satellite-image-deep-learning/techniques

Comprehensive reference covering 17+ specialized task categories—from cloud removal and change detection to SAR processing and foundational models—with curated links to implementations and research across PyTorch, TensorFlow, and domain-specific architectures. Addresses satellite imagery's unique challenges like high-resolution gigapixel processing, multi-temporal analysis, and sparse labeled data through techniques including self-supervised learning, few-shot approaches, and vision-language models. Integrates with standard remote sensing datasets (Sentinel, EuroSAT, UC Merced) and modern frameworks from foundational models to explainable AI.

10,055 stars. Actively maintained with 8 commits in the last 30 days.