Real-time monitoring and visualization of indoor environments based on edge computing and deep learning

Edge computing and deep learning are leveraged in this work to deliver a high-performance, real-time indoor environmental monitoring and visualization system. The architecture uses a hierarchical sensor network to obtain multimodal data streams, and the edge nodes build lightweight neural models for local processing and rapid identification of anomalies. The system improves prediction accuracy and computational efficiency by using technologies such as knowledge distillation, model quantization, and collaborative inference scheduling. The WebGL-based dashboard features fast alerts, an interactive user interface, and dynamic spatiotemporal rendering. Extensive experiments in the workplace validate the sub-second latency, high anomaly detection accuracy, and strong scalability with increasing sensor density. The alerting and visualization modules facilitate proactive facility management, while power consumption and total cost analysis confirm the feasibility of continuous deployment. Combining edge intelligence with advanced neural technologies provides a flexible and adaptable platform for real-time indoor monitoring in complex building scenarios. Due to its modularity and technical precision, the framework can be easily integrated into smart building and urban IoT infrastructures.