Test 2 Flashcards
(136 cards)
1
Q
what does microwave have the ability to do?
A
measure the brightness temperature of the earth’s surface
2
Q
What are the wavelengths of microwave?
A
1 cm to 1 m
3
Q
Passive microwave has…
A
low energy and coarse spatial/ spectral resolution
4
Q
Types of active microwave sensors
A
scatterometer and altimeter- Non-imaging Radar
Real Aperture Radar and synthetic aperture radar- Imaging Radar
5
Q
What does a scatterometer do?
A
Coarse resolution, good for ocean surface, wind, and wave
6
Q
What does an altimeter do?
A
Measure the height of the surface of the earth
7
Q
What are the characteristics of Radar principles?
A
Radar Detection and Ranging
Microwave range
active system
high spatial resolution (same as optical)
day-night & all-weather capabilities
complementary to optical systems
8
Q
How does RADAR work?
A
Radar sends pulses through the transmitter and the target reflects the echo which captured by the receiver
9
Q
What does radar distinguish objects by?
A
They distance but when it looks Nadir, the same distance is on the right and left, this causes problems, leads to side looking geometry
10
Q
How do you improve azimuth resolution?
A
a shorter wavelength pulse will result in improved azimuth resolution
The size of the antenna is inversely proportional to the size of the angular beam width (azimuth resolution)
11
Q
synthetic aperture radar (SAR)
A
makes a relatively small antenna work like it is much larger, by taking advantage of the platform’s motion
12
Q
What does radar measure
A
the ratio between the received electrical field over the field incident to that location on earth
13
Q
What is the backscatter coefficient tell us?
A
amplitude information
14
Q
Some ranges for backscattering
A
very high (>5db): man made objects, terrain slopes towards the radar, very rough surface, radar looks very steep
high backscatter (-10dB to 0dB): rough surface with dense vegetation
moderate backscatter (-20 to -10 dB): medium level of vegetation, agricultural crops, moderately rough surfaces
low backscatter (below -20 dB): smooth surface, water, road, or very dry terrain
15
Q
why do we get speckle?
A
inherent to imaging of distributed scatterers because SAR is a coherent (same wavelength imaging sensor)
16
Q
What is speckle?
A
Salt and peppa. areas with similar land or water cover can have a very “salt and pepper” appearance on radar imagery because of constructive and destructive interference between reflected microwave EM waves
17
Q
geometric distortion or elevation displacement
A
elevation displacement, the image displacement in a remote sensing image toward the nadir point in radar imagery due to sensor/target imaging geometries
main types: foreshortening, layover, shadow
18
Q
Foreshortening
A
mountains are leaning towards the satellite sensor
19
Q
Layoever
A
the top of the mountain covers the glacial valley because of layover
extreme case of foreshortening
easy to fix
20
Q
shadow
A
see paper
21
Q
What does the backscatter coefficient provides information about the imaged surface and is a function of…?
A
atmospheric parameters
22
Q
Radar observation parameters
A
wavelength, polarization, incidence angle of the electromagnetic wave emitted
23
Q
surface parameters:
A
roughness, geometric shape/structure, dielectric properties of the target (moisture content)
24
Q
Polarization of electromagnetic wave
A
the direction of the electric field in relation to the wave propagation
25
What is a dielectric constants
is a measure of the electrical conductivity of material
26
What is the determining factor for backscattered radar energy in terrestrial materials?
moisture content
wet materials- high reflection
dry materials- low reflection
27
The radar return can come from three sources
1. direct scattering from the vegetation
2. direct scattering from the ground
3. Multiple scattering between the ground and the canopy (canopy might absorb some of the microwave energy, must account for attenuation by the canopy)
28
Surface factors that influence radar scattering from vegetated surfaces
1. Changes in soil moisture
2. changes in canopy moisture
3. difference in canopy structures/biomass
4. presence or absence of water on top of soil
29
Monitoring soil moisture variation
1. low soil moisture
a. low direct scattering from canopy
b. low multi-path scattering
c. moderate scattering from soil
2. high soil moisture
a. low direct scattering from canopy
b. low multi-path scattering
c. high scattering from soil
3. flooding or inundation of the soil surface with water
a. low direct scattering from
b. low multi-path scattering
c. no scattering from soil
30
Rules of SAR image interpretation
1. regions of calm water and other smooth surfaces will appear black, because the incident radar reflections away from the spacecraft
2. surface variations near the size of the radar's wavelength cause strong backscattering
3. wind-roughened water can backscatter brightly when the resulting waves are close in size to the radar's wavelength
4. a rough surface backscatters more brightly when it is wet
31
SAR Image interpretation rules continued
1. a particularly strong response, say from a corner reflector, can look like a bright cross in a processed SAR image
2. due to reflectivity and angular structure of buildings, bridges, and other human-made objects, these targets tend to behave as corner reflectors and show up as bright spots in a SAR image
3. Hills and other large-scale surface variations tend to appear bright on the side that faces the sensor and dim on the side that faces away from the sensor
32
What are the dielectric constants of water, soil, and vegitation?
80, 3-6, and 1-3 respectively
33
What do you have to sacrifice in order to get rid of speckle?
spatial resolution
34
What is backscattering dependent on?
relative height or roughness of the surface
very rough surface strong backscatter
35
P band
smoother
doesn't penetrate much through vegetation
Longer wavelength is penetrating the vegetation due
36
C Band
Can't see the roads
37
Passive Systems
use natural energy sources, reflected or emitted energy, photography, thermal, passive MW
38
Active systems
Have own energy source, radar lidar, works in the dark
39
What is LIDAR?
Light Detection and Ranging
emits pulses of light towards a surface
works in the visible and NIR
pulses reflect off surface objects and return to the sensor to be recorded
40
What are the steps to LIDAR?
pulses contain individual quanta of light (photons)
recorded photons are converted to electric currents
electric currents are converted to digital counts
digital counts are measured over fixed time intervals
time intervals are converted to distances
41
What does LIDAR measure?
total time for a light pulse to leave the sensor, hit an object, and then return
42
Pulses
the laser pulse is emitted from the lidar system
then reflects off of every surface/ feature in its path
43
Properties of Lidar Pulses
pulse wavelength: NIR
rate: number of pulses per second in kHz
pulse spacing: distance between lidar pulses
pulse footprint: area converted by a single pulse
discrete return vs. full waveform: how a pulse is processed and stored
44
Pulse spacing
the number of lidar pulses emitted per unit area
depends on: flight speed, scanning pattern, and pulse rate
45
Pulse Footprint
ground area covered by a single pulse
the shorter the distance between the lidar and the surface, then the smaller the footprint
46
Clarification between spacing and footprint
footprint: ground covered by a single pulse
Spacing: distance between footprints
both important to resolution and precision
47
relationship between footprint and spacing
spacing= footprint: ideal
spacing footprint: data gaps
48
discrete return vs. full waveform
each lidar pulse contains many, many photons
these photons produce a waveform based on returned energy amplification
every feature the pulse reflects off of will produce a peak in the waveform
the pulse is digitalized to create individual discrete returns or points
some systems store the full waveform
49
Properties of discrete lidar data
spatial coordinates: represents the spatial location of the point
return number: each pulse can have multiple point "returns"
classification code: similar to classifying features or land use in a digital image
intensity value: relative return energy strength
50
Lidar Data Return number
generated from the waveform discretization
each pulse is a collection of many individual
photons of light. results in multiple data
points that can be measured by each lidar
pulse
Multiple points --> multiple returns
the first point is closest object to lidar system
last return: farthest object from the lidar system
51
Classification of lidar
one can classify lidar points based on the feature
52
Lidar data intensity
generated from the waveform energy amplitudes
a relative intensity measure is provided for each point
amplitude of the pulse energy in the reflected waveform
53
Limitations of lidar intensity values
similar but not the same as passive NIR reflectance
dependent on many variables:
reflectance of object
lidar system and wavelength it uses
distance and scan angle between sensor and
target
rank of the return
normalized 8-bit
54
What happens to Lidar NIR pulses in water
they are absorbed. This causes gaps in dataset, a major limitation of lidar scanners. multiple tools necessary for surveying
55
Why is NIR popular for lidar?
availability of stable materials for laser
most surfaced reflect NIR quite well
low signal-to-noise ratio in sunlight
more eye-safe than other wavelengths
56
Digital Elevation Models
digital terrain models: ground itself
digital surface models: features on the ground
canopy height models: height above ground
57
DTM's
a digital model that only contains the elevation of ground points
last returns and single returns
if ground is classified, then the set of ground returns
58
DSMs
surface model that contains the elevation of ground points as well as any fixed features
includes features such as vegetation, buildings, etc
set of first returns and single returns
set of both ground and non-ground
59
"Normalized" DSMs
calculated as nDSM-DSM-DTM
"normalized" to represent height above ground
ex: canopy height models and building height models
60
Drone Laser Scanning
potential for "best of both worlds"
great coverage and resolution (0.05m)
61
What are the effects of leaf chlorophyll on leaf reflectance?
more chlorophyll content is more absorption (less reflectance) in the visible range
62
What are the effects of leaf water content on leaf reflectance?
more water content is more absorption in the SWIR range
63
what does a spectral transformation of two or more bands allow?
reliable spatial and temporal inter-comparisons or terrestrial photosynthetic activity and canopy structural variations
64
What does healthy vegetation do?
absorbs most of the visible light that hits it, and reflects a large portion of the NIR
65
what does unhealthy do?
reflects more visible and less NIR
66
What do visible, NIR, and SWIR bands do?
give actual images of Earth's surface
67
What does the cirrus band do?
specific cloud detection
68
Mulit-spectral
10s of bands
69
hyper-spectral
100s of bands
70
Vegetation indicies
a spectral transformation of two or more bands in to a single variable
71
Does healthy vegetation have a high or low NDVI?
High
72
What is temporal domain used for?
change detection, temporal composite, and phenology
73
What can temporal domain be used for?
Urban Growth, agriculture expansion, forestry, water, floods, volcanic activity, drought, illegal mining
74
NDVI anomaly can be an indicator of ?
drought
75
What are neighborhood operations?
filtering, segmentation, classification, and object detection
76
Mean value smoothing
pixel values are changed to the average of neighboring pixels
Doesn't preserve edges
77
Median filtering
pixel value is changed to the median of neighboring pixels
78
Sobel filers
something with a gradient, used for boundaries, check lab
79
The larger the swath...
the higher the temporal resolution
80
... but by increasing swath
spatial resolution is coarse
81
Classification
mapping from measurements acquired by a remote sensing instrument to a label(s) for each pixel that identifies it with what's on the ground
82
Quantitative Analysis
VHR image classification by object
83
what the the problems with image classification?
Regression and classification ?
84
Regression
predict continuous variables
85
Classification
predict categorical variables
86
What are the types of classification?
supervised and unsupervised
87
supervised classification
target outputs are provided with input data
87
Unsupervised classification
discover patterns in input data for the purpose of clustering, density estimation, dimensionally reduction, and visualization
88
clustering
to discover groups of similar examples within the data
89
density estimation
determine the distribution of data within the input space
89
Visualization
to project the data from a high-dimensional space down to two or three dimensions
90
supervised classification: the steps
1. class nomenclature definition (general land cover classes: impervious, cropland, etc)
2. features and algorithm selection
3. generation of training data (in situ ground measurements, drones, and photo-interpretation)
4. accuracy assessment
91
what does image classification require?
delineating boundaries of classes in n-dimensional space using class statistics
92
pixel based
each pixel is grouped in a class
multiple changes in land use within a short period of time
best for complete data coverage and needs methods to ensure time series consistency at the pixel level
93
Object-based
pixels with common spectral characteristics are first grouped together
94
Two types of cluster based classification (it is unsupervised)
produce clusters (k-means and ISODATA) and label clusters
95
K-means
assign centroids, classify using nearest centroid, recalculate mean and mean change
96
supervised classification approaches
nearest neighbor, maximum likelihood, decision tree/ random forest, neural network, support vector machine
97
Decision tree
using the concept of information entropy
splitting data based on the normalized information gain
98
supervised classification: generation of training data
number of features: 10s-100s so classes are separable
size: if using n features, >10n pixels of training data should be collected for each class
should be large enough to provide accurate estimates of the properties of each class
location: each class should be represented by several training areas positioned throughout the image
uniformity: each training area should exhibit unimodal frequency distribution for each feature
99
What is a key strength of RS
is enables spatially exhaustive wall-to-wall coverage of an area of interest
100
sampling design
protocol for selecting the subset of spatial units that will form the basis of the accuracy assessment
101
Response design
encompasses all aspects of the protocol that lead to determining whether the map and reference classification are in agreement
101
analysis
protocols for defining how to quantify accuracy along with the formulas and inference framework for estimating accuracy and area and quantifying uncertainty of these estimates
102
what are the properties of optical?
source is the sun
measures reflectance
includes vegetation indices and land cover mapping
doesn't see through the clouds must use cloud masking
cloud top reflectance
103
what are the properties of thermal IR
emitted from the sun
measures temperature
used for fire detection
cloud top height measurements
104
What are the properties of SAR
sensors in mw
backscatter is measured (amplitude and face of the wave)
surface structure and moisture is measured
all sky and all weather
105
What are the properties to passive mw
emitted from the earth
temperature
precipitation, moisture
not must interference from clouds
106
What are the properties of lidar
sensors (light)
reflected or individual photon
topography, tree height, and surface structure
cloud can get in the way
measures reflected photon from the cloud
107
Examples of supervised classification
nearest neighbor
maximum likelihood
decision tree/ random forest
neural network
support vector machine
kNN
108
Examples of unsupervised classification
produce clusters
kmeans
ISODATA
label clusters
109
commission (user's accuracy)
you label it as class j but the reference map is not class j
110
omission (producers accuracy)
should be class j but not labeled as class j. In the reference map it should be j but on your map it is class i
111
User accuracy equation
uu/ total number of u across
112
Producer accuracy equation
uu/ total number of u down
113
user accuracy definition
the proportion of the area mapped as class i that has reference class i
114
producers accuracy
of class j the proportion of the area of reference class j that is mapped as class j
115
emissivity
emitting ability of a real material, compared to that of a blackbody
116
blackbody
an object that absorbs all electromagnetic radiation that falls onto it
117
greybody
imperfect blackbody. The ratio of a gray body's thermal radiation to a black body's thermal radiation at the same temperature is called the emissivity
118
selective emitter
119
Thermal Range
1.30um-12um
120
Thermal Infrared radiation
fire monitoring
sea surface temperature
water stress in agriculture
volcano monitoring
surveillance
assessing inductrial heat efficiency
121
measuring kinetic temperature
contact involved, internal manifestation of temp
122
Measuring radiant temperature
no contact involved -RS
external manifestation of temp
123
Kinetic heat
is the energy of particles in random motion
measured with a thermometer
124
Planck's Law
the temperature of the matter defines the wavelengths of the EM radiation that being emitted
125
Radiant Temperature
the radiant flux being emitted by an object because of its temperature
126
Kirchoff radiation law
the spectral emissivity of an object equals its spectral absorbance
127
The lower an object's reflectance...
the higher its emissivity
128
Emissivity and color
darker colors are better absorbers/ emitters
129
Surface Roughness
as surface roughness increases, emissivity increases
130
Moisture Content
as moisture increases, emissivity increase
131
EM wavelength
emissivity varies with wavelength
132
viewing angle
emissivity varies with viewing angle
133
Sources of temperature variation
the sun, atm, diurnal and seasonal cycles of temperature changes, and thermal capacity
