qplogo
 
Web Quantum Petshop
projectstutorialsaboutqpqpstorepressroomcollaborationcontactlinks

Getting and Using Satellite Image Data
Access raw LANDSAT satellite imagery and use it for graphical work.
By Ryan Foss

This tutorial describes how to get raw Landsat 7 satellite data from NASA repositories and convert it into usable, full color imagery.

 

1


Obtaining Raw Landsat-7 Data

First you have to find your area of interest. Go to the Earth Science Data Interface (ESDI) and click on Map Search. Map searching will allow you to browse the world for various image data using an interactive map.

Since you want to search the datasets for Landsat-7 image data, check the ETM+ checkbox on the upper left.

The Map Search Tool makes it easy to find image data. The tool is pretty self explanatory, so just try it. The easiest way is to just click the map until you're close, then select the arrow+ icon and click again.

What you will end up with is an area selection similar to Figure 2.

Now click on the Preview & Download button to get access to the image datasets. If there is more than one dataset available, you'll need to select one that fits your desires, either by date or area coverage. Select the dataset by clicking the [ID], then click the download button. This will bring you to direct FTP access to the image data.


Figure 1. The Map Search tool at the ESDI web site makes it easy to find image data.

Landat-7 is a multi-band satellite, thus the data is represented with multiple grayscale images, each representing different spectrums of light. Since we want a true-color image, we need to work with the visible spectrum. You'll need to download three images, each representing the basic color spectrums of an RGB image, red, green, blue.

The names of the images may seem confusing, but what's important is the number at the end of the filename.


Figure 2. Single band image from a Landsat-7 dataset.

Download the first three bands (files ending with numbers 10, 20, and 30), uncompress them and open them in Photoshop. Without getting to much in the details, you'll want to download files that look similar in format to these. They (most likely) are compressed TIFF files, so you'll need a program such as WinZIP or WinRAR to extract the images.

  • p027r029_7t20010705_z15_nn10.tif.gz
  • p027r029_7t20010705_z15_nn20.tif.gz
  • p027r029_7t20010705_z15_nn30.tif.gz

FYI, the file names are actually useful, corresponding to Path and Row information, satellite and capture date. The image dataset above, p027r029 is path 27, row 29. This makes it easy to find the dataset again, if you should need to come back to it.

2


Combining Bands into True Color

Each of these bands represents a basic spectrum of color. We need to combine them to produce a true color image.

  • Band 1 is Blue
  • Band 2 is Green
  • Band 3 is Red
With the three images open in Photoshop, go to the channels pallet. Using the little triangle icon on the right, select “Merge Channels”. Merge the channels into a 3 color RGB image. Set the bands appropriately, 3 is red, 2 is green and 1 is blue.

Figure 3. Photoshop channel pallet.

Note: There is another method to combine bands. Essentially, open each image and copy paste it into the appropriate color channel. Here’s step by step how to. Open the band 3 image and convert it to RGB via Image>Mode>RGB Color. Open band 2 image and select all (Ctrl-A) and copy (Ctrl-C). Close the band 2 image and go to the Channel pallet of the band 3 image. Select the Green channel and paste. Open band 1 and repeat the process, copy it to the Blue channel.

3


Color Balance

This is the most difficult step. Depending on the quality of the images from the satellite, and other factors like time of day or haze. Regardless, you'll most likely want to adjust the color. Image>Adjustments>Auto Color usually does a decent job.

There you have it. This image data is in 30m data. This means that each pixel represents approximately 30 meters length. Not really good, but does well for large areas. This can be improved upon by using Band 8, the panchromatic band, to effectively double the resolution to 15 meter. I outline the methods in the next tutorial: Doubling Landsat-7 Image Resolution.

If you’re interested in better resolution imagery, then you’re going to have to pay a service.

 

Don't forget to checkout our entire Tutorials section.

Our Other Popular Tutorials
Found our stuff useful or interesting? Help us function and keep us motivated.




all content © Copyright Quantum Petshop 1998-2006

quantum petshop tresure box
[an error occurred while processing this directive]