1 ECE/OPTI 531 – Image Processing Lab for Remote Sensing Fall 2005 The Nature of Remote Sensing Reading: Chapter 1 Fall 2005 The Nature of Remote Sensing 2 The Nature of Remote Sensing • Introduction • Remote Sensing Systems • Remote Sensing Physics • Sensor Parameters • Display and Data Systems
29
Embed
The Nature of Remote Sensing - University of Arizonadial/ece531/Nature_of...agriculture near Yuma, Arizona TM1 TM2 TM3 TM4 TM5 TM6 (TIR) TM7 The Nature of Remote Sensing 10 Fall 2005
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
ECE/OPTI 531 – Image Processing Lab for Remote Sensing Fall 2005
The Nature of RemoteSensing
Reading: Chapter 1
Fall 2005The Nature of Remote Sensing 2
The Nature of Remote Sensing
• Introduction• Remote Sensing Systems• Remote Sensing Physics• Sensor Parameters• Display and Data Systems
2
Fall 2005The Nature of Remote Sensing 3
Definitions
• Remote Sensing =“Measurement at aDistance”
• This course is aboutEarth remote sensing– Airborne or satellite
platforms– Optical region of the
spectrum• visible (400–700nm) to
thermal (long-wave)infrared wavelengths(8 to 12mm)
Fall 2005The Nature of Remote Sensing 4
Definitions (cont.)• Remote sensing requires
– Active or passive source– Target– Medium (typically lossy)– Sensor (optics, detector)
• Source radiation modeledas a traveling wave– Time-harmonic– c = λν
• C = 2.998 x 108 m/s• λ is the wavelength• ν is the frequency
– Also, wavenumber 1/λ cm-1
• EM spectrum is infiniteand continuous
• Energy interacts with matter– Reflection (Scattering)– Transmission– Absorption (Re-emmitted)
• Sensor characteristics– Spatial (Ground Sample
Interval)– Spectral (Range and width)– Temporal (Revisit time)– Radiometric (Precision)
3
Fall 2005The Nature of Remote Sensing 5
Applications
• Environmental assessment and monitoring• Global change detection• Agriculture• Nonrenewable resources• Renewable resources• Meteorology• Mapping• Military surveillance and reconnaissance• News media• Further reading: Remote sensing tutorial
http://rst.gsfc.nasa.gov
Fall 2005The Nature of Remote Sensing 6
Types of Sensors and Sensing
• Multiangle Imaging SpectroRadiometer (MISR)sensor on NASA Terra satellite (http://www-misr.jpl.nasa.gov/)
4
Fall 2005The Nature of Remote Sensing 7
• Single, broad spectral band,typically 400nm wide in thevisible spectrum
• Often called “panchromatic”• Large number of photons
collected, which allows smallerdetectors, i.e. greater spatialresolution
• Corona was the first globalsatellite reconnaissance mission– high resolution camera– photographic film returned to
Ronald Reagan Washington National Airport (courtesy Space Imaging Inc.)
Fall 2005The Nature of Remote Sensing 24
The Nature of Remote Sensing
• Introduction• Remote Sensing Systems• Remote Sensing Physics• Sensor Parameters• Display and Data Systems
13
Fall 2005The Nature of Remote Sensing 25
Spectral Regions• Determined by:
– "windows" where atmospheric transmittance is relatively high– wavelength regions where detector sensitivity is relatively
high
Temperature(passive)roughness (active)
thermal (passive)artificial (active)
1mm-1mMicrowave, Radar
temperaturethermal8–9.510–14
Thermal Infrared(TIR)
reflectance,temperature
solar, thermal3–44.5–5
Mid-WaveInfrared (MWIR)
reflectancesolar1.1–1.351.4–1.82–2.5
Short-WaveInfrared (SWIR)
reflectancesolar0.7–1.1Near Infrared(NIR)
reflectancesolar0.4–0.7Visible (V)
Surface Propertiesof Interest
Radiation SourceWavelengthRange (µm)
Name
Fall 2005The Nature of Remote Sensing 26
0
0.2
0.4
0.6
0.8
1
2.5 4.5 6.5 8.5 10.5 12.5 14.5
tran
smit
tance
wavelength (µm)
H2O
O3
CO2,
H2O
CO2
CO2
MWIR TIR (LWIR)VNIR SWIR
0
0.2
0.4
0.6
0.8
1
0.4 0.8 1.2 1.6 2 2.4
tran
smit
tance
wavelength (µm)
CO2
H2O
H2O
H2O
H2O
CO2
CO2
CO2,
H2O
CO2,
Atmospheric Transmittance
• Atmospheric “windows” result from energyabsorption by air molecules– Water vapor (H20)– Carbon dioxide (C02)– Ozone (O3)– Others to a lesser extent
14
Fall 2005The Nature of Remote Sensing 27
Radiation Sources
• Approximately equal at the top-of-the-atmosphere (TOA) in the Mid-Wave IR (MWIR)
MWIR
1
10
100
1000
104
1
10
100
1000
104
0.1 1 10
solar irradiance
earth emission
irra
dia
nce
(W
-m-2
-µm
-1)
radian
t exitan
ce (W-m
-2- µm
-1)
wavelength (µm)
Fall 2005The Nature of Remote Sensing 28
Human Vision
• Sensitive over very small range of total solarspectrum
0
500
1000
1500
2000
2500
0
0.2
0.4
0.6
0.8
1
400 900 1400 1900 2400
solar irradiance
daylight sensitivity
sola
r ir
radia
nce
(W
-m-2
-µm
-1)
relative sen
sitivity
wavelength (nm)
solar spectrum and human visual sensitivity
15
Fall 2005The Nature of Remote Sensing 29
Spectral Signatures
• Vegetation spectral reflectance has severaldistinguishing features– “red edge” at 720 – 780nm caused by cellular structure– Low reflectance in the blue and red caused by
chlorophyll absorption; slightly higher reflectance inthe green
– Water absorption features at 1400nm and 1900nm
0
0.1
0.2
0.3
0.4
0.5
0.6
400 800 1200 1600 2000 2400
WheatSugar BeetsOats
refle
ctanc
e
wavelength (nm)
0
0.1
0.2
0.3
0.4
0.5
400 800 1200 1600 2000 2400
Kentucky Blue GrassRed Fescue GrassPerennial Rye Grass
refle
ctanc
e
wavelength (nm)
“red edge”
Fall 2005The Nature of Remote Sensing 30
Spectral Signatures (cont.)
• Soil and geologic minerals show relativelysmooth spectral reflectance– water absorption features in soils at 1400nm and
1900nm– narrow molecular absorption features caused by
• Apply each componentof the spatial responseand downsample to30m
23
Fall 2005The Nature of Remote Sensing 45
Imaging Simulation (cont.)
• Compare to real TM of same area, acquired 4months later
simulated TM real TM
(contrast-adjusted)
real TM
Fall 2005The Nature of Remote Sensing 46
Spatial Resolution
• A “subpixel” object smaller than the GIFOV can bedetected, but not resolved
• Detectability of a subpixel object depends on:– object size relative to the sensor GIFOV– object radiance contrast to the surrounding background– scene noise (“clutter”)– sensor noise
Berkeley Pier: 7m wide, concrete and wood
– detection capabilitydepends on spectral band
Detection of a 7mobject in TM imagery
24
Fall 2005The Nature of Remote Sensing 47
Detectability
• Low-contrast subpixeltargets must bebigger than high-contrast targets fordetection
50% target coverage of GIFOV
minimum-area detectable target
128 15
1 11
111
DN DN
! = 0.08! = 1
! = 0 ! = 0.04
Dependence of detectability on object size and contrast
Fall 2005The Nature of Remote Sensing 48
Sampling
• The measuredradiance of a sub-pixel object dependson the location of theobject relative to thepixel samples
15
20
25
30
35
40
45
6 8 10 12 14 16 18 20
DN
pixel
cba
a b c
Three scans across the Berkeley Pier
25
Fall 2005The Nature of Remote Sensing 49
0
0.2
0.4
0.6
0.8
1
1200 1400 1600 1800 2000 2200 2400
Rel
ativ
e R
esponse
wavelength (nm)
TM5 TM7
0
0.2
0.4
0.6
0.8
1
400 500 600 700 800 900 1000 1100
Rel
ativ
e R
esponse
wavelength (nm)
TM1 TM2 TM3 TM4
Spectral Response
• Individual band spectral response determined by– detector responsivity– filter transmission (discrete spectral band sensors)– spectrometer slit width (hyperspectral sensors)
Example spectral response curves – Landsat Thematic Mapper (TM)
Fall 2005The Nature of Remote Sensing 50
Spectral Resolution
• As in the spatial case, the width of theinstrument spectral response determines itsability to record detail in the spectral signal
0
0.003
0.006
0.009
0.012
0.4 0.5 0.6 0.7 0.8 0.9 1wavelength (µm)
radia
nce
(W
-m-2
-sr-1
-µm
-1)
TM1 TM2TM3
TM4
Simulation of TM band measurements of a vegetation spectral signal