into three parts: client-side,server-side and main control module,and the whole system is designed from two aspects: software and hardware.Among them,the client-side uses Vue and Element UI component combined with node.js technology to complete the design of the management system.The server-side takes Huawei Cloud server as the carrier and uses Spring Cloud Alibaba one-stop solution to complete the back-end development of the system,realizing the remote monitoring of classroom environment and the remote identification of students’ classroom behaviors,and improving the intellectualization of the system.The main control module is designed using IMX6 ULL and STM32 to realize the classroom electronic equipment control and classroom environment regulation function.According to the intelligent demand of intelligent classroom system,the paper research deeply on human behavior recognition algorithm.Based on the structural characteristics of Two-Stream convolutional neural network and 3D convolutional neural network,this paper designs a T-Fusion network model which strengthens the time characteristics of pseudo-3D convolutional network.Furthermore,it adopts the idea of 3D splitting network to combine residual network,T-Fusion network and 3D convolutional neural network.A spatial-temporal fusion pseudo-3D convolutional neural network model TFP3 D is designed to improve the network’s ability to extract video temporal features and global features.Finally,the paper tests the functions of the system,proving the practicability of the smart classroom system and meeting the functional requirements of the smart classroom system put forward by the school.