Gold Exploration using Representations from a Multispectral Autoencoder

TL;DR

This paper introduces a novel approach for gold exploration using multispectral satellite imagery by leveraging a pretrained autoencoder to generate representations that improve classification accuracy over raw data.

Contribution

It presents a new framework utilizing a pretrained autoencoder to extract spectral-spatial features for mineral prospectivity mapping from satellite images.

Findings

Patch-level accuracy improved from 0.51 to 0.68

Image-level accuracy improved from 0.55 to 0.73

Generative embeddings effectively capture mineralogical patterns

Abstract

Satellite imagery is employed for large-scale prospectivity mapping due to the high cost and typically limited availability of on-site mineral exploration data. In this work, we present a proof-of-concept framework that leverages generative representations learned from multispectral Sentinel-2 imagery to identify gold-bearing regions from space. An autoencoder foundation model, called Isometric, which is pretrained on the large-scale FalconSpace-S2 v1.0 dataset, produces information-dense spectral-spatial representations that serve as inputs to a lightweight XGBoost classifier. We compare this representation-based approach with a raw spectral input baseline using a dataset of 63 Sentinel-2 images from known gold and non-gold locations. The proposed method improves patch-level accuracy from 0.51 to 0.68 and image-level accuracy from 0.55 to 0.73, demonstrating that generative embeddings capture transferable mineralogical patterns even with limited labeled data. These results highlight the potential of foundation-model representations to make mineral exploration more efficient, scalable, and globally applicable.