RS-DFM: A Remote Sensing Distributed Foundation Model for Diverse Downstream Tasks

TL;DR

RS-DFM introduces a distributed foundation model for remote sensing that enables collaborative perception across multiple platforms, improving performance on diverse downstream tasks through generalized feature extraction and information interaction.

Contribution

The paper presents a novel remote sensing foundation model that supports online multi-platform collaboration and task-agnostic information interaction, enhancing understanding of large-scale scenarios.

Findings

Achieves state-of-the-art results on multiple remote sensing tasks.

Effectively utilizes geometric priors for feature transformation.

Demonstrates robust multi-UAV collaborative perception.

Abstract

Remote sensing lightweight foundation models have achieved notable success in online perception within remote sensing. However, their capabilities are restricted to performing online inference solely based on their own observations and models, thus lacking a comprehensive understanding of large-scale remote sensing scenarios. To overcome this limitation, we propose a Remote Sensing Distributed Foundation Model (RS-DFM) based on generalized information mapping and interaction. This model can realize online collaborative perception across multiple platforms and various downstream tasks by mapping observations into a unified space and implementing a task-agnostic information interaction strategy. Specifically, we leverage the ground-based geometric prior of remote sensing oblique observations to transform the feature mapping from absolute depth estimation to relative depth estimation, thereby enhancing the model's ability to extract generalized features across diverse heights and perspectives. Additionally, we present a dual-branch information compression module to decouple high-frequency and low-frequency feature information, achieving feature-level compression while preserving essential task-agnostic details. In support of our research, we create a multi-task simulation dataset named AirCo-MultiTasks for multi-UAV collaborative observation. We also conduct extensive experiments, including 3D object detection, instance segmentation, and trajectory prediction. The numerous results demonstrate that our RS-DFM achieves state-of-the-art performance across various downstream tasks.