Performance Evaluation of Deep Learning Algorithms in Intelligent Campus Security Monitoring and Management System

Performance Evaluation of Deep Learning Algorithms in Intelligent Campus Security Monitoring and Management System

연구 분야: Infrastructure

학회: ASENS '24: Proceedings of the International Conference on Algorithms, Software Engineering, and Network Security

With the increase of campus education scale and the improvement of work quality, the demand for school safety work is also increasing. In the process of school safety construction, there are many problems such as weak student safety awareness, need to strengthen safety management, and incomplete systems. This article applies the concept of deep learning (DL) to the analysis and prevention of safety accident modes, based on DL algorithms, with the goal of achieving target tasks within the monitoring area and detecting their motion status. In the motion target trajectory prediction module, the motion target trajectory prediction module is based on the principle of DL algorithms to monitor the surrounding environment of the monitored person in real-time, obtain information such as the position, speed, and direction of the monitored person, and predict their movement direction information. The accuracy of classification tasks includes the accuracy of student behavior recognition, teacher behavior recognition, and item leaving recognition. The accuracy of behavior recognition for test students is 97.8%. The accuracy rate of teacher behavior recognition is 96.3%, and the accuracy rate of item leaving identification is 95.6%. This article helps to improve the level of campus safety management.