Satellite imagery dataset of manure application on pasture fields
Description
Sentinel-2 imagery of freshly manured pasture fields in the northern region of Spain. Each image is in ".tif" format and has 64 channels. The first 13 channels are the raw bands of Sentinel-2. The remaining 51 channels are some of the most common hyperspectral indices in Precision Agriculture. The order of the channels is the following (from 0 to 63): B01 B02 B03 B04 B05 B06 B07 B08 B8A B09 B10 B11 B12 NDVI NSNDVI SDI GNDVI SAVI OSAVI EOMI1 EOMI2 EOMI3 EOMI4 BNR2 RVI DVI RENDVI1 RENDVI2 RENDVI3 CI1 CI2 CI3 NDRE MCARI MCARI1 MCARI2 MTVI1 MTVI2 EVI AVI GCI BSI NBRI NDRE1 NDRE2 NDRE3 MSAVI WDRVI ARVI1 ARVI2 TSAVI CARI1 CARI2 CVI EVI1 EVI2 EVI3 SCI GRNDVI GBNDVI GLI ATSAVI ALTERATION CTVI This data can be used to develop machine learning detectors for freshly manured fields.
Files
Steps to reproduce
1. On-site investigation to verify manure application and observe the plot dimensions. 2. Generate a KML of the plot in Google Earth Engine. 3. Download Sentinel-2 imagery of the plot from Google Earth Engine. (use download_imagery.js from src folder) 4. Create groundtruh masks using the KML of the plot and the georeferenced imagery from Sentinel-2. (use generate_groundtruth.py from src folder) 5. Calculate extra hipersectral indices and add them following the 13 Sentinel-2 bands. (use calculate_indices.py from src folder)