Google Earth Engine code for Data Fusion of Sentinel-2 and NAIP Optical Imagery for High-Resolution Urban Vegetation Classification in the Los Angeles Megacity.
0: Impervious surface
1: Tree (mixed evergreen/deciduous)
2: Grass (assumed irrigated)
3: Shrub
4: Non-photosynthetic vegetation
5: Water (masked using MNDWI/NDWI)
EPSG: 4326/WGS84
- sentinel_naip_FINAL: classifies SoCAB at 0.6 - 10m resolution using supervised machine learning with Sentinel-2 and NAIP imagery in Google Earth Engine
- aggregateNAIP_FINAL: aggregates 0.6 - 10m product from sentinel_naipFINAL script to 30m fractional and majority pixels in Google Earth Engine
- Non-urban and shadow regions classified with ESA Sentinel-2 imagery (10-60 m)
- Urban regions classified with USDA National Agricultural Imagery Program (NAIP) imagery (0.6-1 m)
- Aggregated 30m pixels are aligned with projection of 2016 NLCD classification
classified_SoCAB: 1-band 0.6-10m classification of 6 land cover classes across SoCAB
30m_majority_SoCAB: 1-band 30m classification of majority land cover class within each 30m pixel across all of SoCAB
30m_fractional_SoCAB: 5-band 30m classification of 60cm fractional land cover class within each 30m pixel, where bands correspond to the following labelling (band1 = impervious, band2 = tree, band3 = grass, band4 = shrub, band5 = npv, band6 = water)
SoCAB: Southern California Air Basin
SB: San Bernardino County
RI: Riverside County
OR: Orange County
LA: Los Angeles County
A portion of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Support from the Earth Science Division OCO-2 program is acknowledged. Copyright 2020. All rights reserved.