Physics-Constrained Denoising Autoencoders for Data-Scarce Wildfire UAV Sensing

TL;DR

This paper introduces PC$^2$DAE, a physics-informed autoencoder for denoising UAV wildfire sensor data, effectively handling data scarcity and ensuring physically plausible outputs.

Contribution

The work presents a novel physics-constrained autoencoder architecture that embeds physical laws directly, enabling effective denoising with limited data and ensuring physically admissible results.

Findings

PC$^2$DAE-Lean improves smoothness by 67.3% and noise reduction by 90.7%.

The lean variant outperforms the wide variant by 5.6% in smoothness.

Training completes in under 65 seconds on consumer hardware.

Abstract

Wildfire monitoring requires high-resolution atmospheric measurements, yet low-cost sensors on Unmanned Aerial Vehicles (UAVs) exhibit baseline drift, cross-sensitivity, and response lag that corrupt concentration estimates. Traditional deep learning denoising approaches demand large datasets impractical to obtain from limited UAV flight campaigns. We present PC$^2$DAE, a physics-informed denoising autoencoder that addresses data scarcity by embedding physical constraints directly into the network architecture. Non-negative concentration estimates are enforced via softplus activations and physically plausible temporal smoothing, ensuring outputs are physically admissible by construction rather than relying on loss function penalties. The architecture employs hierarchical decoder heads for Black Carbon, Gas, and CO$_2$ sensor families, with two variants: PC$^2$DAE-Lean (21k parameters) for edge deployment and PC$^2$DAE-Wide (204k parameters) for offline processing. We evaluate on 7,894 synchronized 1 Hz samples collected from UAV flights during prescribed burns in Saskatchewan, Canada (approximately 2.2 hours of flight data), two orders of magnitude below typical deep learning requirements. PC$^2$DAE-Lean achieves 67.3\% smoothness improvement and 90.7\% high-frequency noise reduction with zero physics violations. Five baselines (LSTM-AE, U-Net, Transformer, CBDAE, DeSpaWN) produce 15--23\% negative outputs. The lean variant outperforms wide (+5.6\% smoothness), suggesting reduced capacity with strong inductive bias prevents overfitting in data-scarce regimes. Training completes in under 65 seconds on consumer hardware.