Thermal Anomaly Detection using Physics Aware Neuromorphic Networks: Comparison between Raw and L1C Sentinel-2 Data
Summary: arXiv:2604.18606v1 Announce Type: cross
Abstract
Damage caused by bushfires and volcanic eruptions escalates rapidly when detection is delayed, making fast and reliable early warning capabilities essential. Recent Earth Observation (EO) approaches have shown that thermal anomaly detection can be performed directly on decompressed Level-0 (L0) sensor data, avoiding computationally expensive preprocessing chains.
However, direct exploitation of raw data remains challenging due to several factors such as domain shift, sensor drift, radiometric inconsistencies, and the scarcity of labelled training samples. To address these challenges, this work proposes a Physics-Aware Neuromorphic Network (PANN) framework for onboard thermal anomaly detection.
Key Features of the PANN Framework
- Lightweight Architecture: Inspired by physical neural network principles and neuromorphic computing paradigms.
- Data Evaluation: The framework is evaluated using two Sentinel-2 datasets: decompressed L0 with additional metadata (i.e., raw) and Level-1C (L1C).
- Performance Metrics: Achieves a Matthews Correlation Coefficient (MCC) of 0.809 on raw measurements compared to 0.875 when using ground-processed L1C products.
Processing Latency and Resource Efficiency
The mean processing latency per L0 granule is measured at 2.44 ± 0.09 seconds, which is below the Sentinel-2 acquisition time of 3.6 seconds. This demonstrates the feasibility of real-time, onboard processing.
Furthermore, the projected execution time for the corresponding neuromorphic hardware instantiation is substantially lower at 0.1290 ± 0.0002 seconds. This significant reduction in processing time enhances the potential for timely anomaly detection in critical scenarios.
Memory Usage and Onboard Constraints
Memory usage, including all necessary programs and packages, remains within realistic onboard constraints. The requirements are:
- Software PANN: 0.673 ± 0.007 Gb
- Estimated Hardware Realization: 0.393 ± 0.004 Gb
Conclusion
Overall, these results indicate that the Physics-Aware Neuromorphic Network (PANN) offers a promising pathway toward low-latency and resource-efficient onboard Earth Observation (EO) processing for thermal event detection. The advancements presented in this framework can potentially revolutionize early warning systems, enabling faster responses to natural disasters such as bushfires and volcanic eruptions.