Satellite-image-deep-learning/techniques is a reference guide that catalogs deep learning techniques for satellite and aerial imagery analysis.
The repository addresses the challenge of applying deep learning to satellite and aerial images, which present unique obstacles including vast image sizes and diverse object classes. It provides an exhaustive overview of architectures, models, and algorithms tailored to this domain, organized by task type rather than as a framework or library. The guide covers classification, segmentation, object detection, regression, cloud detection and removal, change detection, time series analysis, crop classification, crop yield and vegetation forecasting, generative networks, autoencoders and dimensionality reduction, few-shot and zero-shot learning, self-supervised and contrastive learning, synthetic aperture radar processing, explainable AI, and large vision and language models.
This is a reference resource rather than a tool to integrate into a project. It suits developers and researchers working on earth observation problems who need to understand which techniques and models apply to their specific task. The repository is organized to enable quick lookup—users search the page for relevant keywords or technique names to find applicable approaches. It does not provide implementations or code samples; instead, it catalogs the landscape of methods available for satellite imagery work.
The project maintains an active curated collection of techniques across a broad range of deep learning approaches for remote sensing. The repository includes a companion website that serves as the primary interface for browsing the techniques catalog.
