| Re: Line index [message #6094 is a reply to message #5990] |
Mon, 25 March 1996 00:00  |
Robert Moss
Messages: 74
Registered: February 1996
|
Member |
|
|
I may have missed your initial question, but if your looking for the
array indices along a line in an image, the following snippet of
code seems to work fine. The two endpoints are (x,y) and (x1,y1), and
the image is in the variable pimage.
dx = float(x1-x) ;delta x
dy = float(y1-y) ;delta y
n = sqrt( dx^2 + dy^2 ) ;length
sy =(y1-y)/n
sx =(x1-x)/n
;
xx = lindgen(n+1)*sx+x ;X indices, make into longwords.
yy = lindgen(n+1)*sy+y ;Y indices
sz = size( pimage ) ; image size
if !order ne 0 then yy=sz(2)-1-yy ;reverse y indices?
ans=pimage(xx,yy) ; image data along the line
You may want to go some fiddling with the (n+1) bit depending on how
you want to handle fractional pixels.
By the way, this is what I use when I want a "profile" of an image, i.e
I want the image data values along the line connecting two points. The
length reported here is equal to the length of the line (within pixel
resolution, your milage may vary).
Note that this is different than the IDL supplied library funtction
called "profile" which actually returns a list of values that are the
length of the x or y projection of the line connecting two points,
depending on which is longer. I did not find the builtin "profile"
results acceptable for my purposes.
Robet M. Moss, Ph.D.
Texaco Inc.
mossrm@texaco.com
-----------
This may not reflect the views of Texaco Inc.
-----------
David Foster wrote:
>
> pit@asterix.kis.uni-freiburg.de (Peter Suetterlin) wrote:
>>
>> In article <31499FB9.2F6D@dlr.de>,
>> Hermann Mannstein <H.Mannstein@dlr.de> writes:
>>
>>> is there a robust function which returns the indices within an image
>>> (like the where function) belonging to a line defined by two or more
>>> points.
>>
>> for shure there is, even built-in IDL:
>> Index=polyfillv(xx,yy,xsize,ysize)
>> Read the manpage for further reference.
>>
>
> According to the "manpage", POLLYFILLV() expects at least 3
> points so it would never work for a simple line. Also, it
> connects each successive line and the first with the last,
> to make a polygon, and then returns the indices of the
> points within this polygon.
>
> You wouldn't be getting the indices of the points on the line(s),
> but of the points within the polygon defined by the points.
>
> Dave Foster
> foster@bial1.ucsd.edu
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
