Landsats detect EMR
The sensors on Landsat satellites detect visible light, and they also detect longer wavelengths of energy which we can not see. All of this energy is electromagnetic radiation (EMR).
EMR is energy "propagated" through space between electric and magnetic fields. It's a tricky thing; it has both electric and magnetic components-- so neither electric current nor magnetism is electromagnetic. In practical terms, we speak of different kinds of EMR, but really these are just different wavelengths of EMR. One could also say that they are different frequencies-- since EMR's frequency times wavelength equals the speed of light, any given frequency has one wavelength, and vice-versa.
These wavelength regions of the EMR spectrum include gamma, X-ray, ultraviolet, visible, infrared (IR), microwave, radio, and heat. Infrared includes far infrared, intermediate infrared, and near infrared (NIR). The radio region includes Low Frequency, Medium Frequency, High Frequency, Very High Frequency (VHF) and Ultra-High Frequency (UHF). Rotating generators and alternating current (AC) lines give off EMR at very long wavelengths.
The micrometer is a common unit of wavelength. Micrometer = micron = one millionth of a meter = 10-6 m = µm. (µ is a mu, the twelfth letter in the Greek alphabet, transliterated to English as M and m, and in the metric system meaning "millionth".)
MSS and TM bands
The Garden City images are from the Multispectral Scanner (MSS), a sensor carried on Landsats 1-5. MSS has 4 bands, and 79-meter resolution. Many of the newer images in Earthshots are TM data-- from the Thematic Mapper, a newer scanner carried on Landsats 4 and 5-- or ETM+ data, from the Enhanced Thematic Scanner Plus, on Landsat 7. These sensors have finer resolution and more bands than MSS, but the images are all comparable and work on the same principles.
The MSS and TM bands (see table below) were selected to maximize their capabilities for detecting and monitoring different types of Earth resources. For example, MSS band 1 (TM band 2) can detect the visible green reflectance of vegetation, and band 2 of MSS (TM band 3) is designed for detecting chlorophyll absorption in vegetation. MSS bands 3 and 4 (TM band 4) are ideal for near-IR (NIR) reflectance peaks in healthy green vegetation and for detecting water-land boundaries. TM band 1 can penetrate water for bathymetric mapping along coastal areas and is useful for soil-vegetation differentiation and for distinguishing forest types. The two mid-infrared red bands on the TM (bands 5 and 7) are useful for vegetation and soil moisture studies, and discriminating between rock and mineral types. The thermal-infrared band on the TM (band 6) is designed to assist in thermal mapping and for soil moisture and vegetation studies.
Table: Landsat bands
Sensor and # Description Landsat Wavelength (µm) Resolution
------------------------------------------------------------------
MSS band 4/1* green 1-5 0.5 - 0.6 79/82 m**
MSS band 5/2* red 1-5 0.6 - 0.7 79/82 m**
MSS band 6/3* near infrared 1-5 0.7 - 0.8 79/82 m**
MSS band 7/4* near infrared 1-5 0.8 - 1.1 79/82 m**
MSS band 8 thermal IR 3 10.4 - 12.6 237 m
TM band 1 blue 4-5 0.45 - 0.52 30 m
TM band 2 green 4-5 0.52 - 0.60 30 m
TM band 3 red 4-5 0.63 - 0.69 30 m
TM band 4 near infrared 4-5 0.76 - 0.90 30 m
TM band 5 shortwave IR 4-5 1.55 - 1.75 30 m
TM band 6 thermal IR 4-5 10.40 - 12.50 120 m
TM band 7 shortwave IR 4-5 2.08 - 2.35 30 m
ETM+ band 1 blue 7 0.45 - 0.515 30 m
ETM+ band 2 green 7 0.525 - 0.605 30 m
ETM+ band 3 red 7 0.63 - 0.690 30 m
ETM+ band 4 near infrared 7 0.75 - 0.90 30 m
ETM+ band 5 shortwave IR 7 1.55 - 1.75 30 m
ETM+ band 6 thermal IR 7 10.40 - 12.5 60 m
ETM+ band 7 shortwave IR 7 2.09 - 2.35 30 m
ETM+ band 8 panchromatic 7 0.52 - 0.90 15 m
* Landsat 1-3's return-beam vidicons were considered bands 1-3.
** Landsat 3 -> 4: altitude 920 -> 705 km; resolution 79 -> 82 m.
(Landsat Data Users Handbook, 1979/1984; NASA L7 Internet files)
The three basic components of visible light are red, green and blue light (RGB). In theory, any color in the visible spectrum can be made by adding these three (the "primary additive colors") together in some proportion. This is how older computer monitors work (you may have heard of "RGB" monitors).
Colored ink or paint works differently; blue paint absorbs (or "subtracts") all light that hits it except blue light, which reflects back to the eye. The primary subtractive colors are cyan, magenta and yellow. These three are also the secondary additive colors-- i.e. what you get by mixing any two primary additive colors.
The Landsat images in Earthshots are color composites, made by assigning the three primary colors to three bands of a sensor (either MSS bands 4, 2 and 1, or the equivalent TM bands, 4, 3 and 2). The mixture of the three colors represents the mixture of the three bands of EMR reflected by the Earth.
So these images are not color photographs-- green fields won't necessarily look
green in an image. And these assignments are arbitrary; someone could have chosen other bands to be represented
by red green and blue, or they may have picked only one or two bands. An image made of just one band is normally shown as shades of
gray, with bright pixels representing high values for that band and dark areas representing low values. But again, that is only
customary; you need to know what each composite color signifies, and then understand how they mix.
So for an example: when you are looking at a false-color MSS composite where:
red = NIR (MSS 4 / TM 4 / ETM+ 4) green = red (MSS 2 / TM 3 / ETM+ 3) blue = green (MSS 1 / TM 2 / ETM+ 2)then what does a yellow area signify? Since yellow is red and green combined, you know that that ground was reflecting a lot of near-infrared, a lot of red energy, and little green energy. (If all values were high the area would appear white.)
For another example: consider one pixel (one tiny single-colored dot) in a digital image. One MSS pixel usually represents an area on the ground 79 meters square, roughly the size of a soccer field. Again, let us say that RGB = MSS bands 4, 2 and 1. The energy reflected off the Earth to the sensor is expressed as an integer (whole number) between 0 and 255 for each band. (It is 0 through 255 because one 8-bit byte can represent 256 values.) If bands 4, 2 and 1 for this soccer field have the values 189, 35 and 27, then this pixel will be a mixture of a lot of red, a little green and a little blue. It will end up looking mostly red. We might suspect that this soccer field has healthy vegetation, since vegetation reflects infrared energy to stay cool and wet, but absorbs visible light for photosynthesis.
A mnemonic for this scheme is, RGB = NRG (Red Blue Green = Near-infrared Red Green, or "energy").
The standard band combination used in Earthshots makes vegetation appear as shades of red-- brighter reds indicating more vigorously growing vegetation. Soils with no vegetation or sparse vegetation range from white (for sand) to greens or browns, depending on moisture and organic matter content. Water appears blue. Deep, clear water is dark blue to black, while sediment-laden or shallow waters will appear lighter. Urban areas appear blue-gray. Clouds and snow are both bright white-- they are usually distinguishable from each other by the clouds' shadows.
About stretching
Almost all the images in Earthshots have their colors stretched. Compare the stretched and unstretched versions of the 1988 Garden City image to see what this means.
The full brightness scale is 0-255. But originally, the 1988 image's NIR values (near infrared, shown as red) were all between 30 and 80, so the image was quite dark, and had little contrast. So we "stretched" these values to the full range; 30 became 0, 80 became 255, and the rest stretched out in between. Doing this for all 3 bands made the darks a bit darker and the brights much brighter.
This sort of manipulation has disadvantages and advantages; the data have been altered, but the images are more legible. (If you paid very close attention, you may also have noticed that the unstretched image compressed to a smaller JPEG file.)
(See the next help article to learn about what we can do with Landsat data besides displaying them as pictures.)
How to cite this article
Campbell, Robert Wellman, ed. 2008. "Help: How images represent Landsat data." Earthshots: Satellite Images of Environmental Change. U.S. Geological Survey. http://earthshots.usgs.gov. This article was first released 14 February 1997, and last revised 14 August 2008.