How to use this type of data
-
I supose I could export .dxf to collada and try it that way....
-
@chris fullmer said:
What does it export? If it will take any other sort of standard file type - csv, obj, anything?, maybe we can write an exporter for it.
Chris
This is website I wanted to get Lidar data from .... http://www.stanfords.co.uk/Business-Mapping/BM/Products/LiDAR_BM-LIDAR.htm -
Any text format file [even DXF] can be parsed and the points extracted in order.
If we know that that say every three of them make a triangular facet we can use that to add geometry.
It'd be a pain in the proverbial to parse a DXF file [it can be done! e.g. I extract 'text' in my dxf-text-importer tool] but other types like OBJ, CSV etc can be readily parsed and used... [DAE doesn't seem to support raw 'points', just facets]
-
Sepo, I have not used it yet but I have installed the tgi3d plugin. I suspect it would handle the stereo pairs quite nicely. This is exactly what their PhotoScan product is supposed to do. Basically you take two or more photos of a surface or surfaces from widely separated areas. Then you mark corresponding points in the photo sets and the software uses these control points to reverse engineer the 3D shape of the surface.
I have also figured out a LIDAR like system of establishing surface control points on small stationary object that does not have strong features to compare photographically. I will post my experiments as time allows. Will be heading to Thailand for a month so it might be a while before I can test the system.
-
@roger said:
Sepo, I have not used it yet but I have installed the tgi3d plugin. I suspect it would handle the stereo pairs quite nicely. This is exactly what their PhotoScan product is supposed to do. Basically you take two or more photos of a surface or surfaces from widely separated areas. Then you mark corresponding points in the photo sets and the software uses these control points to reverse engineer the 3D shape of the surface.
I have also figured out a LIDAR like system of establishing surface control points on small stationary object that does not have strong features to compare photographically. I will post my experiments as time allows. Will be heading to Thailand for a month so it might be a while before I can test the system.
That would be quite interesting. Have you actually tried PhotoScan with tgi3d? I was wondering which other software people would use to create models from 2 photos shot with time delay.
-
@tig said:
Any text format file [even DXF] can be parsed and the points extracted in order.
If we know that that say every three of them make a triangular facet we can use that to add geometry.
It'd be a pain in the proverbial to parse a DXF file [it can be done! e.g. I extract 'text' in my dxf-text-importer tool] but other types like OBJ, CSV etc can be readily parsed and used... [DAE doesn't seem to support raw 'points', just facets]
OK so no joy there as well.
-
@unknownuser said:
That would be quite interesting. Have you actually tried PhotoScan with tgi3d? I was wondering which other software people would use to create models from 2 photos shot with time delay.
Will jump on the learning curve today. And time delay is not strictly necessary if two or more synched cameras are used. I plan to try multiple linked cameras at 4,000th of a second to freeze an athlete in motion.
-
Seriously Roger, I'm going to down to AZ and follow you around for a few days. You always have some cool camera project cooking in the background.
EDIT:
ok, I imported the example dxf and see the point cloud. Very interesting. What are you expecting differently than you achieved previously? The triangulated TIN looks pretty good to me.
-
@chris fullmer said:
EDIT:
ok, I imported the example dxf and see the point cloud. Very interesting. What are you expecting differently than you achieved previously? The triangulated TIN looks pretty good to me.
Well the TIN is unusable as the buildings (in my case) look like melting cakes.
-
The sample TIN points cloud dxf imports easily into a SKP and triangulates into a mesh [even with 40k+ points and some patience!] BUT sepo's valid issue is that while it's fine for getting data on rolling land-surfaces, if there are buildings in it they will mesh into a melted-wedding-cake form [see his image] because 'vertical' surfaces won't always come out coplanar, as some triangles will spring from other corners of the adjacent ground surface's triangles...
-
@chris fullmer said:
Seriously Roger, I'm going to down to AZ and follow you around for a few days. You always have some cool camera project cooking in the background.
That would be easy, but a bit boring as I am moving slower and slower
Advertisement