GEE-S2

A simple tutorial to download Sentinel-2 images over a specific administrative area using Google Earth Engine platform.

Step 0: Introduction

The tutorial is for: new users of GEE, and want to extract Sentinel-2 (or any others) time series over a certain region. There are many useful repositories on github.com to download remote sensing images on GEE, such as

• Sentinel2_GEE
• GEE-to-NPY
• geedim

But as far as I know, none of them are designed for the purpose as this tutorial.

Step 1: generate shape file of your ROI

The first step is to generate a shape file of your ROIs, which can be in various format, such as .shp. There are many ways to do it. One of the ways (that I used) is downloading via DIVA-GIS.

• download the shape files from DIVA-GIS website.
• Load the shape file in QGIS, select the small ROI (eg. 'Bayern, Germany') and save to a shp. file.

Step 2: setup on Google Earth Engine (GEE)

It is suggested to read some guidelines for the new users of GEE before continuing read this tutorial. Some recommandations:

• https://www.google.com/earth/outreach/learn/introduction-to-google-earth-engine/
• https://blog.csdn.net/qq_22865459/article/details/80614822

Then, upload your shape file (in step 1) to GEE Assets. And create a new project and script.

Note: to access to GEE, you need an account; to save the downloaded data, you need a Google drive account.

Step 3: download on GEE

This tutorial is based on the one posted at https://zhuanlan.zhihu.com/p/366744507.

Step a: load the shape file

var roi = ee.FeatureCollection("projects/bayernforest/assets/bayern");

Step b: load the necessary functions

var batch = require('users/fitoprincipe/geetools:batch')
var GuanMethod = require("users/GHX/share:function_Library.js")

Step c: mask-out clouds

function maskS2clouds(image){
  var qa = image.select("QA60");
  
  var cloudBitMask = 1 << 10;
  var cirrusBitMask = 1 << 11;
  
  var mask = qa.bitwiseAnd(cloudBitMask).eq(0)
               .and(qa.bitwiseAnd(cirrusBitMask).eq(0));
  
  return image.updateMask(mask).divide(10000)
         
              .set('system:time_start', image.get('system:time_start'));
}

Step d: collect Sentinel-2 data

var dataset = ee.ImageCollection('COPERNICUS/S2_HARMONIZED') 
                  .filterDate('2020-08-01', '2020-08-30')
                  .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',5)) 
                  .map(maskS2clouds)
                  .filterBounds(roi); 

You can change the date range that you want to get a mosaic of Sentinel-2 image over the ROI. Set the CLOUDY_PIXEL_PERCENTAGE as the rate of cloud coverage.

Step e: save the data to your Google drive folder

Export.image.toDrive({
  image: dataset.mean(),
  description: 'bayern-2015-5-8',
  folder: 'bayern-forest/2015-5-8',
  scale: 10,
  region: roi,
  maxPixels:34e10
});