InSAR Methods and Applications

1
InSAR Methods and Applications

• Howard Zebker
• Stanford University

2
What is a radar?

• Radar Radio Detection and Ranging
• Measures time of flight of EM pulses

3
Early radars
Breit and Tuve (1925)
Taylor and Young (1922)
4
Distance measurements
5
Mapping multiple objects - ppi
6
Imaging geometry
7
Forming an image
8
(No Transcript)
9
(No Transcript)
10
Radar block diagram
11
Imaging radar block diagram
12
Radar equation
?
?
?
?
13
Antenna gain

• Antenna directs energy in one direction
• Gain defined as ratio of illuminated cone to
• total solid angle

L
H
14
Radar cross section

• Radar equation requires cross section
• Distributed targets normalized cross section
  (s0) multiplied by pulse-limited area
• Range illumination
• Azimuth illumination

15
Signal to noise ratio
Absolute level of power not really relevant What
counts is how much larger signal is than noise
Signal to noise ratio (SNR) Psig / Pnoise
16
The dB table
p
s
p
17
Properties of an EM wave

• Observables are frequency, amplitude, phase, and
  direction
• Phasor notation drops explicit time dependence

18
(No Transcript)
19
Phase of an EM wave
Distance traveled is measured as phase of wave
20
Observed phase of a radar echo
Range r - (two-way travel)
21
InSAR geometry and phase
Antenna 2
Antenna 1
r2
r1
22
Phase calculation
23
InSAR geometry
?r
r?r
r
24
InSAR phase - topography
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
InSAR geometry - deformation
?r
r?r
r
?r
29
InSAR deformation phase

• Phase similar to before but now has displacement
  term
• To infer deformation, get difference and
  compensate for topo term, leaving only
  deformation signal

30
(No Transcript)
31
(No Transcript)
32
Phase noise
33
Decorrelation

• Baseline decorrelation
• Temporal decorrelation
• Rotational decorrelation
• Unspecified noises in system
• Thermal effects
• Quantization
• Other nonlinearities

34
Decorrelation sources
35
Quantifying decorrelation
Measured as
Modeled as s1 c n1 s2 c n2 then ?

36
Signal Model
?
y sin ?
P
y
37
Baseline decorrelation
Calculate cross-correlation s1s2 and
since the correlation is proportional to
38
Baseline decorrelation plot
Seasat satellite - L-band
1
Theoretical
Correlation
Observed
0
0
5000
2500
Baseline, m
39
Rotational, temporal decorrelation
1
L-band
Correlation
C-band
0
0
10
20
0
1
2
RMS scatterer motion, cm
Rotation, deg
40
Temporal decorrelation data - L-band
Death Valley floor
1
Oregon lava flows
Correlation
Oregon forest
0
0
20
10
Time, days
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
(No Transcript)
58
(No Transcript)
59
(No Transcript)
60
(No Transcript)
61
(No Transcript)
62
(No Transcript)
63
(No Transcript)
64
(No Transcript)
65
Mass balance
66
(No Transcript)
67
(No Transcript)
68
(No Transcript)
69
(No Transcript)
70
(No Transcript)
71
Volume scattering models
Grain boundary migration
72
Accumulation in layered media
r
s
Layered model
73
Accumulation rate
Inferred layer spacing
74
(No Transcript)
75
(No Transcript)
76
(No Transcript)
77
(No Transcript)
78
(No Transcript)
79
(No Transcript)
80
(No Transcript)
81
(No Transcript)
82
InSAR limited by decorrelation

• Areal coverage limited to areas with high
  correlation, often fails in vegetated regions

Correlation (orange high)
Long Valley caldera
Phase
83
Persistent scatterer principle
Distributed scatterer
Single point scatterer
Dominant scatterer
Pixel phase
2?
Acquisition
84
Amplitude dispersion proxy for phase noise

• Speckle observed as image amplitude variation

? ?
85
Early PS worked well in urban areas
San Francisco Bay Area
From Ferretti et al., EOS, 2004
86
Problems using amplitude dispersion
? ?
PS?
87
Observed phase is sum of many components
PS pixels have low noise
Correlated with baseline
Uncorrelated in time

• ?insar????def ???topo ???atm ???orbit ???noise

Correlated locally- spatially
Correlated globally- spatially
88
Topo error term correlated with baseline
Topo error proportional to red line slope
Estimate topo error and subtract from each
observation
89
Filter to extract spatially correlated signal
Filter calculates variance within circle
90
Pixels that pass filter are retained with phases
91
Long Valley - interferograms and PS
Interferograms
Sept. 1992
Aug. 2000
PS networks
92
Interferogram vs. PS
Good quality data
Poor quality data
Interferogram
PS network
93
Long Valley unwrapped
94
(No Transcript)
95
(No Transcript)
96
(No Transcript)