Research on Campus Behavior Detection Technology for Elementary and Middle School Students Based on YOLOv8 and OpenPose

Aiming at the current challenges in detecting school behaviors of primary and secondary school students, this paper proposes an innovative fusion method. Based on YOLOv8 and OpenPose algorithms, the method integrates open source datasets with homemade datasets to construct a comprehensive dataset covering classroom behaviors and school bullying behaviors. By applying the YOLOv8 algorithm, the preliminary identification of classroom behaviors such as raising hands, reading, and writing, as well as school bullying behaviors such as fighting, bullying, and isolation, is achieved. To further improve the detection accuracy, this paper introduces a customized attention mechanism in the YOLOv8 model and optimizes the loss function. At the same time, the OpenPose algorithm is combined to extract the key points of the human body for secondary identification and classification of behaviors. Experimental results show that this method has significantly improved the accuracy and reliability of campus behavior detection, and it is of great significance for improving the quality of classroom teaching and ensuring campus safety.