Early- and in-season crop type mapping without current-year ground truth: Generating labels from historical information via a topology-based approach

Keyphrases

• Topological Methods 100%
• Historical Information 100%
• Crop Mapping 100%
• Corn-soybean 50%
• Land Cover Classification 50%
• Overall Accuracy 50%
• Paddy Rice 50%
• Multiple Crops 50%
• Northeast China 50%
• F1 Score 50%
• Knowledge Transfer 25%
• Remote Sensing 25%
• Iowa 25%
• Postharvest Treatment 25%
• Automatically Generate 25%
• Landsat 8 25%
• US Midwest 25%
• Crop Classification 25%
• Relative Position 25%
• Interannual Variation 25%
• Histogram 25%
• Decision Boundary 25%
• Topological Information 25%
• Crop Type 25%
• Historical Knowledge 25%
• Classification Decision 25%
• Northern France 25%
• Transferrable 25%
• Random Forest Classifier 25%
• Quality Label 25%
• Shift-invariant 25%
• Harvest Decision 25%
• Sentinel-2 Data 25%
• Silking Stage 25%
• Hauts-de-France 25%
• Subsequent Crop 25%
• Flowering Stage 25%
• Ground-truth Labels 25%
• Spectral Feature Space 25%

Earth and Planetary Sciences

• Land Cover 100%
• France 100%
• China 100%
• Iowa 50%
• Landsat 8 50%
• Sentinel-2 50%
• Remote Sensing 50%
• Annual Variation 50%
• Histogram 50%

Agricultural and Biological Sciences

• Paddies 100%
• Land Cover 100%
• Preharvest 50%
• Anthesis 50%
• Landsat 50%
• Remote Sensing 50%
• Sentinel-2 Satellite 50%