Transform Google Earth Engine script to Python with landsat 8 temporal data

Question

I am doing a study for my work and I need to adapt my google earth engine script in python, but I am encountering some problems. Could they help me?

Here is a google earth engine script:

import ee
ee.Initialize()
 //Importing image and geometry:
var l8 = ee.ImageCollection("LANDSAT/LC08/C01/T1_SR")
//    geometry = /* color: #d63000 */ee.Geometry.Polygon(
//        [[[-80.92489760146748, 25.433457120928352],
//          [-80.64474623427998, 25.488013471687964],
//          [-80.57882826552998, 25.710940372707608],
//          [-81.02377455459248, 25.770317250349557],
//          [-80.95236342177998, 25.552457242621447]]]);

//Filtering date, polygon, and cloudiness
    var image = l8.filterDate ('2010-09-01', '2021-12-31')
                .filterBounds (geometry)
                .filterMetadata ('CLOUD_COVER', 'less_than', 1);

    //NDVI calculation:
    var ndvi_func = function (i) {
      var ndvi = i.normalizedDifference (['B5', 'B4']).rename ('NDVI')
      return i.addBands(ndvi);
    }

    var image_ndvi = image.map(ndvi_func);

    //Calculating year wise NDVI
    var year = ee.List.sequence(2010,2021);
    var year_func = function(y){
      var range = ee.Filter.calendarRange (y, y, 'year');
      return image_ndvi.select('NDVI').filter(range).mean().set ('Year', y)
    };
    var yearwise_ndvi = ee.ImageCollection(year.map(year_func));
    print (yearwise_ndvi);
    Map.addLayer (yearwise_ndvi)

//Creating time-series chart:
var chart = ui.Chart.image.series ({
  imageCollection: image_ndvi.select('NDVI'),
  region: geometry,
  reducer: ee.Reducer.mean(),
  scale: 30})

print(chart);

Figure:

The script above presents an NDVI time series for a region and I need to do it in python. Here is a script that shows the errors in python:

l8 = ee.ImageCollection("LANDSAT/LC08/C01/T1_SR")
geometry = ee.Geometry.Polygon([[[-80.92489760146748, 25.433457120928352],
                                 [-80.64474623427998, 25.488013471687964],
                                 [-80.57882826552998, 25.710940372707608],
                                 [-81.02377455459248, 25.770317250349557],
                                 [-80.95236342177998, 25.552457242621447]]]);

#Filtering date, polygon, and cloudiness
image = l8.filterDate ('2010-09-01', '2021-12-31')
image  = image.filterBounds (geometry)
image  = image.filterMetadata ('CLOUD_COVER', 'less_than', 1);

#NDVI calculation:
def ndvi_func(i):
    ndvi = i.normalizedDifference (['B5', 'B4']).rename ('NDVI')
    return i.addBands(ndvi)

image_ndvi = ndvi_func(image.map)

#Calculating year wise NDVI
year = ee.List.sequence(2010,2021);
def year_func(y):
    range = ee.Filter.calendarRange (y, y, 'year');
    return image_ndvi.select('NDVI').filter(range).mean().set ('Year', y)

yearwise_ndvi = ee.ImageCollection(year.map(year_func));
    print (yearwise_ndvi);
    Map.addLayer (yearwise_ndvi)

#Creating time-series chart:
chart = ui.Chart.image.series ({imageCollection: image_ndvi.select('NDVI'),
                                region: geometry,
                                reducer: ee.Reducer.mean(),
                                scale: 30})

print(chart)

The problem is that the functions do not load the data.

Would you like to help me get through this problem and load NDVI data by date into a DataFrame?

I would also like to have daily NDVI instead of monthly / yearly, but I'm not getting it on Google Earth Engine.

Thank you!

user_7918176 · Answer

I highly recommend checking out eemont, which makes all of this a breeze in python. You can find an example that fits your use case directly in their README:

import ee, eemont

f1 = ee.Feature(ee.Geometry.Point([3.984770,48.767221]).buffer(50),{'ID':'A'})
f2 = ee.Feature(ee.Geometry.Point([4.101367,48.748076]).buffer(50),{'ID':'B'})
fc = ee.FeatureCollection([f1,f2])

S2 = (ee.ImageCollection('COPERNICUS/S2_SR')
   .filterBounds(fc)
   .filterDate('2020-01-01','2021-01-01')
   .maskClouds()
   .scale()
   .index(['EVI','NDVI']))

# By Region
ts = S2.getTimeSeriesByRegion(reducer = [ee.Reducer.mean(),ee.Reducer.median()],
                              geometry = fc,
                              bands = ['EVI','NDVI'],
                              scale = 10)

Please note, that they are using sentinel-2 in their NDVI calculation, but this can easily be change to use LandSat.

On another note: Your script is not Python. If you really want our help to debug your current script, please remove all semicolons, stick to PEP8 formatting and include your imports as well.

Transform Google Earth Engine script to Python with landsat 8 temporal data

1 Answers