This tutorial will help you use downloaded imagery from Radiant.Earth in an open-source GIS environment (specifically Quantum GIS). This process will allow you to use imagery that you find here with other GIS projects.
By the end of this tutorial, you will be able to display images in True Color (the visual display of color that you're used to) and in False Color (which uses the reflective properties of plants to highlight vegetated areas).
Start by locating the scene you want to download from Radiant.Earth, clicking on the scene, and then clicking
Download. In this case, we'll display the San Francisco Bay Area.
Every scene is composed of an overlay of several different bands of light. Landsat 8, for example, has 11 different bands, including the normal Red, Green, and Blue, but also Near-Infrared, Thermal Infrared, and bands targeting aerosols and cloud reflectance. Each band by itself looks like a gray-scale or mono-chromatic image, but when re-displayed in Red, Green, and Blue, the image will have a nice color overlay.
Each band is assigned a number. For this project, you'll want Red (4), Green (3), and Blue (2) for True Color, and the addition of Near Infrared (5) for false color.
To select bands to download, click the Plus Sign (+) next to "Images" to expand the menu and show all the different images that are layered within this scene. Find the files that end with ...
_B3.TIF, and ...
_B2.TIF and click the cloud button next to them to download.
These are large images, so it might take a moment for the download to complete. Radiant.Earth provides "Geotiffs," which are TIFF images that are encoded with geographic location, so they will be displayed properly in GIS.
You might want to move these images into their own folder for processing.
Then, open Quantum GIS (QGIS) or download it here. QGIS is free and open source, and is useful for GIS applications, as well as for displaying raster images. If you want to do deeper analysis of your imagery, check out GRASS GIS.
We'll display the True Color image first.
In QGIS, from the top toolbar select
Input and select bands 2, 3, and 4 (Blue, Green, and Red respectively).
Select an output location and name, and then check the box next to "Place each input file into a separate band." When you're ready, hit
This will project the image, but with the wrong color values assigned. If the "Layer Styles" menu doesn't pop up, double click on the layer name in the Layers Panel.
In the layer styles menu, you'll need to reassign which bands are which. In this case, make the Red "Band 3," keep the Green "Band 2," and make the blue "Band 1." These names do not correspond to the official band numbers.
The Max/Min values are calculated automatically by QGIS for better visual display. In order to get a more accurate picture, you'll need to readjust the max values to include the full radiometric spectrum. These Landsat images are stretched to 16-bit, so you'll want to make the max level
65535. Make sure the style layer matches the box, and then hit
Your display might come out very dark, since you're using the correct maximum, like in the image below.
To fix this, go back into Layer Styles and adjust brightness, saturation, and contrast to your liking. These settings are for display purposes only, and won't affect your analysis. To make it easier to adjust settings, hit
F7 on your keyboard to toggle the layer styles menu on the side.
You can follow the same process for a False Color image. In this case, open the Merge tool and select Bands 3 (Green), 4 (Red), and 5 (Near-Infrared) for your input images. Then in the layer styles, make Red "Band 3," Green "Band 2," and Blue "Band 1."
Make sure to readjust the maximum values, and then click
OK to display the image (like below).
You'll notice that the areas filled with vegetation in the True Color stand out much more in the False Color image (in the red). Vegetation reflects Near-Infrared light at much higher rates than red or even green, so a False Color image can help vegetation stand out.
Now that you've displayed your images using QGIS, you'll be able to integrate them into larger GIS projects or perform heavier analysis.