image transformation

8/10/2019 Image Transformation

1/23

Image transformation


2/23

Image Transformation

Image transformations typically involve themanipulation of multiple bands of data, whether from asingle multispectral image or from two or more imagesof the same area acquired at different times (i.e.multitemporal image data).

Either way, image transformations generate "new"images from two or more sources which highlightparticular features or properties of interest, better than

the original input images


3/23

Image Division

The most common transforms applied to

image data. On a pixel-by-pixel basis carry out thefollowing operation

Band1/Band2 = New band

resultant data are then rescaled to fill the range ofdisplay device

Very popular technique, commonly called

Band Ratio


4/23

Mathematically

BVi,j,r= BVi,j,k / BVi,j,l

Where

BVi,j,k Brightness value at the location line i,pixel j in k band of imagery

BVi,j,l Brightness value at the samelocation in band l

BVi,j,r Ratio value at the same location

(Note: If Denominator is 0 (zero) then Denominator BV is made 1)


5/23

Reasons / Application of Ratios

Undesirable effects on recorded radiances (e.g. variable

illumination) caused by variations in topography.

Sometimes differences in BVs from identical surface material

are caused by topographic slope and aspect, shadows or

seasonal changes

These conditions hamper the ability of an interpreter to correctlyidentify surface material or land use in a remotely sensed image.

Ratio transformations can be used to reduce the effects

of such environmental conditions


6/23

Ratios for Elimination of Topographic Effect

.691611Shadow

.694531Sunlit

Coniferous

.951918Shadow

.965048Sunlit

Decidous

Band BBand A

RatioDigital NumberLandcover/Illumination

Same cover type Radiance at shodow is only 50% of radiance at sunlit

Ratio nearly identical


7/23

Which bands to Ratio

1 2 3 4 5 7


8/23

Which bands to Ratio-example

Healthy vegetation reflects strongly in the near-infrared portion of thespectrum while absorbing strongly in the visible red.

Other surface types, such as soil and water, show near equalreflectances in both the near-infrared and red portions.

Thus, a ratio image of Near-Infrared (0.8 to 1.1 m) divided by Red(0.6 to 0.7 m) would result in ratios much greater than 1.0 forvegetation, and ratios around 1.0 for soil and water.

Thus the discrimination of vegetation from other surface cover types issignificantly enhanced.

Also, we may be better able to identify areas of unhealthy or stressedvegetation, which show low near-infrared reflectance, as the ratioswould be lower than for healthy green vegetation.


9/23

InfraRed Red

RATIO IR/R


10/23

Commonly used Vegetation Indices

Vegetation Index or Ratio Vegetation

Index (RVI) = IR / R Normalized Differential Vegetation Index

(NDVI) = (IR - R)/(IR + R) Transformed Vegetation Index (TVI)

= {(IR - R)/(IR + R) + 0.5}

1/2

x 100


11/23

Principal Component Analysis (PCA)

Different bands of multispectral data are

often highly correlated and thus containsimilar information.

We need to Transforms the original

satellite bands into new bands that

express the greatest amount of variance

(information) from the feature space ofthe original bands


12/23

PCA Cont..

The objective of this transformation is to reduce

the dimensionality (i.e. the number of bands) inthe data, and compress as much of the

information in the original bands into fewer

bands. The "new" bands that result from this statistical

procedure are called components.

This process attempts to maximize (statistically)

the amount of information (or variance) from the

original data into the least number of newcomponents.


13/23

Graphical Conceptualization

PCA is accomplished by a linear

transformation of variables that

corresponds to a rotation and translation

of the original coordinate system


14/23


15/23

Correlation Scatterplot

Da

ta2

Da

ta2

Data 1 Data 1

Positive CorrelationNegative Correlation

Da

ta2

Data 1

No Correlation


16/23

Conceptualization

Translate and/orroute the originalaxes so that theoriginal brightnessvalues on axes

Band 1 and Band 2are redistributed(reprojected) onto anew set of axes ordimensions, Band 1'and Band 2'.

PCA Graphical


17/23

PCA- Graphical

Conceptualization The Band 1' coordinate system is

then be rotated about its new origin(Mean1, Mean2) in the newcoordinate system some degree

so that the axis Band 1' isassociated with the maximumamount of variance in the scatter ofpoints . This new axis is called thefirst principal component (PC1 = 1).

The second principal component(PC2 = 2) is perpendicular(Orthogonal) to PC1. Thus, themajor and minor axes of theellipsoid of points in bands 1 and 2are called the principalcomponents.

The second principal componentdescribes the variance that is not

already described by the first


18/23

Conceptualization

C


19/23

Principal Component 1

The first principal

component,

broadly simulatesstandard black and

white photography

and it contain mostof the pertinent

information

inherent to ascene.


20/23


Thus as is the

conventionthe second

PC has a

smaller

variance than

the first PC


21/23


Some of the graypatterns can be

broadly correlatedwith twocombined classes

of vegetation: Thebrighter tonescome from theagricultural fields.

Moderately darkertones coincidewith some of the

grasslands, forestor tree areas.


22/23


Very Little

InformationContent

Composite PC Image


23/23

Composite PC Image

Forest appearsgreen, river bedin blue, water in

Red

orange ,vegetationappears in

varying shadesof green andfallow

agriculture fieldas pink tomagenta

Color Composite PC1,PC2,PC3

image transformation

Documents