Scribd Logo
Upload

Welcome to Scribd!

  • 31 views

    Summary of Fractal Exponent Calculation

    Uploaded by

    Leomar Acosta Ballesteros
    This document summarizes three methods for calculating the fractal dimension of objects: 1. For self-similar objects made of copies of itself scaled by a factor S, the fractal dimension D is calculated using log(N)/log(1-S), where N is the number of copies. 2. The divider method uses two measurements of the number of divisions N needed at two scale sizes δ1 and δ2 to calculate D using log(N1/N2)/log(δ2/δ1). 3. The box-counting method similarly uses two measurements of the number of boxes N needed to cover the object at two box sizes δ1 and δ2 to calculate D

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Summary of Fractal Exponent Calculation

    Uploaded by

    Leomar Acosta Ballesteros
    31 views4 pages
    This document summarizes three methods for calculating the fractal dimension of objects: 1. For self-similar objects made of copies of itself scaled by a factor S, the fractal dimension D is calculated using log(N)/log(1-S), where N is the number of copies. 2. The divider method uses two measurements of the number of divisions N needed at two scale sizes δ1 and δ2 to calculate D using log(N1/N2)/log(δ2/δ1). 3. The box-counting method similarly uses two measurements of the number of boxes N needed to cover the object at two box sizes δ1 and δ2 to calculate D

    Original Title

    Fractal Summary

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes three methods for calculating the fractal dimension of objects: 1. For self-similar objects made of copies of itself scaled by a factor S, the fractal dimension D is calculated using log(N)/log(1-S), where N is the number of copies. 2. The divider method uses two measurements of the number of divisions N needed at two scale sizes δ1 and δ2 to calculate D using log(N1/N2)/log(δ2/δ1). 3. The box-counting method similarly uses two measurements of the number of boxes N needed to cover the object at two box sizes δ1 and δ2 to calculate D

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    31 views4 pages

    Summary of Fractal Exponent Calculation

    Uploaded by

    Leomar Acosta Ballesteros
    This document summarizes three methods for calculating the fractal dimension of objects: 1. For self-similar objects made of copies of itself scaled by a factor S, the fractal dimension D is calculated using log(N)/log(1-S), where N is the number of copies. 2. The divider method uses two measurements of the number of divisions N needed at two scale sizes δ1 and δ2 to calculate D using log(N1/N2)/log(δ2/δ1). 3. The box-counting method similarly uses two measurements of the number of boxes N needed to cover the object at two box sizes δ1 and δ2 to calculate D

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 4

    1

    Summary of Fractal Exponent Calculation




    I. If an object is made up of N identical copies of itself, shrunk by a factor of S, then the
    fractal dimension D of the object is:

    =
    Iog N
    Log (1S)
    (1)

    The base of the logarithm can be 10 (common log), e (natural log), or whatever you want. For
    the Sierpinski triangle, N = 3 and S = . So D = 1.58496


    II. Divider method of computing the fractal dimension D of a coastline, or other line or
    boundary. Assume we have a ruler of length and that we can cover the entire curve by a
    number N
    6
    of ruler lengths. For a smooth curve (not fractal), we expect that the length of the
    line L is a constant when we take the limit as 0:

    lim
    6-0
    o N
    6
    = I(o) = constant (2)

    For a fractal curve, which is not smooth in the sense of having unique slope at each point, this
    doesnt work. As the ruler length becomes smaller and smaller, we typically find that

    lim
    6-0
    o N
    6
    = I(o) (3)

    So the answer here is that we postulate or assume a relationship involving a fractal
    dimension D such that

    lim
    6-0
    o

    N
    6
    = B = constant (4)

    Then to find the fractal dimension of the line, we measure its length o N
    6
    I at 2 different
    length scales, o
    1
    and o
    2
    . We then use equation (4):

    o
    1

    N
    6
    1
    = B = o
    2

    N
    6
    2
    (5)

    Rearranging terms, we have:

    (
    6
    2
    6
    1
    )

    =
    N
    6
    1
    N
    6
    2
    (6)

    To solve this for D, we need to use some properties of logarithms.


    Brief digression into logarithms
    2


    Lets look at some properties of logarithms. If y = x

    then:

    x = 1u
    Iog
    10
    x
    (7)

    So y = x

    becomes:

    y = x

    = (1u
    Iog
    10
    x
    )

    = 1u
    Iog
    10
    x
    (8)

    Note that we could also have written (8) as:

    y = x

    = (c
    Iog
    c
    x
    )

    = c
    Iog
    c
    x
    (9)

    using the base of natural logarithms e. Alternately, we could write (9) in slightly different
    notation as:

    y = x

    = (c
    Inx
    )

    = c
    Inx
    (10)

    We can solve (8) as:

    log
    10
    y = log
    10
    x or =
    Iog
    10

    Iog
    10
    x
    (11)

    We can solve (9) as:

    log
    c
    y = log
    c
    x or =
    Iog
    c

    Iog
    c
    x
    (12)

    and we can solve (10) as:

    lny = lnx or =
    In
    Inx
    (13)


    Lets now return to solving for D in equation (6).


    Since (
    6
    2
    6
    1
    )

    =
    N
    6
    1
    N
    6
    2
    , we can use, for example, (11):


    3

    =
    Iog
    10
    (
    N
    6
    1
    N
    6
    2
    )
    Iog
    10
    (
    6
    2
    6
    1
    )
    (14)
    Alternatively, we could use equation (12):


    =
    Iog
    c
    (
    N
    6
    1
    N
    6
    2
    )
    Iog
    c
    (
    6
    2
    6
    1
    )
    (15)

    So it really doesnt matter which base of the logarithm we use, it could be base 10 or base e or
    for that matter any other base.



    III. Box covering method of computing fractal dimension of a curve or shape. This is
    basically the same as the divider method, but is often easier to implement. Here what you do is
    to cover the shape with boxes of side length . The number of such boxes is N
    6
    . That is, N
    6

    represents the number of boxes in which any part of the curve appears. We then use two
    different box sizes o
    1
    and o
    2
    and find 2 different numbers N
    6
    1
    and N
    6
    2
    . To find D we
    then use (15).

    A final note. If you apply these methods to compute the fractal dimension of a normal smooth
    line or shape, you will find that the fractal dimension D that you compute is the normal
    embedding dimension d of the space the space in which it appears. For example, the
    embedding dimension of the space in which the Sierpinski triangle is embedded is d = 2. So if
    you use these methods to compute the fractal dimension of an ordinary right triangle, you will
    find D = d = 2.














    4

    You might also like