[GIS] Google Earth Engine: different stretch options

google-earth-enginesentinel-1

I'm working on a map view that detects new ice using Sentinel-1 together with Google Earth Engine.

When adding images to Google Earth Engine I can set the min and max values for my layer in numbers but then in the map view I have the possibility to change change these values and also select different stretch options.

Is it possible to choose one of these options instead of entering a max and min view in the code? Right now my code end looks like this:

Map.addLayer(hh2017, {

bands: ["HH"],
max: -5,
min: -19,
opacity: 1,
palette: ['002bff','a4d5ff','979797','080358','000029' ]

}, 'sentinel-1');

I want it to be Stretch: 90% The reason I need to stretch the image each time is that they all look a bit different.

Best Answer

As described here, you may want to try a Styled Layer Descriptor. The previous answer also works, but requires you to use getInfo(), which can cause your browser to lock (see Client-Server doc for details). Here's another approach:

var i = ee.Image(ee.ImageCollection('LANDSAT/LC08/C01/T1_TOA')
    .filterBounds(Map.getCenter())
    .sort('CLOUD_COVER')
    .first());
Map.centerObject(i);
Map.addLayer(i);

function setDisplay(dictionary) {
  var visParams = {
    bands: ['B4', 'B3', 'B2'], 
    min: [
      dictionary['B4_p5'], dictionary['B3_p5'], dictionary['B2_p5']
    ],
    max: [
      dictionary['B4_p95'], dictionary['B3_p95'], dictionary['B2_p95']
    ]
  };
  Map.layers().set(0, ui.Map.Layer(i, visParams));
}

Map.onChangeBounds(function() {
  var params = i.reduceRegion({
    reducer: ee.Reducer.percentile([5, 95]), 
    geometry: Map.getBounds(true), 
    scale: Map.getScale(),
  });
  params.evaluate(setDisplay);
});

Related Solutions

Google Earth Engine – How to Perform Linear Regression on Seasonal Dataset

I guess you were trying to follow this example by Google: example

I am not familiar with the results of this regression, but I adapted your code so it works similar to the example.

In your for loop, you ended up with one image and added every years precipitation as a new band. The input of the linearFit you are using in the example asked for an image collection as input with bandnames similar in every image. If you change this lines at the end of your loop, you will return an image collection without the dummy image:

// create a blank image
var imageblank = ee.Image();

var list = [];
  for (var i = startDate; i <= endDate; i++) {
    var iMod = i - 1;
    var collectionDates = collection.filterDate(iMod+startMonthDay, i+endMonthDay);
    var collectionMean = collectionDates.sum().clip(Zimbabwe);

    var collectionDates2 = collection.filterDate(i+startMonthDay2, i+endMonthDay2);
    var collectionMean = collectionDates2.sum().clip(Zimbabwe).set('year', ee.Number(iMod).add(1));

 //create blank image and use .merge to add images to the image collection
 imageblank = ee.ImageCollection(imageblank).merge(collectionMean);
 }

// filter out the blank image
var collectionPrec = imageblank.filter(ee.Filter.notNull(['year']));

You can then follow the example by Google to calculate and display the linear regression using linearFit:

// Reduce the collection with the linear fit reducer.
// Independent variable are followed by dependent variables.
var linearFit = collectionPrec.select(['system:time_start', 'precipitation'])
  .reduce(ee.Reducer.linearFit());  

// Display the results.
Map.addLayer(linearFit, 
  {min: [0, 150, 0], max: [-0.9, 1700, 1], bands: ['scale',  'offset', 'scale']}, 'fit');

Link to full code

Google Earth Engine Map.addLayer – Fixing Issue with Adding Bands

(Note: I know nothing about appropriate visualization of "EVI" data; my suggestions in this answer are general advice on Earth Engine programming and not specifically recommended as good visualizations.)

I know that these bands do exist…

You checked img161, where those band names do exist, but you're trying to visualize evi, which does not have those bands — it has one band, which you just computed using an expression:

var evi = img161.expression(
    '2.5 * ((NIR - RED) / (NIR + 6 * RED - 7.5 * BLUE + 1))', {
      'NIR': img161.select('B8'),
      'RED': img161.select('B4'),
      'BLUE': img161.select('B2')
});

When you compute an image by doing arithmetic on single band images (as you have provided here using .select()), the output is also single-band (and the band name is often irrelevant — here it is 'constant' because the very first input is the constant-image-value 2.5).

You have one band, but your evivisparam specifies three bands. The question, then, is: what kind of visualization do you want?

If grayscale will do, remove bands from the visualization parameters.
```
var evivisparam = {min: -1, max: 1};
```

If you want a color gradient, specify a palette in the visualization parameters.

var evivisparam = {min: -1, max: 1, palette: ['FF0000', '000000', '00FF44']};

If you want one color channel to be the EVI value and others to be the original bands, then you need to recombine the computed EVI band with the original image bands, and specify suitable visualization parameters for red, green, and blue:
```
var evivisparam = {
  min: [-11000, -1, -11000],
  max: [11000, 1, 11000],
  bands: ['B4', 'evi', 'B2']

};
var img161WithEvi = img161.addBands(evi.rename('evi'));
Map.addLayer(img161WithEvi, evivisparam, 'EVI img161');
```

Best Answer

Related Solutions

Google Earth Engine – How to Perform Linear Regression on Seasonal Dataset

Google Earth Engine Map.addLayer – Fixing Issue with Adding Bands

Related Question