| Re: A Contour Tracking Problem [message #71122] |
Wed, 02 June 2010 09:19  |
penteado
Messages: 866
Registered: February 2018
|
Senior Member
Administrator |
|
|
If you know the value at some reference point and know how much it
changes on each contour step, you know the value at every point in one
of the contours. Taking these values, you could then interpolate for
the other points. One way to do it would be make a arrays with the
location and value of every contour point, and then give those to
IDL's triangulate and trigrid(), to obtain the rest.
|
|
|
|
| Re: A Contour Tracking Problem [message #71127 is a reply to message #71122] |
Wed, 02 June 2010 06:56  |
Chip Eastham
Messages: 1
Registered: June 2010
|
Junior Member |
|
|
On Jun 2, 9:04 am, jgrimm...@yahoo.com wrote:
> I would very appreciate if I could get help on this problem. It is
> mostly an imaging problem, but may involve some mathematical
> issues. Hence the crosspost. Rather than be very general, I
> will explain the actual example I am confronted with to keep
> things simpler and clearer.
>
> I have an image (digitally acquired), that represents the
> contours of an unknown function. In this particular case, the
> contours are interference fringes of a thin film and hence
> represent contours of constant film thickness. I now wish
> to get a map of the actual thickness, given that I know
> the real thickness at some reference point and I can somehow
> differentiate between going 'uphill' vs 'downhill'. This is just
> the reverse of the usual plotting problem where one *knows*
> a function z = z(x, y) and then gets a contour plot of z.
> Assume that we can process the image to the point that
> we have just black or white regions and so we can clearly
> determine when a fringe is crossed while moving along a
> particular direction.
>
> While one can keep track of contour crossings as one moves
> along a straight line, the part that I cannot get a handle on
> is how to keep track of the contours and know when one is
> back at a contour that one has already crossed. In my case,
> the contours are closed and there are multiple local maxima
> and minima to deal with.
>
> Any pointers will be appreciated.
>
> Thanks.
>
> J. Grimmond
Do you have colors of polarization to tell
when you are going "uphill" vs. "downhill"?
regards, chip
|
|
|
|
| Re: A Contour Tracking Problem [message #71169 is a reply to message #71127] |
Mon, 07 June 2010 21:20 |
jgrimmond
Messages: 4
Registered: September 2008
|
Junior Member |
|
|
On Wed, 2 Jun 2010 06:56:36 -0700 (PDT), Chip Eastham
<hardmath@gmail.com> wrote:
> On Jun 2, 9:04�am, jgrimm...@yahoo.com wrote:
>> I would very appreciate if I could get help on this problem. It is
>> mostly an imaging problem, but may involve some mathematical
>> issues. Hence the crosspost. Rather than be very general, I
>> will explain the actual example I am confronted with to keep
>> things simpler and clearer.
>>
>> I have an image (digitally acquired), that represents the
>> contours of an unknown function. In this particular case, the
>> contours are interference fringes of a thin film and hence
>> represent contours of constant film thickness. I now wish
>> to get a map of the actual thickness, given that I know
>> the real thickness at some reference point and I can somehow
>> differentiate between going 'uphill' vs 'downhill'. This is just
>> the reverse of the usual plotting problem where one *knows*
>> a function z = z(x, y) and then gets a contour plot of z.
>> Assume that we can process the image to the point that
>> we have just black or white regions and so we can clearly
>> determine when a fringe is crossed while moving along a
>> particular direction.
>>
>> While one can keep track of contour crossings as one moves
>> along a straight line, the part that I cannot get a handle on
>> is how to keep track of the contours and know when one is
>> back at a contour that one has already crossed. In my case,
>> the contours are closed and there are multiple local maxima
>> and minima to deal with.
>>
>> Any pointers will be appreciated.
>>
>> Thanks.
>>
>> J. Grimmond
>
> Do you have colors of polarization to tell
> when you are going "uphill" vs. "downhill"?
Unfortunately, no.
J. Grimmond
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
