Wide-Area Visual Monitoring System Based on NB-IoT

Abstract

Translate Article

Take Notes

HighlightsAn integrated architecture for real-time event detection and response has been proposed, addressing the critical challenge of effective detection of unexpected events in wide-area surveillance. Utilizing the wide coverage of narrowband cellular networks (e.g., NB-IoT), the system ensures efficient image data transmission even in remote or rural areas, providing real-time alerts.What are the main findings?The system detects abnormal events by analyzing sequential image frames using intelligent algorithms and stores images only upon anomaly detection, improving storage efficiency.By using the CoAP protocol to transmit encapsulated JPEG images and leveraging the MQTT protocol to deliver image data to client applications, the system achieves efficient data transmission and processing.What is the implication of the main finding?This study offers an intelligent, scalable, and responsive solution for wide-area surveillance systems, overcoming limitations of traditional systems such as low storage efficiency, limited transmission range, and complex operation.The intelligent anomaly detection algorithms reduce the risks and costs associated with manual monitoring, enhancing both the efficiency and accuracy of anomaly detection.Effective detection of unexpected events in wide-area surveillance remains a critical challenge in the development of intelligent monitoring systems. Recent advancements in Narrowband Internet of Things (NB-IoT) and 5G technologies provide a robust foundation to address this issue. This study presents an integrated architecture for real-time event detection and response. The system utilizes the Constrained Application Protocol (CoAP) to transmit encapsulated JPEG images from NB-IoT modules to an Internet of Things (IoT) server. Upon receipt, images are decoded, processed, and archived in a centralized database. Subsequently, image data are transmitted to client applications via WebSocket, leveraging the Message Queuing Telemetry Transport (MQTT) protocol. By performing temporal image comparison, the system identifies abnormal events within the monitored area. Once an anomaly is detected, a visual alert is generated and presented through an interactive interface. The test results show that the image recognition accuracy is consistently above 98%. This approach enables intelligent, scalable, and responsive wide-area surveillance reliably, overcoming the constraints of conventional isolated and passive monitoring systems.