Image enhancement and interpretation

IImmaaggee EEnnhhaanncceemmeenntt aanndd IInntteerrpprreettaattiioonn
Mahesh Kumar Jat
Mahesh Kumar Jat
Department of Civil Engineering
Department of Civil Engineering
Malaviya National Institute of Technology,
Malaviya National Institute of Technology,
Jaipur
Jaipur

IImmaaggee EEnnhhaanncceemmeenntt
 Reduction
 Magnification
 Spatial Profiles
 Spectral Profiles
 Ratioing
 Contrast Stretching
 Frequency Filtering
 Edge Enhancement
 Vegetation Indices
 Texture
 Reduction
 Magnification
 Spatial Profiles
 Spectral Profiles
 Ratioing
 Contrast Stretching
 Frequency Filtering
 Edge Enhancement
 Vegetation Indices
 Texture

Integer Image
Reduction
Integer Image
Reduction

Integer Image
Magnification
Integer Image
Magnification

BBaanndd RRaattiiooiinngg
i j k
i j ratio BV
, ,
i j l
BV
BV
, ,
, , =
where:
- BVi,j,k is the original input brightness value in band k
- BVi,j,l is the original input brightness value in band l
- BVi,j,ratio is the ratio output brightness value
where:
- BVi,j,k is the original input brightness value in band k
- BVi,j,l is the original input brightness value in band l
- BVi,j,ratio is the ratio output brightness value

BBaanndd RRaattiiooiinngg
Ratio values within the range 1/255 to 1 are assigned
values between 1 and 128 by the function:
Ratio values within the range 1/255 to 1 are assigned
values between 1 and 128 by the function:
[( 127) 1] , , , , = ´ + i j n i j r BV Int BV
Ratio values from 1 to 255 are assigned values within
the range 128 to 255 by the function:
Ratio values from 1 to 255 are assigned values within
the range 128 to 255 by the function:
ö
÷ ÷ø
æ
ç çè
128 , ,
= +
2
, ,
i j r
i j n
BV
BV Int

Band Ratioing
of Charleston,
SC Landsat
Thematic
Mapper Data
Band Ratioing
of Charleston,
SC Landsat
Thematic
Mapper Data

BBaanndd RRaattiioo IImmaaggee
Landsat TM
Band 4 / Band 3
Landsat TM
Band 4 / Band 3

Band
Ratio
Band
Ratio
SPOT HRV
Band 3 / Pan
SPOT HRV
Band 3 / Pan

Band
Ratio
Band
Ratio
Landsat TM
Band 40 /
Band 15
Landsat TM
Band 40 /
Band 15

Spatial Profile
-Transect-
Spatial Profile
-Transect-
22000044

Spectral Profile
of SPOT 20 x 20 m
Multispectral Data of
Marco Island, Florida
Spectral Profile
of SPOT 20 x 20 m
Multispectral Data of
Marco Island, Florida
22000044

Spectral Profile
of HYMAP 3 x 3 m
Hyperspectral Data of
Debordieu Colony near
Spectral Profile
of HYMAP 3 x 3 m
Hyperspectral Data of
Debordieu Colony near
North Inlet, SC
North Inlet, SC
22000044

SSttaannddaarrdd DDeevviiaattiioonn CCoonnttrraasstt SSttrreettcchh

Common
Common
Symmetric and
Symmetric and
Skewed
Skewed
Distributions in
Remotely Sensed
Distributions in
Remotely Sensed
Data
Data

Min-Max
Contrast
Stretch
Min-Max
Contrast
Stretch
+1 Standard
Deviation
Contrast
Stretch
+1 Standard
Deviation
Contrast
Stretch

Linear Contrast Enhancement:
Minimum- Maximum Contrast Stretch
k
BV = BV -
min
in k
÷ø
quant ÷ out k k
ö
ç çè æ
-
max min
where:
- BVin is the original input brightness value
- quantk is the range of the brightness values that can be
where:
- BVin is the original input brightness value
- quantk is the range of the brightness values that can be
displayed on the CRT (e.g., 255),
displayed on the CRT (e.g., 255),
- mink is the minimum value in the image,
- maxk is the maximum value in the image, and
- BVout is the output brightness value
- mink is the minimum value in the image,
- maxk is the maximum value in the image, and
- BVout is the output brightness value

min
ö
255 255
4 4
= -
105 4
ö 105 4
çè
255 0
105 4
max min
= ÷ø
æ
-
= ÷ ÷ø
ç çè æ
-
= -
in
BV
out
in
out
BV
All other original brightness values
between 5 and 104 are linearly
distributed between 0 and 255.
All other original brightness values
between 5 and 104 are linearly
distributed between 0 and 255.

Contrast Stretch of
Charleston, SC Landsat
Contrast Stretch of
Charleston, SC Landsat
Thematic Mapper Band 4 Data
Thematic Mapper Band 4 Data
OOrirgigininaal l
Minimum-maximum
Minimum-maximum
+1 standard
deviation
+1 standard
deviation

Contrast Stretching of Charleston, SC
Landsat Thematic Mapper Band 4 Data
Contrast Stretching of Charleston, SC
Landsat Thematic Mapper Band 4 Data
Specific percentage
linear contrast stretch
designed to highlight
Specific percentage
designed to highlight
wetland
wetland
HHisistotoggraramm E Eqquuaalilzizaatitoionn

Contrast Stretching of Predawn
Thermal Infrared Data of the
Contrast Stretching of Predawn
Thermal Infrared Data of the
the Savannah River
the Savannah River
OOrirgigininaal l
Minimum-maximum
Minimum-maximum
+1 standard
deviation
+1 standard
deviation

Specific percentage
designed to highlight the
Specific percentage
designed to highlight the
thermal plume
thermal plume
HHisistotoggraramm E Eqquuaalilzizaatitoionn
Contrast Stretching of Predawn Thermal
Infrared Data of the the Savannah River
Contrast Stretching of Predawn Thermal
Infrared Data of the the Savannah River

NNoonn--lliinneeaarr CCoonnttrraasstt SSttrreettcchhiinngg
PPiieecceewwiissee

Piecewise Linear
Contrast Stretching
Piecewise Linear
Contrast Stretching

NNoonn--lliinneeaarr CCoonnttrraasstt SSttrreettcchhiinngg
Logarithmic and
Inverse Log
Logarithmic and
Inverse Log

Spatial Filtering to Enhance Low- and
High-Frequency Detail and Edges
A characteristics of remotely sensed
images is a parameter called spatial
frequency, defined as the number of
changes in brightness value per unit
distance for any particular part of an
image.
A characteristics of remotely sensed
images is a parameter called spatial
frequency, defined as the number of
changes in brightness value per unit
distance for any particular part of an
image.

Spatial frequency in remotely sensed imagery may be
enhanced or subdued using two different approaches:
Spatial frequency in remotely sensed imagery may be
enhanced or subdued using two different approaches:
- Spatial convolution filtering based primarily on the
use of convolution masks, and
- Spatial convolution filtering based primarily on the
use of convolution masks, and
- Fourier analysis which mathematically separates an
image into its spatial frequency components.
- Fourier analysis which mathematically separates an
image into its spatial frequency components.

SSppaattiiaall CCoonnvvoolluuttiioonn FFiilltteerriinngg
A linear spatial filter is a filter for which the brightness
value (BVi,j,out) at location i,j in the output image is a
function of some weighted average (linear
combination) of brightness values located in a
particular spatial pattern around the i,j location in the
input image.
A linear spatial filter is a filter for which the brightness
value (BVi,j,out) at location i,j in the output image is a
function of some weighted average (linear
combination) of brightness values located in a
particular spatial pattern around the i,j location in the
input image.
The process of evaluating the weighted neighboring
pixel values is called two-dimensional convolution
filtering.
The process of evaluating the weighted neighboring
pixel values is called two-dimensional convolution
filtering.

The size of the neighborhood convolution mask or
kernel (n) is usually 3 x 3, 5 x 5, 7 x 7, or 9 x 9.
The size of the neighborhood convolution mask or
kernel (n) is usually 3 x 3, 5 x 5, 7 x 7, or 9 x 9.
We will constrain our discussion to 3 x 3 convolution
masks with nine coefficients, ci, defined at the
following locations:
We will constrain our discussion to 3 x 3 convolution
masks with nine coefficients, ci, defined at the
following locations:
c1 c2 c3
c1 c2 c3
Mask template = c4 c5 c6
Mask template = c4 c5 c6
c7 c8 c9
c7 c8 c9
1 1 1
1 1 1
1 1 1

The coefficients, c1, in the mask are multiplied by the
following individual brightness values (BVi) in the
input image:
The coefficients, c1, in the mask are multiplied by the
following individual brightness values (BVi) in the
input image:
c1 x BV1 c2 x BV2 c3 x BV3
Mask template = c4 x BV4 c5 x BV5 c6 x BV6
Mask template = c4 x BV4 c5 x BV5 c6 x BV6
The primary input pixel under investigation at any one
time is BV5 = BVi,j
The primary input pixel under investigation at any one
time is BV5 = BVi,j

Various
Convolution
Various
Convolution
Mask
Kernels
Mask
Kernels

Spatial Convolution Filtering:
Low Frequency Filter
Low Frequency Filter
ö
æ
9
1
c x BV
n
ö çè= æ + + +
÷ø
÷ ÷ ÷ ÷
ø
ç ç ç ç
è
=
å=
int ...
9
int
1 2 3 9
5,
BV BV BV BV
LFF
i
i
i
out
1
1
1
1
1
1
1
1
1

LLooww PPaassss FFiilltteerr
3 = 27
9
4 = 36
9
5 = 45
9

Spatial
Frequency
Filtering
Spatial
Frequency
Filtering

Median Filter
Median Filter
A median filter has certain advantages when compared
with weighted convolution filters, including: 1) it does
not shift boundaries, and 2) the minimal degradation to
edges allows the median filter to be applied repeatedly
which allows fine detail to be erased and large regions
to take on the same brightness value (often called
posterization).
A median filter has certain advantages when compared
with weighted convolution filters, including: 1) it does
not shift boundaries, and 2) the minimal degradation to
edges allows the median filter to be applied repeatedly
which allows fine detail to be erased and large regions
to take on the same brightness value (often called
posterization).

Minimum or Maximum Filters
Minimum or Maximum Filters
Operating on one pixel at a time, these filters examine
the brightness values of adjacent pixels in a user-specified
Operating on one pixel at a time, these filters examine
the brightness values of adjacent pixels in a user-specified
radius (e.g., 3 x 3 pixels) and replace the
radius (e.g., 3 x 3 pixels) and replace the
brightness value of the current pixel with the minimum
or maximum brightness value encountered, respectively.
brightness value of the current pixel with the minimum
or maximum brightness value encountered, respectively.

High Frequency Filter
High Frequency Filter
High-pass filtering is applied to imagery to remove the
slowly varying components and enhance the high-frequency
High-pass filtering is applied to imagery to remove the
slowly varying components and enhance the high-frequency
local variations. One high-frequency filter
local variations. One high-frequency filter
(HFF5,out) is computed by subtracting the output of the
low-frequency filter (LFF5,out) from twice the value of
the original central pixel value, BV5:
(HFF5,out) is computed by subtracting the output of the
low-frequency filter (LFF5,out) from twice the value of
the original central pixel value, BV5:
out out HFF x BV LFF 5, 5 5, = (2 ) -

Unequal-weighted smoothing Filter
Unequal-weighted smoothing Filter
0.25 0.50 0.25
0.50 1 0.50
0.25 0.50 0.25
1 1 1
1 2 1
1 1 1

Edge Enhancement
Edge Enhancement
For many remote sensing Earth science applications, the
most valuable information that may be derived from an
image is contained in the edges surrounding various
objects of interest. Edge enhancement delineates these
edges and makes the shapes and details comprising the
image more conspicuous and perhaps easier to analyze.
Edges may be enhanced using either linear or nonlinear
edge enhancement techniques.
For many remote sensing Earth science applications, the
most valuable information that may be derived from an
image is contained in the edges surrounding various
objects of interest. Edge enhancement delineates these
edges and makes the shapes and details comprising the
image more conspicuous and perhaps easier to analyze.
Edges may be enhanced using either linear or nonlinear
edge enhancement techniques.

Spatial Convolution Filtering: Directional
First-Difference Linear Edge Enhancement
Spatial Convolution Filtering: Directional
First-Difference Linear Edge Enhancement
Vertical = BV - BV +
K
i j i j
, , 1
Horizontal = BV - BV +
K
i , j i 1,
j
NE Diagonal = BV - BV +
K
i j i j
+ +
, 1, 1
SE Diagonal = BV - BV +
K
i j i j
- +
-
+
, 1, 1
The result of the subtraction can be either negative or
possible, therefore a constant, K (usually 127) is added to
make all values positive and centered between 0 and 255
The result of the subtraction can be either negative or
possible, therefore a constant, K (usually 127) is added to
make all values positive and centered between 0 and 255

Spatial Convolution Filtering: High-pass
Filters that Accentuate or Sharpen Edges
Spatial Convolution Filtering: High-pass
Filters that Accentuate or Sharpen Edges
-1 -1 -1
-1 9 -1
-1 -1 -1
1 -2 1
-2 5 -2
1 -2 1

Linear Edge Enhancement - Embossing
Linear Edge Enhancement - Embossing
0 0 0
1 0 -1
0 0 0
EEmmbboossss EEaasstt
0 0 1
0 0 0
-1 0 0
EEmmbboossss NNWW

Compass Gradient Masks
Compass Gradient Masks
1 1 1
1 -2 1
-1 -1 -1
NNoorrtthh
1 1 1
-1 -2 1
-1 -1 1
NNoorrtthheeaasstt
-1 1 1
-1 -2 1
-1 1 1
EEaasstt

Edge Enhancement Using
Laplacian Convolution Masks
Edge Enhancement Using
The Laplacian is a second derivative (as opposed to the
gradient which is a first derivative) and is invariant to
rotation, meaning that it is insensitive to the direction in
which the discontinuities (point, line, and edges) run.
The Laplacian is a second derivative (as opposed to the
gradient which is a first derivative) and is invariant to
rotation, meaning that it is insensitive to the direction in
which the discontinuities (point, line, and edges) run.

0 -1 0
-1 4 -1
0 -1 0
-1 -1 -1
-1 8 -1
-1 -1 -1
1 -2 1
-2 4 -2
1 -2 1

Spatial Convolution Filtering: Edge
Enhancement Using Laplacian Convolution Masks
The following Laplacian operator may be used to subtract
the Laplacian edges from the original image:
The following Laplacian operator may be used to subtract
the Laplacian edges from the original image:
1 1 1
1 -7 1
1 1 1

By itself, the Laplacian image may be difficult to interpret.
Therefore, a Laplacian edge enhancement may be added back to the
original image using the following mask:
By itself, the Laplacian image may be difficult to interpret.
Therefore, a Laplacian edge enhancement may be added back to the
original image using the following mask:
0 -1 0
-1 5 -1
0 -1 0

Spatial Convolution Filtering: Non-linear
Edge Enhancement Using the Sobel Operator
Spatial Convolution Filtering: Non-linear
Edge Enhancement Using the Sobel Operator
2 2
Sobel = X +
Y out
( ) ( )
( ) ( ) 1 2 3 7 8 9
X = BV + 2 BV + BV - BV + 2
BV +
BV
3 6 9 1 4 7
5,
Y BV 2 BV BV BV 2
BV BV
where
= + + - + +
1
2
4
7 8
3
6
9
order

The Sobel operator may also be computed by
simultaneously applying the following 3 x 3
The Sobel operator may also be computed by
simultaneously applying the following 3 x 3
templates across the image:
templates across the image:
-1 0 1
-2 0 2
-1 0 1
1 2 1
0 0 0
-1 -2 -1
XX == YY = =

Spatial Convolution Filtering: Non-linear Edge
Enhancement Using the Robert’’s Edge Detector
Enhancement Using the Robert’’s Edge Detector
The Robert’’s edge detector is based on the
use of only four elements of a 3 x 3 mask.
The Robert’’s edge detector is based on the
use of only four elements of a 3 x 3 mask.
Roberts X Y 5,
out
X = BV -
BV
5 9
Y BV BV
6 8
where
= -
= +
1
2
4
7 8
3
6
9
order
5

The Robert’s Edge operator may also be
computed by simultaneously applying the
following 3 x 3 templates across the image:
The Robert’s Edge operator may also be
computed by simultaneously applying the
following 3 x 3 templates across the image:
0 0 0
0 1 0
0 0 -1
0 0 0
0 0 1
0 -1 0
XX == YY = =

Enhancement Using the Kirsch Edge Detector
The Kirsch nonlinear edge enhancement
calculates the gradient at pixel location BVi,j
. To apply this operator, however, it is first
necessary to designate a different 3 x 3
window numbering scheme.
The Kirsch nonlinear edge enhancement
calculates the gradient at pixel location BVi,j
. To apply this operator, however, it is first
necessary to designate a different 3 x 3
window numbering scheme.
BV0 BV1 BV2
BV0 BV1 BV2
Kirsh window = BV7 BVi,j BV3
Kirsh window = BV7 BVi,j BV3
BV6 BV5 BV4
BV6 BV5 BV4

î í ì - =
[ ( )]
7
BV Abs S T
, 0 max 1,max 5 3
i j i =
i i
S = BV + BV +
Bv
i i i i
+ +
1 2
T = BV + BV + Bv + BV +
Bv
þ ý ü
i i i i i i
+ + + + +
3 4 5 6 7
where
The subscripts of BV are evaluated modulo 8, meaning that the computation moves
around the perimeter of the mask in eight steps. The edge enhancement computes
the maximal compass gradient magnitude about input image points although the
input pixel value BVi,j is never used in the computation
The subscripts of BV are evaluated modulo 8, meaning that the computation moves
around the perimeter of the mask in eight steps. The edge enhancement computes
the maximal compass gradient magnitude about input image points although the
input pixel value BVi,j is never used in the computation

HHiissttooggrraamm EEqquuaalliizzaattiioonn
• eevvaalluuaatteess tthhee iinnddiivviidduuaall bbrriigghhttnneessss vvaalluueess iinn aa bbaanndd ooff
iimmaaggeerryy aanndd aassssiiggnnss aapppprrooxxiimmaatteellyy aann eeqquuaall nnuummbbeerr ooff
ppiixxeellss ttoo eeaacchh ooff tthhee uusseerr--ssppeecciiffiieedd oouuttppuutt ggrraayy--ssccaallee ccllaasssseess
((ee..gg..,, 3322,, 6644,, aanndd 225566))..
• aapppplliieess tthhee ggrreeaatteesstt ccoonnttrraasstt eennhhaanncceemmeenntt ttoo tthhee mmoosstt
ppooppuullaatteedd rraannggee ooff bbrriigghhttnneessss vvaalluueess iinn tthhee iimmaaggee..
• rreedduucceess tthhee ccoonnttrraasstt iinn tthhee vveerryy lliigghhtt oorr ddaarrkk ppaarrttss ooff tthhee
iimmaaggee aassssoocciiaatteedd wwiitthh tthhee ttaaiillss ooff aa nnoorrmmaallllyy ddiissttrriibbuutteedd
hhiissttooggrraamm..
• eevvaalluuaatteess tthhee iinnddiivviidduuaall bbrriigghhttnneessss vvaalluueess iinn aa bbaanndd ooff
iimmaaggeerryy aanndd aassssiiggnnss aapppprrooxxiimmaatteellyy aann eeqquuaall nnuummbbeerr ooff
ppiixxeellss ttoo eeaacchh ooff tthhee uusseerr--ssppeecciiffiieedd oouuttppuutt ggrraayy--ssccaallee ccllaasssseess
((ee..gg..,, 3322,, 6644,, aanndd 225566))..
• aapppplliieess tthhee ggrreeaatteesstt ccoonnttrraasstt eennhhaanncceemmeenntt ttoo tthhee mmoosstt
ppooppuullaatteedd rraannggee ooff bbrriigghhttnneessss vvaalluueess iinn tthhee iimmaaggee..
• rreedduucceess tthhee ccoonnttrraasstt iinn tthhee vveerryy lliigghhtt oorr ddaarrkk ppaarrttss ooff tthhee
iimmaaggee aassssoocciiaatteedd wwiitthh tthhee ttaaiillss ooff aa nnoorrmmaallllyy ddiissttrriibbuutteedd
hhiissttooggrraamm..

Statistics for a 64 x 64 Hypothetical Image
Statistics for a 64 x 64 Hypothetical Image
with Brightness Values from 0 to 7
with Brightness Values from 0 to 7
44009966 t otototatatall

HHiissttooggrraamm EEqquuaalliizzaattiioonn

Transformation Function, ki for each
Transformation Function, ki for each
individual brightness value
individual brightness value
For each brightness value level BVi in the quantk
range of 0 to 7 of the original histogram, a new
cumulative frequency value ki is calculated:
For each brightness value level BVi in the quantk
range of 0 to 7 of the original histogram, a new
cumulative frequency value ki is calculated:
( ) å=
k =
f BV
quantk
i
i
i n
0
where the summation counts the frequency of pixels
in the image with brightness values equal to or less
than BVi, and n is the total number of pixels in the
where the summation counts the frequency of pixels
in the image with brightness values equal to or less
than BVi, and n is the total number of pixels in the
entire scene (4,096 in this example).
entire scene (4,096 in this example).

The histogram equalization process iteratively compares the
transformation function ki with the original values of li, to determine
which are closest in value. The closest match is reassigned to the
appropriate brightness value.
FFoorr eexxaammppllee,, wwee sseeee tthhaatt kk00 == 00..1199 iiss cclloosseesstt ttoo LL11 == 00..1144.. TThheerreeffoorree,,
aallll ppiixxeellss iinn BBVV00 ((779900 ooff tthheemm)) wwiillll bbee aassssiiggnneedd ttoo BBVV11.. SSiimmiillaarrllyy,,
tthhee 11002233 ppiixxeellss iinn BBVV11 wwiillll bbee aassssiiggnneedd ttoo BBVV33,, tthhee 885500 ppiixxeellss iinn BBVV22
wwiillll bbee aassssiiggnneedd ttoo BBVV55,, tthhee 665566 ppiixxeellss iinn BBVV33 wwiillll bbee aassssiiggnneedd ttoo
BBVV66,, tthhee 332299 ppiixxeellss iinn BBVV44 wwiillll aallssoo bbee aassssiiggnneedd ttoo BBVV66,, aanndd aallll 444488
bbrriigghhttnneessss vvaalluueess iinn BBVV55––77 wwiillll bbee aassssiiggnneedd ttoo BBVV77.. TThhee nneeww iimmaaggee
wwiillll nnoott hhaavvee aannyy ppiixxeellss wwiitthh bbrriigghhttnneessss vvaalluueess ooff 00,, 22,, oorr 44.. TThhiiss iiss
eevviiddeenntt wwhheenn eevvaalluuaattiinngg tthhee nneeww hhiissttooggrraamm.. WWhheenn aannaallyyssttss sseeee ssuucchh
ggaappss iinn iimmaaggee hhiissttooggrraammss,, iitt iiss uussuuaallllyy aa ggoooodd iinnddiiccaattiioonn tthhaatt
hhiissttooggrraamm eeqquuaalliizzaattiioonn oorr ssoommee ootthheerr ooppeerraattiioonn hhaass bbeeeenn aapppplliieedd.
pr The histogram equalization process iteratively compares the
transformation function ki with the original values of li, to determine
which are closest in value. The closest match is reassigned to the
appropriate brightness value.
FFoorr eexxaammppllee,, wwee sseeee tthhaatt kk00 == 00..1199 iiss cclloosseesstt ttoo LL11 == 00..1144.. TThheerreeffoorree,,
aallll ppiixxeellss iinn BBVV00 ((779900 ooff tthheemm)) wwiillll bbee aassssiiggnneedd ttoo BBVV11.. SSiimmiillaarrllyy,,
tthhee 11002233 ppiixxeellss iinn BBVV11 wwiillll bbee aassssiiggnneedd ttoo BBVV33,, tthhee 885500 ppiixxeellss iinn BBVV22
wwiillll bbee aassssiiggnneedd ttoo BBVV55,, tthhee 665566 ppiixxeellss iinn BBVV33 wwiillll bbee aassssiiggnneedd ttoo
BBVV66,, tthhee 332299 ppiixxeellss iinn BBVV44 wwiillll aallssoo bbee aassssiiggnneedd ttoo BBVV66,, aanndd aallll 444488
bbrriigghhttnneessss vvaalluueess iinn BBVV55––77 wwiillll bbee aassssiiggnneedd ttoo BBVV77.. TThhee nneeww iimmaaggee
wwiillll nnoott hhaavvee aannyy ppiixxeellss wwiitthh bbrriigghhttnneessss vvaalluueess ooff 00,, 22,, oorr 44.. TThhiiss iiss
eevviiddeenntt wwhheenn eevvaalluuaattiinngg tthhee nneeww hhiissttooggrraamm.. WWhheenn aannaallyyssttss sseeee ssuucchh
ggaappss iinn iimmaaggee hhiissttooggrraammss,, iitt iiss uussuuaallllyy aa ggoooodd iinnddiiccaattiioonn tthhaatt
hhiissttooggrraamm eeqquuaalliizzaattiioonn oorr ssoommee ootthheerr ooppeerraattiioonn hhaass bbeeeenn aapppplliieedd.

Statistics of How a a 64 x 64 Hypothetical Image with
Brightness Values from 0 to 7 is Histogram Equalized
Statistics of How a a 64 x 64 Hypothetical Image with
Brightness Values from 0 to 7 is Histogram Equalized

PPrriinncciippaall CCoommppoonneennttss AAnnaallyyssiiss
• ttrraannssffoorrmmaattiioonn ooff tthhee rraaww rreemmoottee sseennssoorr ddaattaa uussiinngg PPCCAA
ccaann rreessuulltt iinn nneeww pprriinncciippaall ccoommppoonneenntt iimmaaggeess tthhaatt mmaayy bbee
mmoorree iinntteerrpprreettaabbllee tthhaann tthhee oorriiggiinnaall ddaattaa..
• mmaayy aallssoo bbee uusseedd ttoo ccoommpprreessss tthhee iinnffoorrmmaattiioonn ccoonntteenntt ooff aa
nnuummbbeerr ooff bbaannddss ooff iimmaaggeerryy ((ee..gg..,, sseevveenn TThheemmaattiicc MMaappppeerr
bbaannddss)) iinnttoo jjuusstt ttwwoo oorr tthhrreeee ttrraannssffoorrmmeedd pprriinncciippaall
ccoommppoonneenntt iimmaaggeess.. TThhee aabbiilliittyy ttoo rreedduuccee tthhee ddiimmeennssiioonnaalliittyy
((ii..ee..,, tthhee nnuummbbeerr ooff bbaannddss iinn tthhee ddaattaasseett tthhaatt mmuusstt bbee
aannaallyyzzeedd ttoo pprroodduuccee uussaabbllee rreessuullttss)) ffrroomm nn ttoo ttwwoo oorr tthhrreeee
bbaannddss iiss aann iimmppoorrttaanntt eeccoonnoommiicc ccoonnssiiddeerraattiioonn,, eessppeecciiaallllyy iiff
tthhee ppootteennttiiaall iinnffoorrmmaattiioonn rreeccoovveerraabbllee ffrroomm tthhee ttrraannssffoorrmmeedd
ddaattaa iiss jjuusstt aass ggoooodd aass tthhee oorriiggiinnaall rreemmoottee sseennssoorr ddaattaa..
• ttrraannssffoorrmmaattiioonn ooff tthhee rraaww rreemmoottee sseennssoorr ddaattaa uussiinngg PPCCAA
ccaann rreessuulltt iinn nneeww pprriinncciippaall ccoommppoonneenntt iimmaaggeess tthhaatt mmaayy bbee
mmoorree iinntteerrpprreettaabbllee tthhaann tthhee oorriiggiinnaall ddaattaa..
• mmaayy aallssoo bbee uusseedd ttoo ccoommpprreessss tthhee iinnffoorrmmaattiioonn ccoonntteenntt ooff aa
nnuummbbeerr ooff bbaannddss ooff iimmaaggeerryy ((ee..gg..,, sseevveenn TThheemmaattiicc MMaappppeerr
bbaannddss)) iinnttoo jjuusstt ttwwoo oorr tthhrreeee ttrraannssffoorrmmeedd pprriinncciippaall
ccoommppoonneenntt iimmaaggeess.. TThhee aabbiilliittyy ttoo rreedduuccee tthhee ddiimmeennssiioonnaalliittyy
((ii..ee..,, tthhee nnuummbbeerr ooff bbaannddss iinn tthhee ddaattaasseett tthhaatt mmuusstt bbee
aannaallyyzzeedd ttoo pprroodduuccee uussaabbllee rreessuullttss)) ffrroomm nn ttoo ttwwoo oorr tthhrreeee
bbaannddss iiss aann iimmppoorrttaanntt eeccoonnoommiicc ccoonnssiiddeerraattiioonn,, eessppeecciiaallllyy iiff
tthhee ppootteennttiiaall iinnffoorrmmaattiioonn rreeccoovveerraabbllee ffrroomm tthhee ttrraannssffoorrmmeedd
ddaattaa iiss jjuusstt aass ggoooodd aass tthhee oorriiggiinnaall rreemmoottee sseennssoorr ddaattaa..

The spatial relationship between the first two principal components: (a) Scatter-plot of data points collected from
two remotely bands labeled X1 and X2 with the means of the distribution labeled μ1 and μ2. (b) A new coordinate
system is created by shifting the axes to an X¢ system. The values for the new data points are found by the
relationship X1¢ = X1 – μ1 and X2¢ = X2 – μ2. (c) The X¢ axis system is then rotated about its origin (μ1, μ2) so that PC1 is
projected through the semi-major axis of the distribution of points and the variance of PC1 is a maximum. PC2 must
be perpendicular to PC1. The PC axes are the principal components of this two-dimensional data space. Component
1 usually accounts for approximately 90% of the variance, with component 2 accounting for approximately 5%.
The spatial relationship between the first two principal components: (a) Scatter-plot of data points collected from
two remotely bands labeled X1 and X2 with the means of the distribution labeled μ1 and μ2. (b) A new coordinate
system is created by shifting the axes to an X¢ system. The values for the new data points are found by the
relationship X1¢ = X1 – μ1 and X2¢ = X2 – μ2. (c) The X¢ axis system is then rotated about its origin (μ1, μ2) so that PC1 is
projected through the semi-major axis of the distribution of points and the variance of PC1 is a maximum. PC2 must
be perpendicular to PC1. The PC axes are the principal components of this two-dimensional data space. Component
1 usually accounts for approximately 90% of the variance, with component 2 accounting for approximately 5%.

SSttaattiissttiiccss UUsseedd iinn tthhee PPrriinncciippaall CCoommppoonneennttss AAnnaallyyssiiss

1. The n ´ n covariance matrix, Cov, of
the n-dimensional remote sensing
data set to be transformed is
computed. Use of the covariance
matrix results in an unstandardized
PCA, whereas use of the
correlation matrix results in a
standardized PCA.
2. The eigenvalues, E = [l1,1, l2,2, l3,3,
…, ln.n], and eigenvectors EV = [akp
… for k = 1 to n bands, and p = 1 to
n components] of the covariance
matrix are computed such that:

E E
where EVT is the transpose of the eigenvector matrix, EV, and E is a diagonal
covariance matrix whose elements, li,i, called eigenvalues, are the variances
where EVT is the transpose of the eigenvector matrix, EV, and E is a diagonal
covariance matrix whose elements, li
,i, called eigenvalues, are the variances
of the pth principal components, where p = 1 to n components.
of the pth principal components, where p = 1 to n components.

Eigenvalues Computed Eigenvalues Computed ffoorr tthhee CCoovvaarriiaannccee MMaattrriixx
p = 1
7
p = 1

Eigenvectors (ap) (Factor Scores) Computed
Eigenvectors (ap) (Factor Scores) Computed
for the Covariance Matrix
for the Covariance Matrix

Correlation, Rk,pCorrelation, R , Between Band k and Each Principal Component p k,p, Between Band k and Each Principal Component p
where:
ak,p = eigenvector for band k and component p
lp = pth eigenvalue
Vark = variance of band k in the
where:
ak,p = eigenvector for band k and component p
lp = pth eigenvalue
Vark = variance of band k in the
covariance matrix
covariance matrix

It is possible to compute a new value for pixel 1,1 (it has 7 bands) in
principal component number 1 using the following equation:
It is possible to compute a new value for pixel 1,1 (it has 7 bands) in
principal component number 1 using the following equation:
n
å=
newBV =
a BV
i , j , p kp i , j ,
k k
1
original
original
remote sensor
data for pixel
remote sensor
data for pixel
1,1
1,1
where akp = eigenvectors, BVi,j,k = brightness value in band k for the pixel at
row i, column j, and n = number of bands.
where akp = eigenvectors, BVi,j,k = brightness value in band k for the pixel at
row i, column j, and n = number of bands.

1,1,1 1,1 1,1,1 2,1 1,1,2 3,1 1,1,3 newBV = a BV + a BV + a BV
4,1 1,1,4 5,1 1,1,5 6,1 1,1,6 7,1 1,1,7 + a BV + a BV + a BV + a BV
0.205(20) 0.127(30) 0.204(22) 1,1,1 newBV = + +
+ 0.443(60) + 0.742(70) + 0.376(62) + 0.106(50)
=119.53

Principal
Components
Analysis (PCA)
Principal
Components
Analysis (PCA)

VVeeggeettaattiioonn TTrraannssffoorrmmaattiioonn ((IInnddiicceess))

RReemmoottee SSeennssiinngg ooff VVeeggeettaattiioonn
Spectral
Characteristics

Dominant Factors DDoommiinnaanntt FFaaccttoorrss CCoonnttrroolllliinngg LLeeaaff RReefflleeccttaannccee
Water
Water
absorption bands:
absorption bands:
0.97 mm
1.19 mm
1.45 mm
1.94 mm
2.70 mm
0.97 mm
1.19 mm
1.45 mm
1.94 mm
2.70 mm
JeJnesnesne,n ,2 0200044

Cross-section Through A
Hypothetical and Real
Leaf Revealing the Major
Structural Components
Cross-section Through A
Hypothetical and Real
Leaf Revealing the Major
Structural Components
that Determine the
Spectral Reflectance
that Determine the
of Vegetation
of Vegetation
JeJnesnesne,n 2, 0200404

Absorption Spectra of Chlorophyll a and b, b-
carotene, Pycoerythrin, and Phycocyanin Pigments
Absorption Spectra of Chlorophyll a and b, b-
carotene, Pycoerythrin, and Phycocyanin Pigments
lack of
absorption
lack of
absorption
• Chlorophyll a peak absorption is at 0.43 and 0.66 mm.
• Chlorophyll b peak absorption is at 0.45 and 0.65 mm.
• Optimum chlorophyll absorption windows: 0.45 - 0.52 mm and 0.63 - 0.69 mm
• Chlorophyll a peak absorption is at 0.43 and 0.66 mm.
• Chlorophyll b peak absorption is at 0.45 and 0.65 mm.
• Optimum chlorophyll absorption windows: 0.45 - 0.52 mm and 0.63 - 0.69 mm

Litton Emerge Spatial, Inc., CIR image
(RGB = NIR,R,G) of Dunkirk, NY, at 1 x
1 m obtained on December 12, 1998.
Litton Emerge Spatial, Inc., CIR image
(RGB = NIR,R,G) of Dunkirk, NY, at 1 x
1 m obtained on December 12, 1998.
Natural color image (RGB = RGB) of a N.Y.
Power Authority lake at 1 x 1 ft obtained on
Natural color image (RGB = RGB) of a N.Y.
Power Authority lake at 1 x 1 ft obtained on
October 13, 1997.
October 13, 1997.

Characteristics of
Sweetgum Leaves
Characteristics of
Sweetgum Leaves
(Liquidambar
styraciflua L.)
(Liquidambar
styraciflua L.)

1 2
a
3
4
35
35
30
30
25
25
20
20
15
15
10
10
5
0
Blue
Percent Reflectance
Percent Reflectance
5
(0.45 - 0.52 m m)
Green leaf
Yellow
Red/orange
Brown
1
2
3
4
45
40
Green
(0.52 - 0.60 m m)
Red
(0.63 - 0.69 m m)
1
2
3
Near-Infrared
(0.70 - 0.92 m m)
a.
b.
c.
d.
1 2
a
3
4
0
Blue
(0.45 - 0.52 m m)
Green leaf
Yellow
Red/orange
Brown
4
45
40
Green
(0.52 - 0.60 m m)
Red
(0.63 - 0.69 m m)
Near-Infrared
(0.70 - 0.92 m m)
a.
b.
c.
d.
Characteristics of
Selected Areas of
Blackjack Oak Leaves
Characteristics of
Selected Areas of
Blackjack Oak Leaves

Hypothetical
Example of
Additive
Reflectance from
A Canopy with
Two Leaf Layers
Hypothetical
Example of
Additive
Reflectance from
A Canopy with
Two Leaf Layers

Distribution of Pixels in a Scene in
Distribution of Pixels in a Scene in
Red and Near-infrared Multispectral Feature Space
Red and Near-infrared Multispectral Feature Space

Reflectance Response of a Single Magnolia Leaf
Reflectance Response of a Single Magnolia Leaf
(Magnolia grandiflora) to Decreased Relative Water Content
(Magnolia grandiflora) to Decreased Relative Water Content

Airborne Visible
Infrared Imaging
Airborne Visible
Infrared Imaging
Spectrometer (AVIRIS)
Datacube of Sullivan’s
Island Obtained on
Spectrometer (AVIRIS)
Datacube of Sullivan’s
Island Obtained on
October 26, 1998
October 26, 1998

Imaging Spectrometer Data of Healthy Green Vegetation in the San
Luis Valley of Colorado Obtained on September 3, 1993 Using AVIRIS
Imaging Spectrometer Data of Healthy Green Vegetation in the San
Luis Valley of Colorado Obtained on September 3, 1993 Using AVIRIS
JeJnesnesne,n 2, 0200000 222244 c chhaannnneelsls e eaacchh 1 100 n nmm w wididee w witihth 2 200 x x 2 200 m m p pixixeelsls

Hyperspectral Analysis
Hyperspectral Analysis
of AVIRIS Data
of AVIRIS Data
Obtained on September
3, 1993 of San Luis
Obtained on September
3, 1993 of San Luis
Valley, Colorado
Valley, Colorado
JeJnesnsnesne,n,n 2,2, 0202000000000

RReemmoottee SSeennssiinngg ooff VVeeggeettaattiioonn
Temporal
(Phenological)
Characteristics
Temporal
(Phenological)
Characteristics

Predicted Percent PPrreeddiicctteedd PPeerrcceenntt C Clloloouudd C Coovveerr i ininn F Foouurr A Arreeaass i ininn t ththhee U Unniititteteedd S Sttataatteteess

Phenological Cycle PPhheennoollooggiiccaall CCyyccllee ooff HHaarrdd RReedd W Wiininntteteerr W Whheeaatt i ininn t ththhee G Grreeaatt P Pllalaaiininnss
SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG
crop establishment
50 108 days 28 34 29 21
10 14
greening up heading mature
50 108 days 28 34 29 21
14 14 21 13 25 4 7 9 5
26
Sow Tillering
Emergence
Dormancy Growth
resumes
Heading
Boot
Harvest
Dead
ripe
Harvest
Soft Hard dough
dough
Jointing
Maximum Coverage
Winter Wheat
Phenology
snow cover
SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG
crop establishment
10 14
greening up heading mature
14 14 21 13 25 4 7 9 5
26
Sow Tillering
Emergence
Dormancy Growth
resumes
Heading
Boot
Dead
ripe
Soft Hard dough
dough
Jointing
Maximum Coverage
Winter Wheat
Phenology
snow cover

Phenological Cycles of
Phenological Cycles of
San Joaquin and
Imperial Valley,
California Crops and
Landsat Multispectral
Scanner Images of One
San Joaquin and
Imperial Valley,
California Crops and
Landsat Multispectral
Scanner Images of One
Field During A
Growing Season
Field During A
Growing Season

Landsat Thematic
Mapper Imagery of
Landsat Thematic
Mapper Imagery of
the Imperial
the Imperial
Valley, California
Valley, California
Obtained on
Obtained on
December 10, 1982
December 10, 1982
Band 1 (blue; 0.45 – 0.52 mm) Band 2 (green; 0.52 – 0.60 mm) Band 3 (red; 0.63 – 0.69 mm)
Band 1 (blue; 0.45 – 0.52 mm) Band 2 (green; 0.52 – 0.60 mm) Band 3 (red; 0.63 – 0.69 mm)
Band 4 (near-infrared; 0.76 – 0.90 mm) Band 5 (mid-infrared; 1.55 – 1.75 mm) Band 7 (mid-infrared; 2.08 – 2.35 mm)
Band 4 (near-infrared; 0.76 – 0.90 mm) Band 5 (mid-infrared; 1.55 – 1.75 mm) Band 7 (mid-infrared; 2.08 – 2.35 mm)
Band 6 (thermal infrared; 10.4 – 12.5 mm)
Sugarbeets
Alfalfa
Cotton
Fallow
feed
lot
fl
Ground Reference
Landsat Thematic Mapper
Imagery of
Imperial Valley, California,
December 10, 1982
Band 6 (thermal infrared; 10.4 – 12.5 mm)
Sugarbeets
Alfalfa
Cotton
Fallow
feed
lot
fl
Ground Reference
Imagery of
Imperial Valley, California,
December 10, 1982

Landsat Thematic
Mapper Color
Composites and
Classification Map of a
Portion of the Imperial
Valley, California
Landsat Thematic
Mapper Color
Composites and
Classification Map of a
Portion of the Imperial
Valley, California

Phenological Cycles
of Soybeans and
Phenological Cycles
of Soybeans and
Corn in South
Carolina
Corn in South
Carolina
snow cover
snow cover
125
125
100
100
75
75
50
50
50%
50%
JAN FEB MAR APR MAY JUN SEP OCT NOV DEC
Initial growth Maturity Harvest
25 cm height
Dormant or multicropped
Soybeans
100%
ground
cover
JUL AUG
Development
snow cover
snow cover
300
300
250
250
200
200
150
150
125
125
75
75
50
50
12-14
leaf
12-14
leaf
8-leaf Tassle Dent/Harvest
25 cm height
JUL AUG
100
10 - 12 leaf
Blister
Corn
100%
50%
a.
b.
Initial growth Maturity Harvest
25 cm height
Soybeans
100%
ground
cover
JUL AUG
Development
8-leaf Tassle Dent/Harvest
25 cm height
JUL AUG
100
10 - 12 leaf
Blister
Corn
100%
50%
a.
b.
SSooyybbeeaannss
CCoornrn

Phenological Cycles
of Winter Wheat,
Cotton, and Tobacco
in South Carolina
Phenological Cycles
of Winter Wheat,
Cotton, and Tobacco
in South Carolina
Winter Wheat
Winter Wheat
100%
ground
cover
100%
ground
cover
snow cover
snow cover
100
100
75
75
50
50
25 cm
25 cm
50%
50%
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Tillering Jointing Booting Head Harvest Dormant or multicropped
Seed
Winter Wheat
Phenology
Winter Wheat
Phenology
150
150
125
125
100
100
75
75
50
50
b.
b.
Seeding Boll Maturity/harvest
snow cover
25 cm height
Cotton
100%
ground
cover
JUL AUG
Fruiting
Pre-bloom
50%
snow cover
snow cover
125
125
100
100
75
75
50
50
50%
50%
Transplanting Development Maturity/harvest
25 cm height
Tobacco
100%
JUL AUG
Topping
a.
c.
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Tillering Jointing Booting Head Harvest Dormant or multicropped
Seed
Seeding Boll Maturity/harvest
snow cover
25 cm height
Cotton
100%
ground
cover
JUL AUG
Fruiting
Pre-bloom
50%
Transplanting Development Maturity/harvest
25 cm height
Tobacco
100%
JUL AUG
Topping
a.
c.
WWininteter rW Whheeaatt
CCootttotonn
TToobbaaccccoo

VVeeggeettaattiioonn IInnddiicceess

Reflectance Curves Reflectance Curves ffoorr SSeelleecctteedd PPhheennoommeennaa

Infrared/R IInnffrraarreedd//RReedd RRaattiioo VVeeggeettaattiioonn IInnddeexx
The near-infrared (NIR) to red simple ratio (SR) is the first true vegetation
index:
The near-infrared (NIR) to red simple ratio (SR) is the first true vegetation
index:
SR = NIR
red
It takes advantage of the inverse relationship between chlorophyll absorption of
red radiant energy and increased reflectance of near-infrared energy for healthy
plant canopies (Cohen, 1991) .
It takes advantage of the inverse relationship between chlorophyll absorption of
red radiant energy and increased reflectance of near-infrared energy for healthy
plant canopies (Cohen, 1991) .

Normalized NNoorrmmaalliizzeedd DDiiffffeerreennccee VVeeggeettaattiioonn IInnddeexx
The generic normalized difference vegetation index (NDVI):
The generic normalized difference vegetation index (NDVI):
NDVI = NIR - red
NIR + red
has provided a method of estimating net primary production over varying
biome types (e.g. Lenney et al., 1996), identifying ecoregions (Ramsey et
al., 1995), monitoring phenological patterns of the earth’’s vegetative
surface, and of assessing the length of the growing season and dry-down
periods (Huete and Liu, 1994).
has provided a method of estimating net primary production over varying
biome types (e.g. Lenney et al., 1996), identifying ecoregions (Ramsey et
al., 1995), monitoring phenological patterns of the earth’’s vegetative
surface, and of assessing the length of the growing season and dry-down
periods (Huete and Liu, 1994).

Vegetation
Indices
Vegetation
Indices

Time Series of 1984 and 1988 NDVI Measurements Derived from AVHRR Global
Area Coverage (GAC) Data for the Region around El Obeid, Sudan, in Sub-Saharan
Time Series of 1984 and 1988 NDVI Measurements Derived from AVHRR Global
Area Coverage (GAC) Data for the Region around El Obeid, Sudan, in Sub-Saharan
Africa
Africa

VVeeggeettaattiioonn IInnddiicceess
Simple ratio red
NIR
NDVI = MSS -
MSS
6 6 5
MSS +
MSS
6 5
NDVI = MSS -
MSS
7 7 5
MSS +
MSS
7 5
NDVI = TM -
TM
4 3
TM +
TM TM
4 3
=

IInnffrraarreedd IInnddeexx
An Infrared Index (II) that incorporates both near and middle-infrared
bands is sensitive to changes in plant biomass and water stress in smooth
cordgrass studies (Hardisky et al., 1983; 1986):
An Infrared Index (II) that incorporates both near and middle-infrared
bands is sensitive to changes in plant biomass and water stress in smooth
cordgrass studies (Hardisky et al., 1983; 1986):
Healthy, mono-specific stands of tidal wetland such as Spartina often
exhibit much lower reflectance in the visible (blue, green, and red)
wavelengths than typical terrestrial vegetation due to the saturated tidal
flat understory. In effect, the moist soil absorbs almost all energy incident
to it. This is why wetland often appear surprisingly dark on traditional
infrared color composites.
Healthy, mono-specific stands of tidal wetland such as Spartina often
exhibit much lower reflectance in the visible (blue, green, and red)
wavelengths than typical terrestrial vegetation due to the saturated tidal
flat understory. In effect, the moist soil absorbs almost all energy incident
to it. This is why wetland often appear surprisingly dark on traditional
infrared color composites.
II = NIRTM4 - MIRTM5
NIRTM4 + MIRTM5

Kauth-Thomas
Transformation
Kauth-Thomas
Transformation

Kauth-Thomas “ Kauth-Thomas “TTaasssseelleedd CCaapp”” TTrraannssffoorrmmaattiioonn

Kauth-Thomas
Transformation
Kauth-Thomas
Transformation
November 9, 1982
November 9, 1982

SSooiilill AAddjjujuusstteteedd V Veeggeettataattitiioioonn I Innddeexx ( (SSAAVVII))
Recent emphasis has been given to the development of improved vegetation indices that
may take advantage of calibrated hyperspectral sensor systems such as the moderate
resolution imaging spectrometer - MODIS (Running et al., 1994). The improved indices
incorporate a soil adjustment factor and/or a blue band for atmospheric normalization.
The soil adjusted vegetation index (SAVI) introduces a soil calibration factor, L, to the
NDVI equation to minimize soil background influences resulting from first order soil-plant
Recent emphasis has been given to the development of improved vegetation indices that
may take advantage of calibrated hyperspectral sensor systems such as the moderate
resolution imaging spectrometer - MODIS (Running et al., 1994). The improved indices
incorporate a soil adjustment factor and/or a blue band for atmospheric normalization.
The soil adjusted vegetation index (SAVI) introduces a soil calibration factor, L, to the
NDVI equation to minimize soil background influences resulting from first order soil-plant
spectral interactions (Huete et al., 1994):
spectral interactions (Huete et al., 1994):
SAVI = (1 + L)(NIR - red)
NIR + red + L
An L value of 0.5 minimizes soil brightness variations and eliminates the need for
additional calibration for different soils (Huete and Liu, 1994).
An L value of 0.5 minimizes soil brightness variations and eliminates the need for
additional calibration for different soils (Huete and Liu, 1994).

Soil and Atmospherically Adjusted SSooilil l a anndd A Atmtmmoosspphheerriciccaallllylyy A Addjujuussteteedd V Veeggeetatatatititoioionn I Innddeexx ( (S(SSAARRVVII)I))
Huete and Liu (1994) integrated the L function from SAVI and a blue-band
normalization to derive a soil and atmospherically resistant vegetation index
(SARVI) that corrects for both soil and atmospheric noise:
Huete and Liu (1994) integrated the L function from SAVI and a blue-band
normalization to derive a soil and atmospherically resistant vegetation index
(SARVI) that corrects for both soil and atmospheric noise:
where
where
SARVI = p * nir - p * rb
p * nir + p * rb
p * rb = p* red -g (p * blue - p * red)
The technique requires prior correction for molecular scattering and ozone
absorption of the blue, red, and near-infrared remote sensor data, hence the term p*.
The technique requires prior correction for molecular scattering and ozone
absorption of the blue, red, and near-infrared remote sensor data, hence the term p*.

EEnnhhaanncceedd VVeeggeettaattiioonn IInnddeexx ((EEVVII))
The MODIS Land Discipline Group proposed the Enhanced Vegetation Index (EVI) for use
with MODIS Data:
The MODIS Land Discipline Group proposed the Enhanced Vegetation Index (EVI) for use
with MODIS Data:
EVI = p * nir - p * red
p * nir + C1p * red - C2p * blue + L
The EVI is a modified NDVI with a soil adjustment factor, L, and two coefficients, C1 and C2
which describe the use of the blue band in correction of the red band for atmsoperhic aerosol
scattering. The coefficients, C1 , C2 , and L, are empirically determined as 6.0, 7.5, and 1.0,
respectively. This algorithm has improved sensitivity to high biomass regions and improved
vegetation monitoring thorugh a de-coupling of the canopy background signal and a
reduction in atmospheric influences (Huete and Justice, 1999).
The EVI is a modified NDVI with a soil adjustment factor, L, and two coefficients, C1 and C2
which describe the use of the blue band in correction of the red band for atmsoperhic aerosol
scattering. The coefficients, C1 , C2 , and L, are empirically determined as 6.0, 7.5, and 1.0,
respectively. This algorithm has improved sensitivity to high biomass regions and improved
vegetation monitoring thorugh a de-coupling of the canopy background signal and a
reduction in atmospheric influences (Huete and Justice, 1999).

Murrells
Inlet
Murrells
Inlet
Murrells
Inlet
Murrells
Inlet
MMuurrrreellllss IInnlleett i inn SSoouutthh CCaarroolliinnaa

Phenological Cycle of Smooth Cordgrass
Phenological Cycle of Smooth Cordgrass
(Spartina alterniflora) Biomass in South Carolina
(Spartina alterniflora) Biomass in South Carolina
Smooth Cordgrass (Spartina alterniflora)
Smooth Cordgrass (Spartina alterniflora)
Live Biomass
Dead Biomass
Live Biomass
Dead Biomass
J A S O N
J A S O N JeJnesnesne,n ,2 0200000
1500
1500
1250
1250
1000
1000
750
750
500
500
250
250
0
DryWeightBiomass, g/m2
F M A M J J D
0
DryWeightBiomass, g/m2
F M A M J J D

Phenological Cycle of Cattails PPhheennoollooggiiccaall CCyyccllee ooff CCaattttaaiillss a anndd W Waatteteerrlliliililliliieieess i ininn P Paarr P Poonndd, ,S S.C.C..

Characteristics of the NASA Calibrated Airborne Multispectral
Scanner (CAMS) Mission of Murrells Inlet, S.C. on August 2, 1997
Characteristics of the NASA Calibrated Airborne Multispectral
Scanner (CAMS) Mission of Murrells Inlet, S.C. on August 2, 1997
Altitude CAMS
Mission Relative above- Spatial CAMS
Date Visibility Humidity ground-level Resolution Spectral Resolution
8/2/97 clear 45% 4000’’ 3.08 x 3.08 Band 1 (0.42 - 0.52 mm); blue
Band 2 (0.52 - 0.60 mm); green
Band 3 (0.60 - 0.63 mm); red
Band 4 (0.63 - 0.69 mm); red
Band 5 (0.69 - 0.76 mm); near-
IR
Band 6 (0.76 - 0.90 mm); near-
IR
Band 7 (1.55 - 1.75 mm); mid-IR
Band 8 (2.08 - 2.35 mm); mid-IR
Band 9 (10.5 - 12.5 mm); TIR
Altitude CAMS
Mission Relative above- Spatial CAMS
Date Visibility Humidity ground-level Resolution Spectral Resolution
8/2/97 clear 45% 4000’’ 3.08 x 3.08 Band 1 (0.42 - 0.52 mm); blue
Band 2 (0.52 - 0.60 mm); green
Band 3 (0.60 - 0.63 mm); red
Band 4 (0.63 - 0.69 mm); red
Band 5 (0.69 - 0.76 mm); near-
IR
Band 6 (0.76 - 0.90 mm); near-
IR
Band 7 (1.55 - 1.75 mm); mid-IR
Band 8 (2.08 - 2.35 mm); mid-IR
Band 9 (10.5 - 12.5 mm); TIR

Nine Bands of 3 x 3 m
Calibrated Airborne
Multispectral Scanner
(CAMS) Data of Murrells
Inlet, SC Obtained on
Nine Bands of 3 x 3 m
Calibrated Airborne
(CAMS) Data of Murrells
August 2, 1997
August 2, 1997
JeJnesnesnen, ,2 2000000
Band 1 (blue; 0.45 – 0.52 mm) Band 2 (green; 0.52 – 0.60 mm) Band 3 (red; 0.60 – 0.63mm)
Band 1 (blue; 0.45 – 0.52 mm) Band 2 (green; 0.52 – 0.60 mm) Band 3 (red; 0.60 – 0.63mm)
Band 4 (red; 0.63 – 0.69 mm) Band 5 (near-infrared; 0.69 – 0.76 mm) Band 6 (near-infrared; 0.76 – 0.90 mm)
Band 4 (red; 0.63 – 0.69 mm) Band 5 (near-infrared; 0.69 – 0.76 mm) Band 6 (near-infrared; 0.76 – 0.90 mm)
Band 7 (mid-infrared; 1.55 – 1.75 mm) Band 8 (mid-infrared; 2.08 – 2.35 mm) Band 9 (thermal-infrared; 10.4 – 12.5 mm)
Band 7 (mid-infrared; 1.55 – 1.75 mm) Band 8 (mid-infrared; 2.08 – 2.35 mm) Band 9 (thermal-infrared; 10.4 – 12.5 mm)

Calibrated Airborne Multispectral Scanner Data of
Murrells Inlet, S.C. Obtained on August 2, 1997
Natural Color Composite
(Bands 3,2,1 = RGB)
Natural Color Composite
(Bands 3,2,1 = RGB)
Masked and Contrast
Stretched Color
Masked and Contrast
Stretched Color
Composite
Composite

Color Infrared Composite
(Bands 3,2,1 = RGB)
Color Infrared Composite
(Bands 3,2,1 = RGB)
Masked and Contrast
Stretched Color
Masked and Contrast
Stretched Color
Composite
Composite

IInn SSiitittutuu CCeeppttotoommeetteteerr L Leeaaff--AArreeaa--IInnddeexx M Meeaassuurreemmeenntt
•• LAI may be computed using a Decagon Accupar Ceptometer™ that consists of
a linear array of 80 adjacent 1 cm2 photosynthetically active radiation (PAR)
sensors along a bar.
•• LAI may be computed using a Decagon Accupar Ceptometer™ that consists of
a linear array of 80 adjacent 1 cm2 photosynthetically active radiation (PAR)
sensors along a bar.
•• Incident sunlight above the canopy, Qa, and the amount of direct solar energy
incident to the ceptometer, Qb, when it was laid at the bottom of the canopy
directly on the mud is used to compute LAI.
•• Incident sunlight above the canopy, Qa, and the amount of direct solar energy
incident to the ceptometer, Qb, when it was laid at the bottom of the canopy
directly on the mud is used to compute LAI.

IInInn S Sititiututu C Ceepptototommeeteteterr L Leeaaff--A-AArrereeaa--I-IInInnddeexx M Meeaassuurrereemmeenntt

Relationship Between
Calibrated Airborne
Relationship Between
Calibrated Airborne
(CAMS) Band 6 Brightness
(CAMS) Band 6 Brightness
Values and in situ
Values and in situ
Measurements of Spartina
alterniflora Total Dry
Measurements of Spartina
alterniflora Total Dry
Biomass (g/m2) at
Biomass (g/m2) at
27 Locations in Murrells
August 2 and 3, 1997
27 Locations in Murrells
August 2 and 3, 1997

NASA Calibrated
Airborne Multispectral
Scanner Imagery
(3 x 3 m) and Derived
Biomass Map of a
Portion of Murrells
Inlet, South Carolina
on August 2, 1997 CAMS Bands 1,2,3 (RGB) CAMS Bands 6,4,2 (RGB)
NASA Calibrated
Airborne Multispectral
Scanner Imagery
(3 x 3 m) and Derived
Biomass Map of a
Portion of Murrells
Inlet, South Carolina
on August 2, 1997
CAMS Bands 1,2,3 (RGB) CAMS Bands 6,4,2 (RGB)
TM Bands 5,3,2 (RGB)
Biomass in a Portion of Murrells
Inlet, SC Derived from 3 x 3 m
Calibrated Airborne Multispectral
Scanner (CAMS) Data Obtained on
August 2, 1997
Total Biomass (grams/m 2)
500 - 749
750 - 999
1000 - 1499
1500 - 1999
2000 - 2499
2500 - 2999
TM Bands 5,3,2 (RGB)
Biomass in a Portion of Murrells
Inlet, SC Derived from 3 x 3 m
Calibrated Airborne Multispectral
Scanner (CAMS) Data Obtained on
August 2, 1997
Total Biomass (grams/m 2)
500 - 749
750 - 999
1000 - 1499
1500 - 1999
2000 - 2499
2500 - 2999

Total Above-ground Biomass
Total Above-ground Biomass
in Murrells Inlet, S. C.
Extracted from Calibrated
Airborne Multispectral Scanner
in Murrells Inlet, S. C.
Extracted from Calibrated
Airborne Multispectral Scanner
Data on August 2, 1997
Data on August 2, 1997
Total Biomass (grams/m2)
500 - 749
750 - 999
1000 - 1499
1500 - 1999
2000 - 2499
2500 - 2999

Remote Remote SSeennssiinngg UUnniivvaarriiaattee SSttaattiissttiiccss -- VVaarriiaannccee
The variance of a sample is the average squared deviation of all
possible observations from the sample mean. The variance of a band
of imagery, vark, is computed using the equation:
The variance of a sample is the average squared deviation of all
possible observations from the sample mean. The variance of a band
of imagery, vark, is computed using the equation:
( BV
-
m
)
ik k
n
n
å=
= i
1
var
k
2
The numerator of the expression is the corrected sum of squares (SS).
If the sample mean (mk) were actually the population mean, this would
be an accurate measurement of the variance. The standard deviation is
the positive square root of the variance.
The numerator of the expression is the corrected sum of squares (SS).
If the sample mean (mk) were actually the population mean, this would
be an accurate measurement of the variance. The standard deviation is
the positive square root of the variance.

Texture
Texture
Transformation
Transformation

Second-Order Second-Order SSttaattiissttiiccss iinn tthhee SSppaattiiaall DDoommaaiinn

Second-Order Statistics Second-Order Statistics iinn tthhee SSppaattiiaall DDoommaaiinn
Original image =
0 1 1 2 3
0 0 2 3 3
0 1 2 2 3
1 2 3 2 2
2 2 3 3 2
Original image =
0 1 1 2 3
0 0 2 3 3
0 1 2 2 3
1 2 3 2 2
2 2 3 3 2
h c =
0 1 2 3
0 1 2 1 0
1 0 1 3 0
2 0 0 3 5
3 0 0 2 2
h c =
0 1 2 3
0 1 2 1 0
1 0 1 3 0
2 0 0 3 5
3 0 0 2 2
SSppaatitaial lD Deeppeennddeennccyy M Maatrtrixix

Second-Order Statistics in the Spatial Domain
Second-Order Statistics in the Spatial Domain
The Angular Second Moment (ASM):
The Angular Second Moment (ASM):
( )
2
å å ,
= =
ASM h i j
0 0
=
k quantk
j
quant
i
where,
quantk = quantization level of band k (e.g., 28 = 0 to 255)
hc(i, j) = the (i, j)th entry in one of the angular brightness
value spatial-dependency matrices,
where,
quantk = quantization level of band k (e.g., 28 = 0 to 255)
hc(i, j) = the (i, j)th entry in one of the angular brightness
value spatial-dependency matrices,

TTeexxttuurree MMeeaassuurreemmeenntt

MMooiissttuurree VVeeggeettaattiioonn IInnddeexx
Rock et al (199) utilized a Moisture Stress Index (MSI):
Rock et al (199) utilized a Moisture Stress Index (MSI):
MSI = MidIRTM5
NIRTM4
based on the Landsat Thematic Mapper near-ifnrared and middle-infrared bands
based on the Landsat Thematic Mapper near-ifnrared and middle-infrared bands

Image enhancement and interpretation

More Related Content

What's hot (20)

Similar to Image enhancement and interpretation (20)

More from DocumentStory (20)

Image enhancement and interpretation