[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: extrapolation of surface

Subject: Re: extrapolation of surface
From: "Dick Jackson" <dick(at)d-jackson.com>
Date: Thu, 02 Aug 2001 04:17:25 GMT
Newsgroups: comp.lang.idl-pvwave
Organization: Shaw Communications
References: <Pine.HPX.4.10.10108011734400.24623-100000@stud4.tuwien.ac.at>
Reply-To: "Dick Jackson" <dick(at)d-jackson.com>
Xref: news.doit.wisc.edu comp.lang.idl-pvwave:25969

Hi Alexander,

"Alexander Rauscher" <e9121161@stud4.tuwien.ac.at> wrote:
>
> i have a function f(x,y) which is f(x,y) != 0 on a more or less circular
> region which lies within an e.g. 128x128 array and 0 outside this region.
> now i want to extrapolate this function (over a range of e.g 10 pixles)
> into the region outside this circular region without the result being
> influenced by all the f(x,y)=0.
> sfit also takes into account where f(x,y)=0, which makes the fitted
> surface fit(x,y) decrease towards the edge of my array which is not what i
> want. (knowing from the shape of the original f(x,y) that it should
> increase...)

A few months ago, I came up against a similar problem and thought: "SFIT is
nice, but what if I want to fit a set of points that aren't on a full,
regular, X-Y grid". So I massaged the matrix algebra magic in the SFIT code
(which I still don't fully understand!) to work on an arbitrary set of XYZ
points.

Then you need a way to use the polynomial to get Z values from X, Y and the
KX parameters, so I'm throwing in an even less-fully-documented function to
do that.

An example program to test this is attached as well, it shows how you'd use
it if you really *had* an XY grid full of data (easier to see the result). I
wish I had more time to make this pretty or to apply it to your exact
problem, but I have a pressing deadline to meet. Let me know if you need a
hand with applying it to your problem.

I hope this is of some help, but you never know, there may be enhancements
to fitting features like this in upcoming releases of IDL...

--
-Dick

Dick Jackson                   /            dick@d-jackson.com
D-Jackson Software Consulting /       http://www.d-jackson.com
Calgary, Alberta, Canada     / +1-403-242-7398 / Fax: 241-7392


begin 666 SFitXYZ.pro
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E<FTH:W@L(&XR*2 C('5T#0I2971U<FXL(&9I= T*#0I%3D0-"@``
`
end

begin 666 Polynom2DFunction.pro
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9#0I2971U<FXL(')E<W5L= T*#0I%3D0-"@``
`
end

begin 666 SFitXYZTest.pro
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3=7)F86-E+"!Z1FET#0H-"D5.1 ``
`
end

References:
- extrapolation of surface
  - From: Alexander Rauscher

Prev by Date: Re: Editor feature?
Next by Date: Re: Question about using database like access or dbase?
Prev by thread: extrapolation of surface
Next by thread: Color problems in IDL
Index(es):
- Date
- Thread