Wednesday, February 18, 2009

Data and stuff for a script test

well, I've been asked for some data to test .. so here is what I've got on this particular small test project I was working on with that 16bit grayscale script:

first off, there are a few programs worth playing with with this sort of data(LIDAR), its been a while but here's what I think I was using to generate/view/translate the original point data that I received..

Depending on how many trees are in your data, you may have to do so other pre-processing on your point data to get as good a bare earth set of points as possible - that's done with another program .. I've been working with some desert point data so I don't have to worry about that with this data set...

quikgrid; to process the bare earth point data into a surfer .grd - This can take a LONG time if you generate a very detailed grid, and you will have to fiddle with settings depending on your point density/trees, so on sparser point data expect to have to regenerate several times before you get a decent .grd - so set aside a weekend and/or work on small segments at a time - I didn't, even though in stage miles this is a very short bit of road, It still took ages for the 20cm and 30cm tests - Patience, and a Fast computer will both do you good:

landserf 2.3 - now you load that .grd so you can generate the shaded releif you will later use as a background in BTB so you can see where to place track and terrain nodes, also take the opportunity to transfer to a format thats easily readable by scripts and other GIS software .. I've been using arc.asc AKA 'ArcGIS text raster' in the landserf save box

Landserf 2.3 dl:


now of course is the BTB work, load up that shaded relief as a background, remember your grid corners from quickgrid, and edit the .xml of your BTB save to get the background in the right place..

lay down your track, cross-sections and terrain nodes at appropriate locations on the shaded relief ..

and hope that we can get a script to work!

I would also like such a script to affect surface cross-sections, as this would save much time, cause the generated grids are not perfect, since so much of it is interpolated, it would give a massive head start on the inevitable manual fine tuning.


Images and data to follow...

First off, here is a 3d view of a grid I generated from LIDAR, this is the data that we are looking at today.

the location is a road just off of a major fault line in California, heading up into the mountains.

Cloudland Truck Trail:

note that I couldn't get the view level in the 3d view, its not that steep =)

here is a shaded relief of the whole dataset just to give you a idea of the area we are looking at:

OK, got the zip of the shaded relief, the BTB test track to be modified, and the .asc gridfile, 20cm resolution uploaded:

DEM, Track, Background

if anyone can get a script working, I'll regenerate the DEM to try and minimize the point data artifacts ... (the graining)

And here is the point data that I generated this DEM from:

Point Data 6018

No comments:

Post a Comment