MetaEarth: A Generative Foundation Model for Global-Scale Remote Sensing Image Generation

TL;DR

MetaEarth is a groundbreaking generative model capable of creating high-resolution, global-scale remote sensing images at arbitrary locations and resolutions, significantly advancing the scope of image generation in Earth observation.

Contribution

The paper introduces MetaEarth, a novel framework that scales remote sensing image generation to a global level using a resolution-guided self-cascading diffusion model and a new noise sampling strategy.

Findings

Successfully generates worldwide, multi-resolution remote sensing images.

Demonstrates high-quality, realistic images at unbounded scales.

Provides a large dataset for training and downstream applications.

Abstract

The recent advancement of generative foundational models has ushered in a new era of image generation in the realm of natural images, revolutionizing art design, entertainment, environment simulation, and beyond. Despite producing high-quality samples, existing methods are constrained to generating images of scenes at a limited scale. In this paper, we present MetaEarth, a generative foundation model that breaks the barrier by scaling image generation to a global level, exploring the creation of worldwide, multi-resolution, unbounded, and virtually limitless remote sensing images. In MetaEarth, we propose a resolution-guided self-cascading generative framework, which enables the generating of images at any region with a wide range of geographical resolutions. To achieve unbounded and arbitrary-sized image generation, we design a novel noise sampling strategy for denoising diffusion models by analyzing the generation conditions and initial noise. To train MetaEarth, we construct a large dataset comprising multi-resolution optical remote sensing images with geographical information. Experiments have demonstrated the powerful capabilities of our method in generating global-scale images. Additionally, the MetaEarth serves as a data engine that can provide high-quality and rich training data for downstream tasks. Our model opens up new possibilities for constructing generative world models by simulating Earth visuals from an innovative overhead perspective.