Remote Images, Let's Take a Closer Look!

"True Color" Image of the Cedar Falls Workshop Location	"False Color Infrared" Image of the Cedar Falls Workshop Location

Tell me more...

The two images above were created using the LANDSAT V satellite image provided by GLOBE. Multispec, a free image manipulation program, was used to create two very different displays of the same location. Both images represent the same 15 km x 15 km of land. Neither image is a photograph of the land. Instead the images are made up of pixels generated from data collected by LANDSAT V's sensors.

LANDSAT V is a satellite orbiting and collecting digital data about the earth. LANDSAT's sensors 'look' down on the earth and collect data on 5 wavelengths of light that are reflected off the earth's surface. Three of these wavelengths are Blue, Green and Red light. These are wavelengths that you can see. The last two sets of data are collected on Near Infrared (NIR) and Middle Infrared (MIR). The human eye can not see NIR or MIR light. In order for us to study the data LANDSAT collects on these two wavelengths we produce images which represent NIR and MIR light in colors we can see. Different surfaces reflect different amounts of each of the five wavelengths that LANDSAT measures. This means that by creating images from the satellite data we can distinguish between different types of surfaces on the ground (or different landcovers).

The image on the left represents the Cedar Falls Site in "true color". This means that blue light data from LANDSAT is show in blue on the image. In the same way green data is shown in green and red in red. The result is an image which looks very much like a photograph. The image on the right is very different. It shows NIR as red, red light as green, and green light as blue. Why would we want to look at this "false-color infrared" image? Because by looking at the infrared information it is possible to distinguish between some types of landcover that can not be seen in the true-color image. Using Multispec, you can look at any many different combinations of data.

Take a closer look at the images. Below are two close-ups. In the left close-up can you tell which green areas are trees and which are grasses? Now look at the right image. In this image trees tend to be display as a dark "true" red while grasslands show up as a brighter, almost neon red. The real test would be to go to this location and see just what is there!

In the images above, you can see a limit of the LANDSAT image data. The image is made of squares or pixels. Each pixel represents a 30 m x 30 m square of land. This means that the smallest thing that can be seen in a LANDSAT image must be at LEAST 30 m x 30 m. Each pixel is displayed as a weighted average of the data collected from everything within that 30 meter square space. This is another reason that ground validation on LANDSAT data is needed and why it is included in the Landcover Protocols.

Each GLOBE school can receive from The GLOBE Program, LANDSAT images of their study sites. These images are used in the Landcover Protocols, but they can also be used in mapping units or for other projects. The GLOBE Iowa workshop will include training in Multispec and more information on remote imaging and the LANDSAT program.