[WIP] Braselton (Road Atlanta) LIDAR

Status
Not open for further replies.
Mario Morais

Mario Morais

Registered
Hi all

I started this project as a test for the development of my scripts to the max to handle point clouds.

I gathered LIDAR data circuit "road Atlanta" and allows an accuracy of 1 meter that very good data for free. After cutting the cloud I import and get 63.787 points (x, y, z) and intensity. Intensity i can use later for vertex color, alpha.

I need photos of the track with good quality.
If anyone can help I would appreciate.


road_cloud.jpg


roadat.jpg
 
Last edited by a moderator:
Imho you're going to complicate your life a bit too much. Using raw lidars pointcloud formats for such low resolution could be a waste of time when (and if) you can get GRID ASCII data or DEMs at the same resolution, but already converted into a more usable format.

In both cases you've to create your own mesh, using the terrain data, but if you get GRIDS or DEMs you can, at least, use those data as a terrain mesh, working with different exporting resolutions to avoid useless quad density.

Maybe I miss your real intention for using pointclouds...:)
 
Hi Tuttle

My point cloud is for everything.
when I import the points i create a mesh without faces. later i used for terrain walls. etc.

I'm developing scripts as needed

GRIDS and DEMs have walls? bridges? buildings? color?
 
Last edited by a moderator:
I use others scripts .

I import the track as kml path, images tiles (5 images 4096x4096) and point cloud, everything in the correct position automatically.
 
Mario Morais said:
Hi Tuttle

My point cloud is for everything.
when I import the points will not convert mesh is a mesh without faces. I'm developing scripts as needed

GRIDS and DEMs have walls? bridges? buildings? color?
Click to expand...

Of course lidars are not mesh, but points...and at the same time GRIDs are not meshes but formats to transport raster to model...

And yes, DEM can transport buildings and structures if you're working with surfaces instead of terrain.

What do you mean with color? If you've color sensitive data from scanning you can use apart from the mesh...but the pointcloud management will be much more expensive instead a format derived from the same lidar data. Of course if you can't get another format you've to work starting with this...but usually if you have the lidar you can access other formats, like GRIDs or finalized DEMs.
 
Mario Morais said:
I use others scripts .

I import the track as kml path, images tiles (5 images 4096x4096) and point cloud, everything in the correct position automatically.
Click to expand...

Yeah I know. Every time you work into GIS environment you'll get geo references for everything is Geo-Tagged...but remember there are errors in the ortho image positioning. Do not take everything like perfect science, because aerial data (especially for free contents) needs a lot of post-correction and you can get a incorrect synching between a point and a image pixel (we can get errors for meters). BTW, those data are pretty nice to build a track..but I was just thinking about the pointcloud instead other formats more suitable for a mesh management...
 
I never handle DEM files. But what you say i can get 1m resolution?
thank you for your information and you know where i find DEM file in EUROPE?
 
Elevation profiles can handle centimeters...not just meters. :)

The problem is get free data with nice resolutions, the same for pointclouds.

About DEM in Europe there are a lot of sources, depends which area you're searching for to find the proper database and dataset.

Usually free DEM/ASCII and Lidar resources are managed by local owners (and respective public databases), and professional contents by companies.
 
To be honest (and if I understood it correctly) 1m accuracy is very very low one. I'm not sure it is enough even to draw track shape, I'm not talking about curbs and other objects. google maps gives better resolution imo. Check this - google maps view on Brno circuit
 
Regarding to elevations, for one who doesn't have any better reference other than pictures and videos, 1 meter accuracy is an excelent start.
 
1m of grid density is not the same as 1m accuracy. Google maps will not give you any decent elevation data.

Combining the two is what you need and as you can see on screenshots in first post, that's what has been done.
 
Maybe is better to hold this track..

I need photos of the track with good quality.
If anyone can help I would appreciate.
Click to expand...

The original LIDAR have more points but i reduce to import to max. Is the better i can.

The track is ready but i don't have photos for textures and objects modeling. only sat images.

I go for Laguna seca, at least I have the streeview. :-)
 
You don't need textures to begin a track loft! :) I'd love to see a good version of Road Atlanta. Brilliant little circuit!
 
K Szczech said:
1m of grid density is not the same as 1m accuracy. Google maps will not give you any decent elevation data.

Combining the two is what you need and as you can see on screenshots in first post, that's what has been done.
Click to expand...

Yeah, it was a bad word choice.

Anyway, Google Earth elevation I don't even count, is impossible to trust. But what I meant is that the LiDAR data from the site Mario posted is kinda excelent starting point for setting the track elevations. Better then nothing, right?
 
Fabio Pittol said:
Yeah, it was a bad word choice.

Anyway, Google Earth elevation I don't even count, is impossible to trust. But what I meant is that the LiDAR data from the site Mario posted is kinda excelent starting point for setting the track elevations. Better then nothing, right?
Click to expand...

You right..
 
the question of geometry and accuracy is no longer in dispute. I have an optimum resolution with what I have.

The only thing I ask to the community is good quality photos.

The ideal was someone who has gone to the circuit and get photos.

I do not need critics want is photos! someone who says "Do not worry I can go to the circuit and ask to take pictures with detail" :-)
 
Maps are just aerial orthoimages, Lidars are terrain/surface scanning with laser pulses...so two completely different stuff. :)

BTW....1 meter for vertical accuracy is a awful performance, but I think Mario was talking about the Horizontal asset, not vertical.

I don't know this specific dataset but usually modern airborn surveying are pretty efficient for vertical resolutions, even when used for standards applications. Usually they can go for 12cm for FVA (fundamental vertical accuracy) or 35cm for consolidated and supplemental vertical accuracy... Pretty nice stuff for free. :)
 
Always remember though that 12cm is actually 24cm I think? If you are 12cm above the real surface, then 10ft further you are 12cm below, that is a 24cm difference, which can be pretty big.
 
Status
Not open for further replies.
Back
Top