This code is based on Pixel-Based Image Compositing for Large-Area Dense Time Series Applications and Science. (White et al., 2014) http://www.tandfonline.com/doi/full/10.1080/07038992.2014.945827
It uses a series of pixel based scores to generate a composite with the Best Available Pixel, assuming it is the one that has better score.
2017 Rodrigo E. Principe - geebap - https://github.com/fitoprincipe/geebap
Rodrigo E. Principe: fitoprincipe82@gmail.com
To use this package you must have installed and running Google Earth Engine Python API: https://developers.google.com/earth-engine/python_install
Once you have that, proceed
pip install geebap
this will install also geetools that you could use besides geebap
After following Option 1 or 2 in https://developers.google.com/earth-engine/python_install, open a new notebook and write:
import sys
!{sys.executable} -m pip install geebapCollections come from geetools.collection. For examples see: https://github.com/gee-community/gee_tools/tree/master/notebooks/collection
- Satellite
- Distance to clouds and shadows masks
- Atmospheric Opacity
- Day of the year (best_doy)
- Masked pixels percentage
- Outliers
- Absolute value of a vegetation index
- ndvi
- evi
- nbr
- Sites size should not be too big. Works with 300 km2 tiles
If you are using Jupyter, you can download a notebook from https://github.com/fitoprincipe/geebap/blob/master/Best_Available_Pixel_Composite.ipynb
else, if you are using another approach, like Spyder, create an empty script and paste the following code:
import ee
ee.Initialize()
import geebap
from geetools import tools
import pprint
pp = pprint.PrettyPrinter(indent=2)
# SEASON
a_season = geebap.Season('11-15', '03-15')
# MASKS
cld_mask = geebap.masks.Mask()
# Combine masks in a tuple
masks = (cld_mask,)
# FILTERS
filt_cld = geebap.filters.CloudCover()
# filt_mask = geebap.filters.MaskCover() # Doesn't work
# Combine filters in a tuple
filters = (filt_cld,)#, filt_mask)
# SCORES
best_doy = geebap.scores.Doy('01-15', a_season)
sat = geebap.scores.Satellite()
out = geebap.scores.Outliers(("ndvi",))
ind = geebap.scores.Index("ndvi")
maskpercent = geebap.scores.MaskPercentKernel()
dist = geebap.scores.CloudDist()
# Combine scores in a tuple
scores = (
best_doy,
sat,
out,
ind,
maskpercent,
dist
)
# BAP OBJECT
BAP = geebap.Bap(range=(0, 0),
season=a_season,
masks=masks,
scores=scores,
filters=filters)
# SITE
site = ee.Geometry.Polygon([[-71.5,-42.5],
[-71.5,-43],
[-72,-43],
[-72,-42.5]])
# COMPOSITE
composite = BAP.build_composite_best(2019, site=site, indices=("ndvi",))
# `composite` is a ee.Image object, so you can do anything
# from here..
one_value = tools.image.getValue(composite,
site.centroid(),
30, 'client')
pp.pprint(one_value)Prints:
{ 'blue': 733,
'col_id': 29,
'date': 20190201,
'green': 552,
'ndvi': 0.7752976417541504,
'nir': 2524,
'red': 313,
'score': 5.351020336151123,
'swir': 661,
'swir2': 244,
'thermal': 2883}