Supervised Classifications using Google Earth Engine - Revision history

Winaaakabzaa: /* Accessing and Working with Google Earth Image collections */

2022-10-07T20:21:29Z

Accessing and Working with Google Earth Image collections

Winaaakabzaa: /* Adding Legend to the map */

2022-10-07T20:19:42Z

Adding Legend to the map

Winaaakabzaa: /* Adding Legend to the map */

2022-10-07T20:07:07Z

Adding Legend to the map

Winaaakabzaa: /* Conclusion */

2022-10-07T20:05:56Z

Conclusion

Winaaakabzaa: /* Run on Chosen Image */

2022-10-07T19:58:35Z

Run on Chosen Image

Winaaakabzaa: /* Getting Started */

2022-10-07T19:48:57Z

Getting Started

Winaaakabzaa: /* Getting Started */

2022-10-07T19:34:15Z

Getting Started

Winaaakabzaa: /* Getting Started */

2022-10-07T19:22:07Z

Getting Started

Loucasdiamantboustea: /* Image Preperation */

2020-12-06T15:42:24Z

Image Preperation

Loucasdiamantboustea: /* Getting Started */

2020-12-06T15:34:12Z

Getting Started

@@ Line 45: / Line 45: @@
 You can add as many bands as possible (that are available in the data), however we will use Band 4, 3 and 2 to visualize the true color composite of the area. Also, the Map.setCenter, sets the center of the image to our selected area.
-[[File:Picture1.png|750px|thumb|center|Location of image settings, range, and band selection. Note the darkness of the image.]]
+[[File:Picture1.png|1000px|thumb|center]]
 ==Creating Geometry==

@@ Line 508: / Line 508: @@
 // add legend to map (alternatively you can also print the legend to the console)
 Map.add(legend);
-[[File:Picture3.png|500px|thumb|center]]
+[[File:Picture33.png|750px|thumb|center]]
 =Exporting the Classified Image=

@@ Line 508: / Line 508: @@
 // add legend to map (alternatively you can also print the legend to the console)
 Map.add(legend);
+[[File:Picture3.png|500px|thumb|center]]
 =Exporting the Classified Image=

@@ Line 442: / Line 442: @@
 Just like the histograms, the chart is interactable and you can hover over the classes to see their exact area. The chart will look something like this:
 [[File:Barchart.png|500px|thumb|center]]
+=Adding Legend to the map=
+Legends define features in a map, to someone without prior understanding of the color code, the map will pass no specific information. Therefore, it is important to add legends to serve as labels to distinctly differentiate features in the map. The legend represented the different land covers with their respective colors.
+<syntaxhighlight lang="javascript">
+var legend = ui.Panel({
+  style: {
+    position: 'bottom-left',
+    padding: '8px 15px'
+  }
+});</syntaxhighlight>
+ // Create legend title
+<syntaxhighlight lang="javascript">
+var legendTitle = ui.Label({
+  value: 'Legend',
+  style: {
+    fontWeight: 'bold',
+    fontSize: '18px',
+    margin: '0 0 4px 0',
+    padding: '0'
+    }
+});</syntaxhighlight>
+// Add the title to the panel
+<syntaxhighlight lang="javascript">
+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'
+        }
+      });</syntaxhighlight>
+      // Create the label filled with the description text.
+<syntaxhighlight lang="javascript">
+      var description = ui.Label({
+        value: name,
+        style: {margin: '0 0 4px 6px'}
+      });
+      // return the panel
+<syntaxhighlight lang="javascript">
+      return ui.Panel({
+        widgets: [colorBox, description],
+        layout: ui.Panel.Layout.Flow('horizontal')
+      });</syntaxhighlight>
+};
+//  Palette with the colors
+var palette =['1923d6', '6fba00', '000000', 'ef0b0b'];
+// name of the legend
+var names = ['Water','Vegetation','Burn', 'Urban'];
+// Add color and and names
+for (var i = 0; i < 4; 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);
+=Exporting the Classified Image=
+We can also export our result to our drive for further analysis. You can do this by creating a folder in your google drive, I have named mine ‘earth engine’.
+<syntaxhighlight lang="javascript">
+var exportImage = postBurn
+Export.image.toDrive({
+    image: exportImage,
+    description: 'LandCover',
+    folder: 'earthengine',
+    fileNamePrefix: 'LandCover',
+    region: extent,
+    scale: 10,
+    maxPixels: 1e9
+})
+There are so many things that you can do with Google Earth Engine. I have just scratched the surface with a random forest supervised classification. If you followed this tutorial you will have completed your own classification. I encourage you to try using different classification algorithms and experiment with different band combinations.
+If you have any questions about Google Earth Engine, be sure to check out the documentation here. It was written by the developers and is full of useful information and tips & tricks.
 =Conclusion=

@@ Line 310: / Line 310: @@
 As you can see, the classification appears to have worked fairly well. An important thing to note is that I only used 10 trees in my classification, increasing the number of trees may have made my model slightly more accurate. BUT you must remember that over modelling is just as bad as having an underperforming model.
+==Improving the classification results==
+We can improve the classification result, the availability of multiple datasets helps in improving image classification. We sometimes, just need to add more data such as the vegetation indices, built-up indices and all kind of indices that can easily distinguish land cover classes from one another.
+Adding indices
+BSI - Bare soil index
+NDVI - Normalized Difference Vegetation Index
+MNDWI -  Modified Normalized Difference Water Index
+NDBI - Normalized Difference Built_up Index
+<syntaxhighlight lang="javascript">
+var addIndices = function(image) {
+  var ndvi = image.normalizedDifference(['B5', 'B4']).rename(['ndvi']);
+  var ndbi = image.normalizedDifference(['B6', 'B5']).rename(['ndbi']);
+  var mndwi = image.normalizedDifference(['B3', 'B6']).rename(['mndwi']);
+  var bsi = image.expression(
+      '(( X + Y ) - (A + B)) /(( X + Y ) + (A + B)) ', {
+        'X': image.select('B6'), //swir1
+        'Y': image.select('B4'),  //red
+        'A': image.select('B5'), // nir
+        'B': image.select('B2'), // blue
+  }).rename('bsi');
+  return image.addBands(ndvi).addBands(ndbi).addBands(mndwi).addBands(bsi)
+};</syntaxhighlight>
+==Normalizing data inputs==
+The best data inputs for machine learning models are the inputs who are same scaled. By normalizing the inputs, the accuracy of the model is increased. Normalization of input data ensures that each individual inputs have values ranging between 0-1
+<syntaxhighlight lang="javascript">
+function normalize(image){
+  var bandNames = image.bandNames();
+  // Compute min and max of the image
+  var minDict = image.reduceRegion({
+    reducer: ee.Reducer.min(),
+    geometry: boundary,
+    scale: 10,
+    maxPixels: 1e9,
+    bestEffort: true,
+    tileScale: 16
+  });
+  var maxDict = image.reduceRegion({
+    reducer: ee.Reducer.max(),
+    geometry: boundary,
+    scale: 10,
+    maxPixels: 1e9,
+    bestEffort: true,
+    tileScale: 16
+  });
+  var mins = ee.Image.constant(minDict.values(bandNames));
+  var maxs = ee.Image.constant(maxDict.values(bandNames));
+  var normalized = image.subtract(mins).divide(maxs.subtract(mins))
+  return normalized
+} ;</syntaxhighlight>
 ===Calculating Area of Each Class===

@@ Line 63: / Line 63: @@
 '''NOTE:''' Mine is named 'extent'. You will see it in my code.
+=Important Data=
+Also most times, we can have a pre-defined area extent or geometry in form of vector or raster data, such data can be used to subset and define our area extent. The data on our computer, can be imported into google earth engine, by using the Code Editor page,
+●	Go to Assets
+●	Then Click New
+●	You will see Table Upload
+●	Click Shapefiles (if you have a shapefile or CSV)
+[[File:Picture2.png|750px|thumb|center]]
+Select all files accompanying the data you want to import, name it accordingly and click Import.
 ==Displaying Images==

← Older revision		Revision as of 19:34, 7 October 2022
Line 45:		Line 45:
	You can add as many bands as possible (that are available in the data), however we will use Band 4, 3 and 2 to visualize the true color composite of the area. Also, the Map.setCenter, sets the center of the image to our selected area.		You can add as many bands as possible (that are available in the data), however we will use Band 4, 3 and 2 to visualize the true color composite of the area. Also, the Map.setCenter, sets the center of the image to our selected area.

			[[File:Picture1.png\|750px\|thumb\|center\|Location of image settings, range, and band selection. Note the darkness of the image.]]

@@ Line 17: / Line 17: @@
 **It can get a little hectic in the API, so splitting things up is a nice housekeeping habit
 **I am using a single small screen, so I could be biased on this point
+==Accessing and Working with Google Earth Image collections==
+The datasets available on Google earth engine are mostly in form of collections, understanding how to access the wide and numerous datasets on google earth engine, is very important. Image collections are images provided by a single dataset provider but taken at different times and locations. For this tutorial, we will be making use of the Landsat 8 data. You can search the Dataset catalog to find more datasets available to work with. Visit the  Landsat 8 Level 2 data, and copy the code to access it.
+<syntaxhighlight lang="javascript">
+var dataset = ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')
+    .filterDate('2021-05-01', '2021-06-01');
+// Applies scaling factors.
+function applyScaleFactors(image) {
+  var opticalBands = image.select('SR_B.').multiply(0.0000275).add(-0.2);
+  var thermalBands = image.select('ST_B.*').multiply(0.00341802).add(149.0);
+  return image.addBands(opticalBands, null, true)
+              .addBands(thermalBands, null, true);
+}
+dataset = dataset.map(applyScaleFactors);
+var visualization = {
+  bands: ['SR_B4', 'SR_B3', 'SR_B2'],
+  min: 0.0,
+  max: 0.3,
+};
+Map.addLayer(dataset, visualization, 'True Color (432)');
+};</syntaxhighlight>
+You can add as many bands as possible (that are available in the data), however we will use Band 4, 3 and 2 to visualize the true color composite of the area. Also, the Map.setCenter, sets the center of the image to our selected area.
 ==Creating Geometry==

@@ Line 58: / Line 58: @@
 ==Image Preperation==
-Google Earth Engine has a large selection of remote sensing data to choose from. In this example, I will be using the '
+Google Earth Engine has a large selection of remote sensing data to choose from. You can easily search for anything in the search bar. In this example, I will be using the 'USGS Landsat 8 Collection 1 Tier 1 Raw Scenes' collection. This is a collection of the highest quality images available through Landsat 8.
-USGS Landsat 8 Collection 1 Tier 1 Raw Scenes' collection. This is a collection of the highest quality images available through Landsat 8. Throughout the process of choosing imagery, I suggest taking as many opportunities to view the images generated. This will allow you to manually choose if necessary. Adding the layers as we progress through the steps is also recommended as it will give you a better idea of what exactly each step does. By running the following code you can see the data for an image or an image collection. It will include specifics like date, and band or bit values.
+Throughout the process of choosing imagery, I suggest taking as many opportunities to add the images to the map between processing steps. It will help your understanding of each step. Another tip, by running the following code you can see the data for an image or an image collection in the tab on the right of the screen. It will include specifics like image date, and band or bit values.
 <syntaxhighlight lang="javascript">
@@ Line 87: / Line 88: @@
 };</syntaxhighlight>
-Now onto the composite. This is a much simpler function, it simply intakes a collection of images and uses the built in 'simpleComposite' function to combine them into the best possible image.
+Now onto the composite. This is a much simpler function, it intakes a collection of images and uses the built in 'simpleComposite' function to combine them into the best possible image.
 <syntaxhighlight lang="javascript">
 // Function to create composite

@@ Line 11: / Line 11: @@
 Some tips before we get going:
 *Things will save to you google drive, so be sure it is linked by signing in
-*Create a folder that in the left tab
+*Create a folder in the left tab
 *Always save after doing things
 **It will save when you run the program
@@ Line 19: / Line 19: @@
 ==Creating Geometry==
-The first step is to create geometry representing the area you want to look at. This step will comprise of creating a polygon that will be used to find clip images.
+The first step is to create geometry representing the area you want to look at. This step will comprise of creating a polygon that will be used to find and clip images.
 To do this, follow these steps: