Pan-Sharpening on Landsat 8 image in GEE

Pan-sharpening is the process of enhancing a low-resolution multiband image by fusing it with a corresponding high spatial resolution panchromatic (single band) image such as a Landsat 8 image. The resulting multiband image has a resolution similar to the panchromatic image since the two rasters fully overlap. In the Google earth engine, the pan-sharpening process is made possible by using two methods; rgbToHsv () and hsvToRgb ().

This article focuses on outlining step-by-step the pan-sharpening process in the Google earth engine.

Select an area of interest.

Using the geometry tool, select your area of interest.

Add an image collection.

1. In this case, an image from Landsat 8 image collection was used. Perform a sort for the clouds so as to obtain the first image with the least cloud cover.
2. Proceed to perform a filter for the dates, to obtain the date the image was acquired and a filter for bounds to obtain an image for your area of interest.

var image = ee.Image(ee.ImageCollection('LANDSAT/LC08/C01/T1_TOA')
  .filterDate('2017-01-01', '2017-12-31')
  .filterBounds(geometry)
  .sort('CLOUD_COVER')
.first ());

Select the multiband and panchromatic images.

From the image collection, select a multiband image with bands 4, 3 and 2 representing red, green and blue bands respectively and a panchromatic image with only band 8 representing the panchromatic band (Landsat 8 Image)

var rgb = image.select('B4', 'B3', 'B2').clip(geometry);
var pan = image.select ('B8').clip(geometry);

Perform the pan-sharpening.

Pan-sharpening is done by converting the multiband image to HSV, swapping in the panchromatic band, and converting the image back to RGB.

var huesat = rgb.rgbToHsv().select('hue', 'saturation');
var upres = ee.Image.cat(huesat, pan).hsvToRgb();

Display before and after images using the same visualization parameters

Map.addLayer(rgb, {max: 0.3}, 'Original');
Map.addLayer(upres, {max: 0.3}, 'Pansharpened');

Original image

Pan-Sharpened Image