Google Earth Engine

Getting Started

  • Render Sentinel 2 imagery
  • Define region of interest (ROI)
  • Create a cloudless orthomosaic 
  • Export an image
  • import a different image to your project
  • add legend
  • share with the world as an app

What is Google Earth Engine? 

  • Computing platform and runs in your internet browser
    • with a code editor,
    • offers a compiler that runs in the Google cloud,
    • hosts data on Google servers.
  • Integrated Development Environment (IDE)

 

Google Earth Engine is a

Data Repository (data as a service)

Software as a Service (SaaS)

Platform as a service (PaaS)

 

 

 

User Interface

Take a look at available data here

Google Earth Engine Data Catalog 

https://developers.google.com/earth-engine/datasets

 

As an example: 

sentinel 2  It provides data suitable for assessing state and change of vegetation, soil, and water cover.


https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED

 

Let's Look at the Meta Data (data about the data)  together

  • Read the description of the dataset.
    • who supports this mission/dataset?
    • Is this dataset global? 
  • How many bands are there?
  • What are some of the other image properties?

ANSWERS:

  • Read the description of the dataset.
    • who supports this mission/dataset? -->Copernicus Land Monitoring studies- European Space Agency
    • Is this dataset global? -->WARNING: 2017-2018 L2 coverage in the EE collection is not yet global.
  • How many bands are there? --> 12 Bands 
  • What are some of the other image properties?

{JavaScript}

is an interactive front-end scripting language

Comments - explains what code does to the reader but not the compiler

// Displays a map

 

JavaScript is case sensitive

 

 

Statements - individual instruction or steps

statements end with a ;

 

Variable - tell the system to remember or store data or a value

 

First Declare a variable

 

Then assign a value to a variable

 

{Hello World}

Let's try some simple examples

Add Data and show it on the map

cut and paste this code into your code window

// call a country border vector file to set the region of interest (roi) from datasets - you are welcome to pick a different country.
var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Aruba');


////select images only from Jan - the Wet Season in Aruba
var wet_S2_collection = ee.ImageCollection('COPERNICUS/S2_HARMONIZED')
  .filterDate('2025-01-01', '2025-01-31')
  .filterMetadata('CLOUDY_PIXEL_PERCENTAGE', 'less_than', 5);
  
  //reduce to an image to the median value and clip to area
var wet_S2_mosaic = wet_S2_collection.median().clip(roi);

//call bands from Sentinel 2

var S2_display = {bands: ['B4', 'B3', 'B2'], min: 0, max: 3000};


//render map
Map.addLayer(wet_S2_mosaic, S2_display, "Wet True Color");

//center the map on the roi
Map.centerObject(roi);

What do you see? 

select the Inspector tab on the top right

Then Click anywhere in your map  

Notice there is a DN for each band  

Example

In this example, it renders the sentinel images for the entire world, for a given date range. 

 

 

Now let's change

  • the date range
  • the center of the map - by moving the geometry box
  • to save energy - clip to the region of interest set when you manually drew that box

set the geometry manually - draw the box

in the map interface and adjust the transparency - add a line of code and then hit run

//call sentinel 2 images - specify date and cloudless pixels
////notice that the geometry is not set - no geometry variable yet! You use the square box in the map pane to draw your bounding box and those coordinates will be generated. 
var S2_collection = ee.ImageCollection('COPERNICUS/S2_HARMONIZED')
  .filterBounds(geometry)
  .filterDate('2025-05-01', '2025-06-14')
  .filterMetadata('CLOUDY_PIXEL_PERCENTAGE', 'less_than', 10);
 
  
 //calculate and select the median value of each cloudless pixel from the above date range and clip to the roi/geometry set

var S2_mosaic = S2_collection.median().clip(geometry);


 //tell which bands to make visible in the map below
var S2_display = {bands: ['B4', 'B3', 'B2'], min: 0, max: 3000};


 //add the following variable to be visible in the map below
Map.addLayer(S2_mosaic, S2_display, "Sentinel-2");

//center the map on the point of interest defined

Map.centerObject(geometry);

Now change some of the parameters

just to explore

change the band combinations

Change the date ranges

change the band combinations

B8, B4, B3 to see healthy vegetation

Change the date ranges

export the image mosaic

//To export your image add the following code to the bottom of your existing code
// Dont forget to change the name of the file to reflect the region you picked. 
Export.image.toDrive({
image: S2_mosaic,region: geometry,
 description: 'Utrecht May-June2025',
 scale: 30,

});

full code

Subtitle

var geometry = //set this using the method described with drawing a box or bringing in your own shp file

//call sentinel 2 images specify date
var S2_collection = ee.ImageCollection('COPERNICUS/S2_HARMONIZED')
  .filterBounds(geometry)
  .filterDate('2025-05-01', '2025-06-14')
  .filterMetadata('CLOUDY_PIXEL_PERCENTAGE', 'less_than', 10);
 
  
 //calculate and select the median value of each cloudless pixel from the above date range and clip to the roi/geometry set

var S2_mosaic = S2_collection.median().clip(geometry);


 //tell which bands to make visible in the map below
var S2_display = {bands: ['B4', 'B3', 'B2'], min: 0, max: 3000};


 //add the following variable to be visible in the map below
Map.addLayer(S2_mosaic, S2_display, "Sentinel-2");

//center the map on the point of interest defined

Map.centerObject(geometry);

// Export the sentinel 2 image
Export.image.toDrive({
image: S2_mosaic,region: geometry,
 description: 'Utrecht May-June2025',
 scale: 30,

});

I removed geometry variable - set it yourself by drawing a bounding box

  1. Hit Run to run the code
  2. Then the Tasks Tab should turn orange
  3. Click it
  4. You will see the file name will appear with a button that says Run
  5. click run to start the export process
  6. It will be saved to the root directory of your Google Drive (unless you specify otherwise)

To export the image

ArcGIS Pro classification wizard

Now go through the assignment in the word document

answer the questions along the way 

AFTER you have classified your image in ArcGIS Pro

Bring your classifed image into GEE

In GEE navigate to Assets-->New-->GeoTiff and select your classified image

 

may take a few minutes to show up in your assets folder

 

Then click the arrow to put it in your script

Set the visualization parameters and hit run

make your map interactive

import your final classified image into Google Earth Engine

When you want to put your final class image online

enter this code in Google Earth Engine

//first set the var image: by following directions in the video in the last slide - this code alone will not work without it
//your classified image will have as many classes as you set - mine had six - I used colorbrewer2.org for color advice - make yours better :)
var imageViz = {min: 0, max: 6, palette: ['e41a1c', '377eb8', '4daf4a', '984ea3', 'ff7f00', 'ffff33']};
//call your variable image to render, call variable to stylize it, set the visible layer name, and true to render it
Map.addLayer(image, imageViz, 'classes', true);
//tell map to center on your classified image
Map.centerObject(image);

what is missing? 

A Legend! Here is the code to add one. Read comments carefully so you can replace my colors and classes with your own. 


//your classified image will have as many classes as you set - mine had six - I used colorbrewer2.org for color advice - make yours better :)

var imageViz = {min: 0, max: 6, palette: ['e41a1c', '377eb8', '4daf4a', '984ea3', 'ff7f00', 'ffff33']};

//call your variable image to render, call variable to stylize it, set the visible layer name, and true to render it

Map.addLayer(image, imageViz, 'classes', true);
//tell map to center on your classified image
Map.centerObject(image);

//now add a legend
// set position of panel of the legend
var legend = ui.Panel({
  style: {
    position: 'bottom-left',
    padding: '8px 15px'
  }
});
 
// Create legend title
var legendTitle = ui.Label({
  value: 'Land class Legend',
  style: {
    fontWeight: 'bold',
    fontSize: '18px',
    margin: '0 0 4px 0',
    padding: '0'
    }
});
 
// Add the title to the panel
legend.add(legendTitle);
 
// Creates and styles 1 row of the legend.
var makeRow = function(color, name) {
 
      // Create the label that is actually the colored box.
      var colorBox = ui.Label({
        style: {
          backgroundColor: '#' + color,
          // Use padding to give the box height and width.
          padding: '8px',
          margin: '0 0 4px 0'
        }
      });
 
      // Create the label filled with the description text.
      var description = ui.Label({
        value: name,
        style: {margin: '0 0 4px 6px'}
      });
 
      // return the panel
      return ui.Panel({
        widgets: [colorBox, description],
        layout: ui.Panel.Layout.Flow('horizontal')
      });
};
 
//  Palette needs to match your map - copy and paste the colors you used above
var palette =['e41a1c', '377eb8', '4daf4a', '984ea3', 'ff7f00', 'ffff33'];
 
// name of the legend need to match your classes
var names = ['urban','rural','water','agriculture','semi-urban','other'];
 
// Add color and and names change the number 6 to how many classes you have
for (var i = 0; i < 6; i++) {
  legend.add(makeRow(palette[i], names[i]));
  }  
 
// add legend to map (alternatively you can also print the legend to the console)
Map.add(legend);

Alternatively - do the whole assignment in Google Earth Engine   

// Load a Landsat composite for input.
var input = ee.ImageCollection('LANDSAT/COMPOSITES/C02/T1_L2_32DAY')
  .filterDate('2001-05', '2001-06')
  .first()
  .clip(geometry);
  

Map.centerObject(geometry);
Map.addLayer(ee.Image().paint(geometry, 0, 2), {}, 'region');

// Make the training dataset.
var training = input.sample({
  region: geometry,
  scale: 30,
  numPixels: 5000
});

// Instantiate the clusterer and train it.
var clusterer = ee.Clusterer.wekaKMeans(15).train(training);

// Cluster the input using the trained clusterer.
var result = input.cluster(clusterer);

// Display the clusters with random colors.
Map.addLayer(result.randomVisualizer(), {}, 'clusters');

for Sentinel 2 see this good tutorial

And tutorial from google

Give me more!

Read more here

want more tutorials? https://tutorials.geemap.org/Image/image_visualization/

advanced https://courses.spatialthoughts.com/end-to-end-gee.html

 

Made with Slides.com