| Re: Feature identification and measurement [message #73348] |
Tue, 02 November 2010 01:03 |
rogass
Messages: 200
Registered: April 2008
|
Senior Member |
|
|
On 1 Nov., 20:21, wlandsman <wlands...@gmail.com> wrote:
> I am working on testing an optical system in which several features
> (circles, crosses, lines, a "ladder") are imaged onto a CCD. We
> use these fiducials to monitor changes in the optical path (each
> fiducial can shift independently), and so are interested in measuring
> their relative positions.
>
> My first thought was that this might be a case to take advantage of
> all those IDL remote sensing tools that astronomers mostly ignore.
> I started to read about shapefiles, but there seemed to be a lot of
> overhead just to get started.
>
> Right now I am using the Canny function to detect edges, and then
> LABEL_REGION to isolate the different features, and fitting each
> fiducial separately (eg. the lines with LINFIT(), the circles with
> mpfitellipse ) to get their position. But Canny + Label_region
> doesn't work perfectly -- either the fiducial is not entirely within
> one LABEL_REGION value, or it includes pixels that I don't want to
> measure. So I spend a lot of time interactively changing pixel
> values or selecting points to fit. I'd welcome any advice on tools
> that might help with these measurements. Thanks, --Wayne
Hi,
maybe you can provide a test image?
Cheers
CR
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
