STARS: Sensor-agnostic Transformer Architecture for Remote Sensing

TL;DR

This paper introduces a sensor-agnostic spectral transformer using a Universal Spectral Representation that enables a single model to process data from various spectral sensors, enhancing generalization and foundation model development.

Contribution

The paper proposes a novel sensor-agnostic spectral transformer architecture with a universal spectral representation and a self-supervised pre-training method for sensor-independent spectral feature learning.

Findings

Model generalizes to unseen sensors

Effective sensor-independent spectral feature learning

Supports diverse spectral data for foundation models

Abstract

We present a sensor-agnostic spectral transformer as the basis for spectral foundation models. To that end, we introduce a Universal Spectral Representation (USR) that leverages sensor meta-data, such as sensing kernel specifications and sensing wavelengths, to encode spectra obtained from any spectral instrument into a common representation, such that a single model can ingest data from any sensor. Furthermore, we develop a methodology for pre-training such models in a self-supervised manner using a novel random sensor-augmentation and reconstruction pipeline to learn spectral features independent of the sensing paradigm. We demonstrate that our architecture can learn sensor independent spectral features that generalize effectively to sensors not seen during training. This work sets the stage for training foundation models that can both leverage and be effective for the growing diversity of spectral data.