Niect: A Model for Intrusion Security Detection Applied to Campus Video Surveillance Edge Networks
Feng Zhou, Ming Yuan, Yu Liu, Hongbing Zhang, Mingyu Gu, Tongming Zhou · 2024
The Internet of Things is a network of interconnected devices that communicate with each other and use the Internet to send and receive data to users. The Internet of Things is widely used in applications such as automobiles, energy, logistics, transportation, government affairs, and campuses. During our research on network security of campus video surveillance systems, we found that the IoT devices of campus video surveillance systems have long been scanned and invaded by the engines of various search devices. This situation brings great hidden dangers and considerable risks to campus surveillance network security. In order to maintain the network security of IoT devices in the campus video surveillance system, we propose a network intrusion security detection model Niect (Network Intrusion Security Detection Model) based on the convolutional neural network, energy valley optimization algorithm, LightGBM, and CatBoost algorithm. We use the Edge-IIoTset, an integrated real-world cybersecurity dataset for IoT and Industrial IoT applications. The Niect model we proposed achieved an accuracy of 96.35%, a precision of 95.87%, a recall of 93.57%, and an AUC value of 91.68% in the experiment. This experimental result shows that the proposed Niect model can provide strong technical support for network security intrusion detection of IoT devices in campus video surveillance systems.