Module 2 - How Radar Imaging Works

C) Looking at a Radar Image

Objectives

Students will be able to display radar images on the computer using the program Sigma0. Students will learn how to open image files, adjust the histogram for display, and obtain backscatter values from the image.
Students will be able to display different polarization radar images on the computer using the program Sigma0. Students will learn that comparing radar backscatter from different polarizations can give information about the scatterer.
Students will learn how to save a radar image using the program Sigma0.

Image data for Module 2 contained in the /DATA/MODS1TO5/MODULE02/IMAGES/AIRBORNE and /DATA/MODS1TO5/MODULE02/IMAGES/SIR-C directories should be copied over to your computer before beginning this module. The files you will need are:

SanFranc.stk, SanFrkey.gif, SanFranc.gif and SanFranc.txt

To display the SanFranc.stk image, you need to run the Sigma0 program then use the Open Compressed Stokes option under the File menu. If you select the default option (TP) in the menu that pops up and in the histogram window that follows by clicking the OK button, an image should appear on your screen.

What is Sigma0?

A description on how to use Sigma0 is available. When you run Sigma0, you will see a window appear with the words Sigma0 in the title bar. Go to the File menu item. The following window should appear (or something similar):

[SIGMA0 FILE MENU]

Highlight the Open compressed stokes option. A window will then appear for you to select a directory/filename. Go to the CD directory /DATA/MODS1TO5/MODULE02/IMAGES/AIRBORNE and select the file SanFranc.stk to open it. You will now be given the Options window:

[SIGMA0 OPTIONS WINDOW]

Select VV and click on OK. Another window will pop up:

[SIGMA0 HISTOGRAM WINDOW]

Click OK for now. We'll take a look at what this does a little later.

An image of San Francisco as taken in 1985 by a JPL airborne imaging radar will pop up on your screen.

San Francisco image displayed using Sigma0

Run the cursor back and forth onto the Golden Gate bridge. As you are doing that, notice what is happening with the three values at the bottom left of the image. Those values tell you, in order, the X and Y coordinates of the cursor and the reflectivity of parts of the image to radar.

This value is the normalized radar cross section ( ). It ranges from about -40 to +5, where -40 represents almost no reflections of the radar pulses and +5 represents very high reflectiveness. In the Sigma0 images, -40 is represented as black and +5 as white, with gray shades representing the numbers in between.

To display the key to the San Francisco image, use the Finder to locate the SanFrkey.gif file.

Go back to Sigma0 using the Finder or redisplay the SanFranc.stk image if you quit Sigma0. Now go to Image | Zoom in. You have magnified the image and not all of it can be seen any more. You can help this by moving the cursor to the small box at the bottom right of the image, click and hold the mouse while pulling the box towards the bottom right of the computer screen. Now you can see most of the image.

Go to a spot just above and left of the middle of the Golden Gate Bridge. Click and hold the mouse and move the cursor to the right of the bridge and below the middle. You should have created a box with "ANTS" marching around again. Now go to Image | Zoom in once more.

Zoom in once more. Find X,Y coordinates 110, 55 now. You will notice that the image seems to be made up of a lot of little boxes - we call these picture elements or pixels. Each pixel contains one bit of information. Although we can make it bigger, we can't get more information from it.

Now close this image of San Francisco (click in the box at top left or go to File | Close). Now open SanFranc.stk again (File | Open Compressed Stokes..., VV, OK). Again you should see the histogram window below. This is a histogram (a graph) of the number of pixels that have particular reflectiveness, ranging from very dark and unreflective (left) to very light and reflective (right). San Francisco seems to have a lot of unreflective area. Let's take a look at only the portion of the histogram that has a lot of pixels. Bring the tip of the cursor into the gray bar at the bottom of the window, click and hold the mouse, and then move the arrow inside the bar to the position shown below. Let up the mouse button:

[SIGMA0 HISTOGRAM WINDOW]

Click OK. You will find that the image displayed will appear lighter than the previous image, because you have instructed the computer to make the medium gray pixels into light gray and turn all of the old light gray values into white. This change in the gray scale lets you see more contrast between regions of the image that seem to look a lot alike. This is often called Enhancing the contrast. Bring the cursor back to the Golden Gate bridge, at X = 110 and Y = 55. The number for should not have changed.

This is important. Making changes that help you see the things in the image better DOES NOT change the data. The radar reflectiveness of that portion of the Golden Gate bridge HAS NOT CHANGED just because we changed the way the bridge looks on the computer

What are those other choices under Sigma0 Compressed Stokes?

The radar sends out and collects horizontally and vertically polarized waves. Some horizontal waves are return unchanged, as horizontal waves. These are represented in the Compressed Stokes table as HH. These would be bouncing back (backscattering) mostly off of objects that are larger than the waves or are mostly oriented horizontally. In the same way, some vertical waves are sent out and collected unchanged. These are represented as VV and would be backscattered mostly off of objects that are upright.

[SIGMA0 OPTION WINDOW]

There are of course lots of objects that are tipped somewhat sideways and are neither horizontal nor vertical. Tree branches are a good example of these. They may be pointing vertically, horizontally, or any which way in between.

Sometimes a horizontal wave might bounce in such a way that it gets flipped to vertical - or a vertical wave might get flipped to horizontal. This is what is collected as HV.

The correlation coefficient is a measurement of how much alike the HH and the VV scatterers are, expressed as a percent. They could be completely different (0 % alike) or they can be completely similar (100% alike). Thus, correlation coefficient is a measurement expressed from 0 to 100.

Also of interest are the phase measurements, HHVV, HVVV, and HHHV. These give some indication of how much difference there is in the length of the path between two types of collected waves, such as HH and VV.

Open SanFranc.stk again (File | Open compressed Stokes) and this time select Correlation Coefficient. Go to the section of the Golden Gate bridge that we looked at before (X,Y coordinates 110,55).

Make a box around the area of the bridge we have been looking at. (Click and hold the mouse as you drag out a box.) Now zoom in several times. (Image | Zoom in) You should still see the bridge on the screen as you enlarge the image. Be sure to drag out the image so that it fills the entire screen (Click and drag the box at bottom right of the image.) After you zoom several times, you should see the option Image | Display correlation coefficients appear in bold letters. Click on this option.

Try other options in the Dialog box and compare the results with the key given in SanFrkey.gif. You should see that some polarizations are very useful for separating urban areas from water or vegetation.

Teacher's Guide - Table of Contents

Converted to the IBM-PC by Al Wong, sirced03@southport.jpl.nasa.gov

Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109