Deep Feature Gaussian Processes for Single-Scene Aerosol Optical Depth Reconstruction

TL;DR

This paper introduces Deep Feature Gaussian Processes (DFGP), a novel method combining deep learning and Gaussian processes for single-scene aerosol optical depth reconstruction, outperforming existing methods on real datasets.

Contribution

The paper presents a new approach that leverages deep feature learning and Gaussian processes to improve AOD reconstruction without multi-temporal data.

Findings

DFGP outperforms deep CNN and random forest in R^2 scores.

Achieved R^2 of 0.7431 on MODIS AOD and 0.9211 on EMIT AOD.

Increased R^2 by over 0.35 compared to random forest.

Abstract

Remote sensing data provide a low-cost solution for large-scale monitoring of air pollution via the retrieval of aerosol optical depth (AOD), but is often limited by cloud contamination. Existing methods for AOD reconstruction rely on temporal information. However, for remote sensing data at high spatial resolution, multi-temporal observations are often unavailable. In this letter, we take advantage of deep representation learning from convolutional neural networks and propose Deep Feature Gaussian Processes (DFGP) for single-scene AOD reconstruction. By using deep learning, we transform the variables to a feature space with better explainable power. By using Gaussian processes, we explicitly consider the correlation between observed AOD and missing AOD in spatial and feature domains. Experiments on two AOD datasets with real-world cloud patterns showed that the proposed method outperformed deep CNN and random forest, achieving R$^2$ of 0.7431 on MODIS AOD and R$^2$ of 0.9211 on EMIT AOD, compared to deep CNN's R$^2$ of 0.6507 and R$^2$ of 0.8619. The proposed methods increased R$^2$ by over 0.35 compared to the popular random forest in AOD reconstruction. The data and code used in this study are available at \url{https://skrisliu.com/dfgp}.