Research And Implementation Of Attention Monitoring System Based On Raspberry Pi And Head Pose Estimation

In traditional offline teaching,students' inattention is one of the important factors affecting teaching quality,but the observation of attention is a difficult task.The developing computer vision technology provides more possibilities to solve this problem.This dissertation implements a complete attention monitoring system based on the research of algorithms related to head pose estimation,integrating edge computing ideas,and based on open-source hardware and software for computer experimental teaching scenarios.The system consists of an acquisition part,a server part,and a monitoring terminal.The single acquisition part uses Raspberry Pi as the hardware,and develops core programs such as image acquisition,head posture estimation,and attention state determination based on Python language to realize the acquisition and processing of students' attention data;the server part serves as the data collection and service center of the system,receiving and managing attention data from multiple acquisition parts,and also providing attention monitoring services for the monitors through the Java EE-based server.The acquisition part and server part are based on WiFi and use Netty and Protobuf technologies to achieve efficient data communication;the monitoring terminal is the access portal for the monitors,and the students can be monitored by video and the state of attention can be visually monitored in real-time through a browser.Head pose estimation as the core of the system is implemented in two steps: firstly,face keypoint detection is done based on the ERT algorithm,then the EPnP algorithm is used to realize the solution of position and orientation on this basis.Besides,aiming at the problem that the average error of pose estimation is large in the actual test,the evaluation and correction mode of incorporating Kalman filter is established to optimize and improve the algorithm,which improves the accuracy of monitoring while ensuring the stability of operation of the acquisition part.This dissertation takes engineering application as the goal researched,designed and implemented the attention monitoring system is characterized by a clear structure,easy deployment,strong real-time performance,and high visualization,which can better realize the needs of monitoring students' attention in computer experimental teaching scenarios and has reference significance for the research of similar systems.