Earth from Space

Teacher's Guide for
Two Remote Sensing Lessons:
Studying the Earth From Space

These lessons are designed for middle or high school instruction. They feature ocean color data, but can be easily adapted to any Earth data type.

These lessons have been presented at three sessions of the American Meteorological Society's annual Maury Project Summer Teacher Workshop, held at the U.S. Naval Academy in Annapolis Maryland for the past five years. Maury participant comments on these lessons have been very favorable. Following this year's presentation and the comments I received , I believe these lessons are now classroom-ready, so I have decided to make them available on the internet.

Rebecca Farr
August, 1998

Teacher's Guide:

Note: Basic concepts about the electromagnetic spectrum should be understood before teaching these lessons. A simple description of the electromagnetic spectrum can be found in "Space Based Astronomy: A Teacher's Guide with Activities" (EG-102, NASA Headquarters, Education Division, Code FET, Washington DC 20546-0001, August 1994), Unit 2: The Electromagnetic Spectrum (Introduction). The activities in this Lesson were adapted from Unit 4 in this publication.

Lesson 1 A.

Stray Light:

It is very important to use good color filters for this exercise. You may have trouble if you do not use gel filters. Ordinary colored cellophanes and plastics may pass other colors. This is called "stray light". For instance, green objects may actually look green when using the blue filter. This means the blue filter is not blocking all of the yellow light.

Stray light at other color frequencies allows the students see the actual colors of the objects making the exercise much easier and ruining the desired effect. Good filters will pass only the colors desired while minimizing stray light. Inexpensive professional color filters are available from vendors like Edmund Scientific. This chart illustrates the frequencies of light passed by KODAK gel filters:

Kodak Wratten filter spectral chart

Try out your filters ahead of time to make sure they work well with the objects you have chosen. You will find the red filter is most effective at blocking light from other parts of the color spectrum. It is best to start this activity with the red filter first.

For the first part of this exercise, it's fun to use toys that have a certain color associated with them eg: a red fire hat, a yellow school bus, green Kermit the frog etc....I like to raid my son's toybox for large, colorful objects.

For the second part, use geometric shapes, boxes or other uncharacterizable objects for the unknowns eg: a blue swimsuit, a red ball, a green glove, a yellow box etc. Have fun; bring in some unusually colored objects, like a blue banana, a yellow apple...Any fluorescent colored object will throw the class off. These colors reflect light in quite a different way from 'standard' colors like blue, red or green.

Results:

Here is an example of some results drawn as histograms:

Histogram examples

Discussion:

Remote Sensing Realities:

You will find it practically impossible to find two objects from different sources with exactly the same color. The class will even notice a difference in the histograms of objects that appear to be nearly the same color in white light, illustrating the power of this analytical spectral technique for color recognition.

If slight color differences can be seen using this technique with toys and the naked eye, imagine how discerning satellites are!! Satellite digital spectral analysis is sensitive enough to allow analysts to discern the different chemical make-up, and thus origins, of paint on vehicles on the ground miles below.

Students may complain about glare on an object, lack of reflectivity from a furry toy, or lack of discernability because the object is too small. These complaints illustrate three important limiting factors in characterizing objects using satellite remote sensing:

Because of the small sampling size of this exercise, it will not be totally clear to everyone that a 'red' unknown object can be correctly deduced from the histogram of a known red object. Three important remote sensing ideas can be presented here:

Someone may ask 'Why not just use a color camera rather than a whole collection of sensors sensitive to different bands of color?' The early Landsat satellites did indeed have both a black and white camera and a digital sensor called a Multi-Spectral Scanner (MSS). At that time, the camera was actually more dependable than the electronic sensor. Over the years, researchers and engineers have perfected digital sensors and continue to refine them. Today's Earth satellites use digital sensors almost exclusively.

Digital sensors have two main advantages over film or televison cameras:

1. Cameras produce analog images, not digital data. This means you cannot perform mathematical calculations or manipulations on analog images, making it impossible to derive quantities such as chlorophyll concentration from the image, even though cameras can be equipped with color filters to collect data in several color bands.

2. While the resolution of a camera is limited by the chemistry of its film emulsion, the resolution of electronic sensors gets better every year with advances in technology.

Someone may ask if three or more discreet spectral sensors can really see everything. The answer is "No, but it's the best we can do for now.". The ideal sensor would cover a broad and continuous spectral range. Until recently, this has not been possible because of limitations in sensor and spacecraft technologies. However, new designs for future satellite sensor systems employ this idea. These new satellites are called "hyperspectral" satellites.

It is somewhat likely you will have at least one student in the class who cannot distringuish the difference between red and green. This is a genetic trait known as red-green color blindness, and it occurs in approximatey 8/100 of males. His observations will be different from those of the other students. Discuss how they are different. The class should soon realize that he is not perceiving the red and green in the same way.

Lesson 1 B.

Ocean Color Image Chlorophyll Value Answer Key (mg/m3)

.65
.42
.47
.47
.42
.55
.42
.32
.42
.55
.47
.65
.47
.37
.32
.42
.37
.32
.37
.32
.47
.42
.42
.42
.32
.42
.37
.37
.32
.27
.42
.42
.42
.42
.42
.32
.37
.42
.27
.27
.47
.42
.37
.32
.32
.32
.32
.27
.27
.27
.37
.32
.42
.32
.32
.27
.27
.27
.27
.27
.42
.42
.32
.32
.27
.27
.27
.22
.22
.22
.32
.42
.32
.27
.32
.22
.22
.27
.27
.32
.37
.37
.32
.22
.22
.22
.32
.27
.27
.32
.32
.37
.32
.27
.22
.22
.22
.27
.22
.22

100 pixel CZCS scene

The resulting plot should look something like this. Each square pixel covers approximately 16 square kilometers, (an area about the size of Washington D.C.). The whole image of 100 pixels covers about 1600 square kilometers (576 square miles). The next figure illustrates the scale of the problem. The 10x10 square completed in the first part of the exercise is actually only one small part of a much larger and very interesting image of the Gulf Stream:

CZCS Gulf Stream Image with subset

Discussion:

The Gulf Stream image above has four kilometer resolution. In other words, each pixel is four kilometers on a side. This means you cannot resolve the shape of any feature smaller than four kilometers. Thus, a four kilometer resolution satellite image cannot show many small scale features that exist in the ocean. Most satellite images have four kilometer or lower resolutions. One kilometer resolution is considered 'good' resolution, but is more expensive to collect and process. The image resolution of satellite data is increasing as sensor technology and data processing techniques evolve.

Someone in the class may notice that there are actually only eight different chlorophyll values in this exercise. Real data would be a lot more variable...give that student a gold star! You should expect each pixel to reveal a different chlorophyll value in real data. (For this exercise, the author chose to use only eight values in order to reduce student effort.)

You will probably find that the students spend a lot of time agonizing over the accuracy of their calculations. However, even if they make errors, by doing this exercise they will understand the following important principles in remote sensing:

Some additional information on color blindness...

(adapted from an article by Joe Pfeiffer, pfeiffer@nmsu.edu)

Color blindness is a general term for a deficiency in the visual system that leads to a distorted color perception. Speaking very loosely, we have three types of color receptors in our eyes, sensitive to red, green and blue. We also have black and white receptors (these last are more sensitive than the color receptors, which is why we have deficient color vision at night). Color blindness comes as a result of a lack of one or more of the types of color receptors. There is a normal variance in color reception; nobody's color vision is "perfect" -- whatever that might even mean!

Trying to describe how the color sense of a color blind person differs from that of a color sighted person is frustrating and fruitless (a conclusion based on nearly four decades of trying to understand how the world looks to my color blind father!). The best I can do is to try to describe effects.

The most common form of color blindness is red-green color blindness, which affects approximately 8% of the male population and some very small percentage of the female. This form of color blindness is a result of a lack of red receptors (what? not green? just trust me -- if you ask, I'll start talking about opponent color theories). If a red-green color blind person sees a red object and a green object which are about equally bright, the red object will seem much darker than the green one. This person will find the display of an old LED calculator almost unreadable A good approximation to the world of a color blind person can be obtained by unplugging the red cable from a graphics display.

There can also be other forms of color blindness -- yellow-blue is the second most common form, but it's very, very rare. And it is possible to have the color receptors missing entirely, which would indeed result in somebody having monochrome vision."


Privacy Policy and Important Statements

Author and Responsible NOAA official:
Rebecca Farr
Ground Data Operations Manager
NOAA NESDIS
Posted August, 20 1998