SeeFar: Satellite Agnostic Multi-Resolution Dataset for Geospatial Foundation Models

TL;DR

SeeFar addresses the fragmentation of satellite imagery by introducing a satellite-agnostic, multi-resolution dataset standardization pipeline that harmonizes data from Landsat, Sentinel-2, WorldStrat, and NewSat IV into uniform 384×384, 4-band patches across resolutions of 30, 10, 1.5, and 1 m. It details a comprehensive preprocessing framework, including channel alignment, histogram matching, metadata augmentation, and rigorous quality controls (e.g., no-data and cloud thresholds using $NDSI = \frac{Green - NIR}{Green + NIR}$), and creates train/validation/test benchmarks to support geospatial foundation-model training. The dataset is accompanied by licensing clarity (CC-BY 4.0, CC-BY-NC 4.0, CC-BY-SA 3.0 IGO, and USGS Landsat) and a plan for continual growth and refinement, enabling broader access to high-quality, interoperable satellite data. By delivering a scalable, standardized, and diverse multi-source corpus, SeeFar enables robust, satellite-agnostic model development with improved transferability to historical and high-resolution imagery.

Abstract

SeeFar is an evolving collection of multi-resolution satellite images from public and commercial satellites. We specifically curated this dataset for training geospatial foundation models, unconstrained by satellite type. In recent years, advances in technology have made satellite imagery more accessible than ever. More earth-observing satellites have been launched in the last five years than in the previous fifty. Modern commercial satellites now offer up to 100 times the spatial resolution of public access satellites. However, the high cost and limited historical availability of commercial satellite imagery is a barrier to the training of foundational models, impacting what images can be used during inference. The SeeFar dataset represents a step towards training models that are satellite-agnostic by combining multi-resolution commercial and public access pre-processed images. This will enable users to utilize historical data alongside higher-resolution, more expensive satellite imagery, offering greater flexibility during inference. To achieve this, we describe a process for standardizing data from diverse satellite sources, normalizing different data formats, and aligning spectral bands to enhance interoperability. The SeeFar dataset includes images at a resolution of 384x384 pixels, spanning four spectral bands (Blue, Green, Red, and Near-Infrared) and expanding spatial resolutions (starting with 30, 10, 1.5, and 1.0 meters), all in cloud-optimized GeoTIFF format. It also provides consistent and comprehensive metadata to enhance data transparency and reliability. By aggregating data from multiple sources, SeeFar makes processed and consistent satellite data accessible to a wider range of users - from researchers to policymakers - fostering competition and innovation in satellite imagery analysis. The dataset is available at \url{coastalcarbon.ai/seefar}.

Figures (2)

• Figure 1: SeeFar processed data examples from different satellite vendors.
• Figure 2: SeeFar Methodology: an overview of the proposed method utilized to create a standard dataset.