| Re: Q:how to use MESH_OBJ [message #19591] |
Sat, 01 April 2000 00:00 |
Randall Frank
Messages: 17
Registered: October 1999
|
Junior Member |
|
|
Rick,
If I can assume that the data for your fish takes the form
of a number of crosssectional contours that you are trying to connect
end to end (similar to what one would get if you CT scanned the
fish and drew contours on each slice) and you are using IDL 5.3,
you should try the IDLanROIGroup object. Especially, the ComputeMesh
method. This method takes a series of contours (IDLanROI objects)
and triangulates a continuous surface inbetween each one. The output
surface is in the proper form for an IDLgrPolygon object. This will
help "glue" the cylinders you mention from MESH_OBJ together into a
single continuous object.
Hope it helps.
Rick Towler wrote:
>
> Hi all,
>
> First a big thanks to everyone who responded to my post a few weeks back
> on z-buffer ordering. Paraphrasing Emeril, I have been forced to kick
> my IDL programming up a notch and your comments have helped.
>
> Today I am wrestling with the MESH_OBJ routine. I am trying to generate
> a surface that represents a 3d fish. The closest example I can come up
> with is the teapot in the IDL demos except that the code to actually
> generate the teapot mesh isn't included (or I haven't found it. The
> program that displays the teapot loads a file that contains the mesh and
> polygon data.)
>
> I have an array of x,y,z points that make up the fish body. I have had
> success manipulating the data a little and using the "extrusion" surface
> type to generate a collection of thin oval cylinders that when viewed
> from the side look o.k. The problem with this approach is that these
> individual extrusions aren't connected to their neighbors so when the
> polygon normals start to point away from the eye you can start to see in
> between them. And, when viewed head or tail on, the object virtually
> disappears (the extrusions have no thickness). The problem here seems
> to be the polygon list. If I omit the polygon list from the call to
> IDLgrPolygon the resultant object will be interconnected and I will not
> have the problems describe above. But, I do introduce another problem in
> that it takes far to long to generate the IDLgrPolygon object without
> the polygon list information.
>
> Since the problem seems to be with the polygon list (the mesh points are
> fine, they just aren't connected), I tried using a polygon list
> generated from a separate call to MESH_OBJ where I chose the
> triangulated surface. When I pass the mesh data from the MESH_OBJ
> "extrusion" call and the polygon data from the MESH_OBJ "triangulated"
> to my IDLgrPolygon call I get a ton of errors (invalid connectivity list
> detected(invalid vertex reference)) and finally an object that is
> interconnected. The only problem here are the errors, and what seem
> like overly complex meshes (hard to explain).
>
> I thought that the "triangulated" surface type would be the ticket but
> taking this approach yields unexpected results. I must be missing
> something. Even if I define 6 vertices for a simple 3d diamond and pass
> that to MESH_OBJ,1 what I get when I pass that data to IDLgrPologon is
> an object that sort of looks like you took the diamond and cut along the
> edges and laid it out. This isn't exactly what it looks like, but it
> certainly isn't a closed object.
>
> What very basic thing am I missing here? Is this the wrong tool? How
> did the RSI people generate the surface mesh and ploy list for the
> teapot? Is there a way to fix the polygon list generated by the
> MESH_OBJ "extrusion" routine so all the little guys are connected?
>
> And lastly, am I just screwed and do i have to generate the polygon list
> manually?
>
> Thanks!
>
> -Rick Towler
--
rjf.
Randy Frank | ASCI Visualization
Lawrence Livermore National Laboratory | rjfrank@llnl.gov
B451 Room 2039 L-561 | Voice: (925) 423-9399
Livermore, CA 94550 | Fax: (925) 423-8704
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
