Hyperspectral Land-Sea-Cloud Segmentation for Onboard Processing via Spectral Band Selection

Keyphrases

• Atmospheric Effects 33%
• Band Configuration 33%
• Band Selection 100%
• Boundary Delineation 33%
• Class Imbalance 33%
• Cloud Detection 33%
• Cloud Discrimination 33%
• Cloud Segmentation 100%
• Combined Loss Function 33%
• Computational Complexity 33%
• Data Availability 33%
• Dice Loss 33%
• Edge Information 33%
• Fixed-length 33%
• Function-based 33%
• Hyperspectral 100%
• Hyperspectral Mission 33%
• Intersection over Union 33%
• Land-sea 100%
• Land-sea Boundary 33%
• Major Challenges 33%
• Mean Intersection over Union 33%
• Near-infrared 33%
• Near-infrared (NIR) Light 33%
• Near-infrared Channel 33%
• On-board Processing 100%
• Pixel Accuracy 33%
• Preprocessing Step 33%
• Red Edge 33%
• Remote Sensing 33%
• Resource-constrained 33%
• RGB-NIR 33%
• Satellite Imagery 33%
• Semantic Segmentation 33%
• Spectral Analysis 33%
• Spectral Variability 33%
• Weighted Cross-entropy Loss 33%

Earth and Planetary Sciences

• Atmospheric Effect 50%
• Entropy 50%
• Hyperspectral Image 50%
• Patch Size 50%
• Remote Sensing 50%
• Satellite Imagery 50%
• Spectral Band 100%