Search papers, labs, and topics across Lattice.
This paper introduces a lightweight edge-oriented framework for anomaly detection and localization in geospatial data streams, addressing the challenges of data storage, processing, and communication in real-time applications. By utilizing the H3 discrete global grid system and a multi-scale drill-down logic, the approach achieves a remarkable 99.7% reduction in computational evaluations compared to traditional methods, while effectively filtering out noise to identify persistent anomalies. The findings highlight the framework's capability to transform vast geospatial data into actionable insights, making it highly relevant for critical applications across various domains.
Achieving a staggering 99.7% reduction in computational overhead, this framework revolutionizes real-time anomaly detection in geospatial data streams.
As an increasing number of critical applications, including environmental, emergency, meteorological, and agricultural, rely on real-time anomaly detection in geospatial data streams, challenges related to the storage, processing, and communication of this data arise. Traditionally, large volumes of data have been sent to centralized processing locations for insight extraction. Given the big data context of these applications, this approach becomes increasingly infeasible as data volume and velocity continue to increase. This paper proposes a lightweight edge-oriented approach for anomaly detection and localization for geospatial data streams. By leveraging the H3 discrete global grid system and a multi-scale drill-down logic, the proposed approach significantly reduces computational overhead, achieving a 99.7\% reduction in evaluations compared to traditional flat-scan methods. Furthermore, by filtering out noise-induced flickering anomalies at lower resolutions, spatially-persistent anomalous signals can be efficiently identified. The results demonstrate that the proposed framework effectively distills massive geospatial data into actionable insights.
