A Sensor Agnostic Domain Generalization Framework for Leveraging Geospatial Foundation Models: Enhancing Semantic Segmentation viaSynergistic Pseudo-Labeling and Generative Learning

2025/05/02 by Anan Yaghmour, Yaghmour, Anan, Melba M. Crawford +3 · 2 citations
Computer Science · Earth and Planetary Sciences · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Geological Modeling and Analysis #Semantic Web and Ontologies

paper · pdf · doi:10.48550/arxiv.2505.01558

openalex publication_date 2025/05/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Remote sensing enables a wide range of critical applications such as land cover and land use mapping, crop yield prediction, and environmental monitoring. Advances in satellite technology have expanded remote sensing datasets, yet high-performance segmentation models remain dependent on extensive labeled data, challenged by annotation scarcity and variability across sensors, illumination, and geography. Domain adaptation offers a promising solution to improve model generalization. This paper introduces a domain generalization approach to leveraging emerging geospatial foundation models by combining soft-alignment pseudo-labeling with source-to-target generative pre-training. We further provide new mathematical insights into MAE-based generative learning for domain-invariant feature learning. Experiments with hyperspectral and multispectral remote sensing datasets confirm our method's effectiveness in enhancing adaptability and segmentation.

Citations

• ExPLoRA: Parameter-Efficient Extended Pre-Training to Adapt Vision Transformers under Domain Shifts
• MetaEarth: A Generative Foundation Model for Global-Scale Remote Sensing Image Generation
• On the Generalization Ability of Unsupervised Pretraining
• SpectralGPT: Spectral Remote Sensing Foundation Model
• FLAIR: a Country-Scale Land Cover Semantic Segmentation Dataset From Multi-Source Optical Imagery
• Cross-City Matters: A Multimodal Remote Sensing Benchmark Dataset for Cross-City Semantic Segmentation using High-Resolution Domain Adaptation Networks
• Vision-Language Models in Remote Sensing: Current Progress and Future Trends
• LLaMA: Open and Efficient Foundation Language Models
• MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation
• Florence: A New Foundation Model for Computer Vision
• Masked Autoencoders Are Scalable Vision Learners
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
• Parameter-Efficient Fine-Tuning for Pre-Trained Vision Models: A Survey and Benchmark

Cited by

• IRSAMap:Towards Large-Scale, High-Resolution Land Cover Map Vectorization
• From Pixels to Images: A Structural Survey of Deep Learning Paradigms in Remote Sensing Image Semantic Segmentation

Related

• CrossEarth: Geospatial Vision Foundation Model for Domain Generalizable Remote Sensing Semantic Segmentation
• Leveraging Domain Adaptation for Low-Resource Geospatial Machine Learning
• StandardGAN: Multi-source Domain Adaptation for Semantic Segmentation of Very High Resolution Satellite Images by Data Standardization
• SiamSeg: Self-Training with Contrastive Learning for Unsupervised Domain Adaptation Semantic Segmentation in Remote Sensing
• GeoMultiTaskNet: remote sensing unsupervised domain adaptation using geographical coordinates