Untitled

Interpolation
How to estimate unknown values at specific locations

Spatial Interpolation :
– Trend surfaces
– Nearest neighbours: Thiessen(voronoi)
– Inverse distance weighting (IDW)
– Splines
– Kriging
Example:
We would like to estimate the variable value at (5,5)

Inverse Distance Weighting (IDW)
 Value of z(x) is estimated from all known values

of z at all n points.
(Weighted Moving Average technique)
and
 Weights usually add to 1 :

IDW Solution :
Geostatistics
 Techniques which are used for mapping of

surfaces from limited sample data and the
estimation of values at unsampled locations
 Geostatistics is used for:
– spatial data modelling
– characterizing the spatial variation
– spatial interpolation
– simulation
– optimization of sampling
– characterizing the uncertainty
 The idea of geostatistics is the points which are
close to each other in the space should be likely
close in values.
Geostatistics is used in:
Mining
Geography
Geology
Geophysics
Oceanography
Hydrography
Meterology
Biotechnology
Enviromental studies
Agriculture
Etc.
Geostatistical methods provide
 How to deal with the limitations of deterministic

interpolation
 The prediction of attribute values at unvisited
points is optimal
 BLUE (Best Linear Unbiased Estimate)
Geostatistical method for interpolation
 Reconigtion that the spatial variation of any

continuous attribute is often too irregular to be
modelled by a simple mathematical function.
 The variation can be described better by a
stochastic surface.
 The interpolation with geostatistics is known as
kriging
Ordinary kriging
 In ordinary kriging, a probability model is used in
which the bias and error variance can be
computed and select weights for the neighbour
sample locations that the everage error for the
model is 0 and the error variance is minimized.
 The procedure of ordinary kriging is similar to
weighted moving average except the weights are
derived from geostatistical analysis
 The estimationi by ordinary kringing can be
expressed by:
 The minimum variance of z(x0 ) is
 And it is obtained when

n
Z2
KRIGING (Spasial Functioin) = Z* = ∑ Zi . i
i=1
Z3
 = weighted factor
Z1 d2 = 12
d1 = 10 Concept :
d 3 = 14
z* • Z* – Z smallest, B L U E
d 4 = 12
Var [Z* - Z] smallest
d 5 = 18
Z4
• Var [Z* - Z] =
Var [Z*] + Var [Z] – 2.Covar [Z*,Z]
Z5
n n n
• Var [Z*] = Var [ ∑ Zi .i ] = ∑ ∑
i=1 i=1 j=1
i . j . Cov { Z i . Zj }
• Var [Z] = σ 2
n
• Covar [Z*. Z] = ∑
i=1
i . Cov { Zi . Z* }
So the deviation :
n n n
Var [Z* - Z] = σ 2 + ∑ ∑
i=1 j=1
i . j . Cov { Zi . Zj } – 2.
i=1
∑ i . Cov { Zi . Z* }
minimized deviation DIFFERENSIAL Var [Z* - Z] = 0
Help function LAGRANGE MULTIPLIER, μ

•KRIGING DEVIATION
n n n n
2 2
σK = σ + ∑ ∑
i=1j=1
i . j. Cov { Zi. Zj } – 2.i=1∑i . Cov { Zi . Z*} +2. (∑
i=1
i - 1 )
• MINIMIZED KRIGING DEVIATION

=0
(σ K)
2
n
= 2. ∑ j . Cov { Zn . Zj } – 2. Cov {Zn . Z*} + 2.  = 0
n j=1
n
∑ i = 1
i=1
(σ K2) n
= 2. ∑  - 2
 i=1
n
1. ∑ j . Cov {Zi . Zj } +  = Cov {Z . Z} with i = 1,2,3,…..,n
j=1
n n
2. ∑ i = 1 3. σk2
= σ
2
- ∑ i . Cov {Zi . Z*} +
i=1

i=1
described with matrix formula :
 (Z1 . Z1 )  (Z1 . Z 2 ) …………..  (Z1 . Z n ) 1  (Z1 . Z*)
 (Z2 . Z1 )  (Z2 . Z2 ) …………..  (Z2 . Z n ) 2  (Z2 . Z*)
…………… …………………………......…………. …. = ………….
 (Zn . Z1 )  (Zn . Z2 ) …………..  (Zn . Zn ) …. ………….

1 1 0  1
Kriging Solution :
 Computing kriging weights for the unsampled

point x = 5, y = 5.
 Let the spatial variation of the attribute sampled at
the five points be modelled by a spherical
variogram with parameters c 0 =2.5, c =7.5 and
range a = 10.
 The data at the five sampled points are:
 Value at (5,5) = weights * z = 4.3985
 With estimation variance = (weights*b)+ Ф =
4.2177 + (-0.1544) = 4.0628
 Note: The estimation error variance is also known
as kriging variance
Block kriging
 The modification of kriging equations to estimate
an average value z(B) of the variable z over a
block of area B
Example showing a regular 2x3 grid of point
locations within a block. Each discretizing point
accounts for the same area
 The average value of z(B) over the block B is
given by
is estimated by
with
 The minimum variance is
and is obtained with

Simple kriging
 It is similar to ordinary kriging except that the
weights sum equation (=1) is not added.
 The mean is a known constant.
 It uses the average of the entire data set.
(ordinary kriging uses local average :
the average of the points in the subset for a
particular interpolation point)
Cokriging
 It is an extension of ordinary kriging where two or
more variables are interdependent.
 How:
• U and V are spatial correlated
• Variable U can be used to predict variable V that
is information about spatial variation of U can
help to map V.
– Why:
• V data may be expensive to measure or
collect or have some limitations in data
collection process so the data may be
infrequent.
• U data, on the other hand, may be cheap to
measure and possible to collect more
observations.
Indicator kriging
 Binary value
 From a continuous variable z(x), an indicator can
be created by indicating it 1 for z(x) is less than
or equal to a cut-off value, zc , and 0 otherwise
Kriging: Step by step
 Studying the gathered data: data analysis

 Fitting variogram models:
experimental variogram and theoretical
variogram models
 Estimating values at those locations which have
not been sampled (kriging)
e.g. ordinary kriging, simple kriging, indicator
kriging and so on
 Examining standard error which may be used to
quantify confidence levels
 Kriging interpolation
Example
 Elevation data set in 2000 laser scanning points
The variogram calculated by
Isotropy
Semivariogram
Searching Neightborhood
Anisotropy Semivariogram
Searching Neightborhood
Kriging interpolation
References:
1. Geographic information Analysis by Sullivan, D.

And Unwin, D.
2. Geostatistics for Environmental Scientists by
Richard Webster and Margaret Oliver.
3. Principle of Geographical Information Systems,
Chapter 5 and 6 by Peter Burrough and Rachael
McDonnell.
4. An Introduction to Applied Geostatistics by
Edward Isaaks and Mohan Srivastava.
5. Quality Aspects in Spatial Data Mining by Alfred
Stein, Wenzhong Shi and Wietske Bijker.
6. Geostatistics: Kriging by Rangsima Sunila, D.Sc,
Email : [email protected]
Hp : 08174125876

Untitled

Uploaded by

Copyright:

Available Formats

Untitled

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Untitled

Uploaded by

Copyright:

Available Formats

Interpolation

How to estimate unknown values at specific locations

– Nearest neighbours: Thiessen(voronoi)

– Inverse distance weighting (IDW)

We would like to estimate the variable value at (5,5)

 Value of z(x) is estimated from all known values

 Weights usually add to 1 :

 Techniques which are used for mapping of

 How to deal with the limitations of deterministic

 Reconigtion that the spatial variation of any

 The minimum variance of z(x0 ) is

 And it is obtained when

Var [Z*] + Var [Z] – 2.Covar [Z*,Z]

minimized deviation DIFFERENSIAL Var [Z* - Z] = 0

Help function LAGRANGE MULTIPLIER, μ

• MINIMIZED KRIGING DEVIATION

described with matrix formula :

 (Z1 . Z1 )  (Z1 . Z 2 ) …………..  (Z1 . Z n ) 1  (Z1 . Z*)

 (Z2 . Z1 )  (Z2 . Z2 ) …………..  (Z2 . Z n ) 2  (Z2 . Z*)

…………… …………………………......…………. …. = ………….

 (Zn . Z1 )  (Zn . Z2 ) …………..  (Zn . Zn ) …. ………….

 Computing kriging weights for the unsampled

and is obtained with

 Studying the gathered data: data analysis

1. Geographic information Analysis by Sullivan, D.

You might also like

Var [Z] + Var [Z] – 2.Covar [Z,Z]