Remote Sensing

1
Remote Sensing

• January 20, 2006
• Geog 258 Maps and GIS

2
Outlines

• Principles of remote sensing
• Three key aspects of resolution of remotely
  sensed image
• Aerial photo vs. satellite image
• Classifying aerial photograph
• Geometric distortion of aerial photograph
• Classifying satellite image

3
Principles of remote sensing

• What you see is the amount of electromagnetic
  energy reflected by object needs energy source,
  object, sensor
• Electromagnetic energy can be divided into
  different spectral bands (visible light, NIR,
  microwave) given its wavelength
• Electromagnetic spectrum
• Electromagnetic energy interacts with object
  differently
• Snow reflects most of energy (thats why it looks
  bright)
• Dry soil absorbs most of energy (thats why it
  looks dark)
• How object interacts with electromagnetic energy
  is a function of wavelengths ? basis of image
  interpretation
• Spectral reflectance signature
• Electromagnetic radiation interacts with
  atmosphere as well. Some radiations (in visible,
  NIR and microwave bands) pass through atmosphere
  relatively well
• Atmospheric window

4
Electromagnetic spectrum
5
Spectral signature
Explain why water looks darkish blue Explain why
vegetation looks greenish Explain why sand looks
reddish yellow
6
Atmospheric window

• It shows how much electromagnetic radiation
  transmits atmosphere
• In which bands, is transmission high?
• Advantage of using those bands?
• Compare aerial photo and satellite image in terms
  of spectral bands

7
Aerial photograph
Optical sensor
Infrared sensor
Radar sensor
8
Different kinds of image

• Pancromatic image
• True-color image
• False-color image

9
Pancromatic image

• If airborne cameras use black/white film or
  satellite sensors use a single band, it produces
  pancromatic image (gray scale image)

10
Color composite

• Color primaries RGB (Red, Green, Blue)
• Many colors are formed by combining color
  primaries in various proportions
• Look inside your inkjet color printer. Did you
  really buy 100 different color cartridges?
• Same principles apply to producing color images
  taken from airborne cameras or satellite sensors

11
True-color image

• See Table 9.1 (p. 182)
• How many bands are used for Landsat 4 and 5?
• Whats the spectral range (and corresponding
  colors) for band name 3, 2, 1?
• Each band collects different images (See Figure
  9.27 at p. 182)
• To produce color-image, bands 321 are assigned to
  RGB respectively
• The color of resulting image resembles what would
  be observed by human eyes
• Thats why its called true-color image

12
False-color image

• If three bands are arbitrarily (i.e. doesnt have
  to be RGB bands) assigned to RGB layers, it
  produces false-color image

SPOT XS1 (green band)
SPOT XS2 (red band)
SPOT XS3 (Near IR band)
R XS3 (NIR band)G XS2 (red band)B XS1
(green band)
Why does vegetation appear dark in green band in
pancromatic image?
Why does vegetation appear red in this color
composite?
Why does vegetation appear bright in NIR band in
pancromatic image?
13
Three key aspects of resolutions

• Spatial pixel (grain) size minimum distance
  which can be recorded
• Aerial photo has of course higher resolution
• Temporal how often is the image recorded
• For satellite image, it can be regular
  (satellites are orbiting the earth in regular
  time interval)
• Spectral how many bands are used
• Satellite sensors use different bands most of
  them less than 10 (e.g. TM uses 7 bands) sensors
  with band gt 10 are called hyperspectral (e.g.
  MODIS)
• Important in determining the fitness of use

14
Types of sensor systems

• Passive
• Rely on external energy source
• Optical sensor (reliant on solar energy)
• Active
• Sensor itself sends out radiation (like flash
  camera)
• Radar sensor (reliant on microwave)
• Which one can record the image of earth at night,
  and which one cant?
• Another reason for the popularity of Radar
  imaging is the ability of microwave to penetrate
  atmosphere (e.g. clouds) see Figure 9.4 (p. 164)
• So what would be the advantage of Radar imaging
  compared to optical imaging?

15
Aerial photography vs. Satellite
imaging(Airborne RS vs. Spaceborne RS)

• Which began earlier? (with the development of
  what?)
• Which would have a higher spatial resolution? (in
  which altitude?)
• Are they analog or digital? (output format)
• Which is more flexible? (can satellite be
  launched any time?)
• Which would provide geographic coverage in a
  systematic manner?
• Which is equipped with wider spectral bands?
  (does air photo has thermal-infrared sensors?)
• How would post-processing be different?

16
Classifying aerial photo

• Film emulsion
• black-and-white, true-color, color-infrared
• Cameras height
• high, mid, low-altitude
• Cameras vantage point
• vertical, oblique
• When color-infrared films were invented, and for
  which purpose?

17
Geometric distortion of air photo

• Aerial photo gives us perspective view (it
  distorts geometry of geographic features)
• Transformation from central to parallel
  perspective results in planimetrically correct
  photo or orthophoto

18
Sources of geometric distortion

• Camera tilts

Scale is not uniform
19
Sources of geometric distortion

• Variations in terrain

Remember photo is flat (2D) and objects should be
seen from the point of infinite height in the
planimetrically correct photo. In the aerial
photo, a is displaced outward to a b is
displaced inward to b due to terrain variations
20
Classifying satellite image

• Spatial resolution
• high, low
• Types of energy source
• passive, active
• Types of spectral bands
• pancromatic, multispectral, hyperspectral
• Special purposes
• earth-observing, meterological,
  thermal-infrared
• From the list of earth observing satellite,
  http//www.ersc.wisc.edu/resources/EOSC.php,
  place the sensors in classification schemes above

21
Review questions

• What is the difference between aerial photo,
  orthophoto, and DOQs?
• Discuss advantage and disadvantage of aerial
  photo vs. satellite imagery
• What is the advantage of using active sensor
  systems?
• What is the advantage of using spectral bands
  beyond visible wavelength ranges (0.4-0.7 micron)?