基于计算机视觉的课堂情况分析系统

打开文本图片集

中图分类号：TP391;TP399 文献标志码：A 文章编号：1674-2605（2025）05-0006-07DOI： 10.12475/aie.20250506 开放获取

Classroom Situation Analysis System Based on Computer Vision

WANG Yongqiang LUO Hongye ZHAO Xian （Guizhou Electronic Technology College， Guiyang 55ooo3， China）

Abstract： Aimingatcurrent problems inuiversitiessuchaslarge classizessusceptibilitytoatendance fraud，andteachers' difculty in promptlyunderstanding eachstudent'slearing state，this paperproposesaclassroomsituationanalysissystembasedon computervision.Thesystemutlizesthemulti-taskcascadedconvolutionalneuralnetwork（MCNN）todetectfacesfromtheealtimevideostreamofclassroom studentscaptured byacamera.Itemploysthefacerecognition model toextractfacialfeatures for studentidentityverification.Keypointdetectionalgorithms andtheYOLOv10object detectionalgorithmareused todetectstudent behaviorslikeslepingandusingmobilephones inclassrespectively.Dynamic thresholdsandastatepersistence mechanismare introduced toavoid misjudgmentof transientbehaviors，andthe thresholds for detecting slepingand phone usagebehaviors are dynamically adjusted.Testresultsshow that thesystemcanaccurately perform functions including automatic clasroomattendance， abnormalbeaviordetetionisalzatiooftedacersults，ndutomaticgneratinoflassroomsiatioaalysiseprtss helpsteachersgaspstudents'learingstatesineal-time，alowingthemtoadjust teachingstrategies promptly，terebyiproving teaching quality.

Keywords： computer vision; clasroom situationanalysis; facerecognition; abnomalbehaviordetection; dynamic hreshold; YOLOv10; multi-task cascaded convolutional neural network

0 引言

课堂管理是教育教学中的一项基本工作，也是提高教学质量的重要环节。（剩余6168字）