Lightweight Cloud Masking Models for On-Board Inference in Hyperspectral Imaging
This addresses the need for real-time preprocessing in satellite imaging, though it is incremental as it applies existing methods to optimize for lightweight deployment.
The study tackled cloud and cloud shadow masking in hyperspectral satellite imaging by evaluating machine learning models like XGBoost, LightGBM, and CNNs, with the CNN with feature reduction achieving over 93% accuracy and as few as 597 parameters for efficient on-board inference.
Cloud and cloud shadow masking is a crucial preprocessing step in hyperspectral satellite imaging, enabling the extraction of high-quality, analysis-ready data. This study evaluates various machine learning approaches, including gradient boosting methods such as XGBoost and LightGBM as well as convolutional neural networks (CNNs). All boosting and CNN models achieved accuracies exceeding 93%. Among the investigated models, the CNN with feature reduction emerged as the most efficient, offering a balance of high accuracy, low storage requirements, and rapid inference times on both CPUs and GPUs. Variations of this version, with only up to 597 trainable parameters, demonstrated the best trade-off in terms of deployment feasibility, accuracy, and computational efficiency. These results demonstrate the potential of lightweight artificial intelligence (AI) models for real-time hyperspectral image processing, supporting the development of on-board satellite AI systems for space-based applications.