I am not looking for any kind of computationally insentive operations (for example as using 3D buffer onall points to get bigger points). Is there a direct hook into FME's rendering engine with some kind of transformers or other memory efficient tricks?
Thanks!
As far as i known, and i tried looking for same, there is none.
Renderengine used is a pretty basic thing.
U are stuck wiht exporting to a format and using a 3d program with more capable renderengine (or just a render engine, like brt or something. Or povray etc.)
Gio
Could you expand on what render engine you have used to solve this type of problem? I dont think using a raytracing engine such as povray would work well, since this would require converting all the points to spheres (or something with volume), which is a) memory intensive and b) points farther away would appread smaller or invisible.
Thanks for any leads!
Ther is simply no way to control transparency or "size" of points in de the datainspector.
Points are representation of a spatiallocation, they do not have size (i'll take a wild guess and say "you know this, for sure").
What you can do tho, is use a PointCloudOnRasterComponentSetter (seems to be the transformer with the longest name..)
. Looks pretty nice if u got a pc with good denisty.
All you need to do is align some raster with the pc. You could make any raste in any color.
You still cant control the size of the pints though. As there is no size, ther is no tranparency possible.
PC's are usualy large, so factor procestime is alwyas an issue.
This image is an extraction, using tiling and parallelprocessing, from 270501199 + 381808275 points
I control the proces with 12 published parameters.
At the end an ecw is projected on the resulting cloud (also tiled).
Ssynchronised tiling (fascilitates the use of paralellprocessing, all my cores went up!).
FME Session Duration: 15 minutes 2.3 seconds. (CPU: 377.8s user, 37.5s system)
END - ProcessID: 3748, peak process memory usage: 2973196 kB, current process memory usage: 172084 kB
Translation was SUCCESSFUL
Leiden station, Holland
Gio