Re: Z-buffer and plot/oplot issues [message #89775 is a reply to message #89757] |
Wed, 03 December 2014 06:49  |
Phillip Bitzer
Messages: 223 Registered: June 2006
|
Senior Member |
|
|
On Monday, December 1, 2014 12:08:40 PM UTC-6, Antonio Carlos Varela Saraiva wrote:
> Hello folks,
>
> I'm working with Google tiles generation using IDL and I'm facing a weird problem with the Z-buffer device.
>
> Background: As some of you should know, but others don't, google tiles is the way Google maps/Earth and other map servers draw maps on the screen. A series of images are posed side by side and on top of each other in order to assemble a mosaic of the desired image/data. I believe they do things that way to maximize speed. Every tile is georeferenced, usually in Mercator coords and WGS84 Datum.
>
> My problem: I'm using IDL to create tiles to show lightning data over a given region. To maximize speed I'm using 256x256 tiles and the lightning data is expressed as filled square polygons. I use very simple commands like map_set to set the map coords on my tile and plot or oplot procedures to plot the lightning data. After that I just copy the tile to a variable using tvrd and save as png.
>
> Using the X window device everything is perfect, but when I use the Z-buffer device there is an artifact on the final figure that I cannot remove, and that only happen when I plot filled symbols with plot or oplot (the same doesn't happen with plots procedure). The artifact is a blank like on top of the tile. However if I use an open square symbol, or anything "open", there is no such artifact. If I use plots there is no artifact as well, but it take forever to plot the data and I can't afford that. The problem is that some tiles share portions of the same "lightning" and when I put one figure on top of the other, there is a line cutting the filled square.
>
> I'm mostly sure that I'm setting the z-buffer correctly because some procedures work as they should and plot (and oplot) produce that blank line at the top when using special filled symbols.
>
> Any help is welcomed.
>
> Cheers,
> Antonio
Can't say I've ever seen this happen. Do you have some sample code (MWE) that illustrates the issue?
|
|
|