General question . . .

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

General question . . .

blammo

All,

 

I've been a lurker on this list for a few months now, steeling myself for what's to come.  I'm going to be receiving a large point cloud dataset very soon covering a rather large area (for me/us).  So, I'm asking for some ideas about how to handle . . . .

 

The specifics of the collection:

·         76+ square miles (Maybe triple this size if I can find the funding to process a neighboring collection as well)

·         8pts per sq. meter (~8in post spacing?, using a simple 2D grid at 8in spacing starts to approach the 5 billion point mark)

·         The collection utilized a  side looking LIDAR rig (will include shots up the sides of buildings, hopefully enough to produce good building volumes)

·         Two separate datasets, bare earth (buildings removed) and the buildings.

·         There will be additional collects for identified urban canyons, etc.

 

Needs:

·         Data storage and distribution ideas.

·         Processing of the data, on the fly (feasible?) and/or as derived datasets (Contours, Slopes, drainage basins, 3d structure, etc.)

·         Over the Web distribution is the target end use.

o   For AutoCAD users, RAW data, AutoCAD Civil 3D Features, general 3D constructs, Derived datasets, including contours etc.

o   For Use in a Web based viewer (to be defined) and integrated with our 2D data distribution.

§  Our existing custom viewer (2d only):  http://gis.ci.stpaul.mn.us/gis/gismo_public/html/

§  This is all available as WMS and WFS where possible.

·         Hopefully some sort of standardized format (something OGC like, similar to WMS/WFS, etc.)

 

This is all somewhat new to me from the LIDAR perspective, especially in the data management side of the equation.  I have some ideas about automating the processing from a RAW LIDAR data store on the fly from Web calls.  Some sort of 3D gridding/tiling, probably with generalized (pyramided?) layers for simplification and Web distribution).  There are obvious derived products to be built from this type and density of data.  We'll also need some method of adding to this "seamless" data store over time with additional flights as well as street level LIDAR collections.

 

We currently have a Web based infrastructure in place that feeds our AutoCAD (90+ seats) users as well as our Web based Data visualization services.  this is mostly 2D, but we've had some experience in 3D visualizing over the web in the past with Plugins.  The new Browser capabilities are demonstrating that this is now possible without the plugins.

 

Thanks

bobb

 

 

 

 


_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: General question . . .

Radu B Rusu
Administrator
Bob,


On 06/14/2012 09:26 AM, Basques, Bob (CI-StPaul) wrote:

> All,
>
> I've been a lurker on this list for a few months now, steeling myself for what's to come. I'm going to be receiving a
> large point cloud dataset very soon covering a rather large area (for me/us). So, I'm asking for some ideas about how to
> handle . . . .
>
> The specifics of the collection:
>
> ·76+ square miles (Maybe triple this size if I can find the funding to process a neighboring collection as well)
>
> ·8pts per sq. meter (~8in post spacing?, using a simple 2D grid at 8in spacing starts to approach the 5 billion point mark)
>
> ·The collection utilized a side looking LIDAR rig (will include shots up the sides of buildings, hopefully enough to
> produce good building volumes)
>
> ·Two separate datasets, bare earth (buildings removed) and the buildings.
>
> ·There will be additional collects for identified urban canyons, etc.

Sounds like a very exciting project!

> Needs:
>
> ·Data storage and distribution ideas.
>
> ·Processing of the data, on the fly (feasible?) and/or as derived datasets (Contours, Slopes, drainage basins, 3d
> structure, etc.)
>
> ·Over the Web distribution is the target end use.


All these are sort of on the PCL roadmap for out-of-core data structures and manipulation. The topic is hot, and our
developers are doing an excellent work integrating various ideas into the new libpcl_outofcore library. With more usage
scenarios and more minds behind it, I'm sure we'll have even better solutions faster.

>
> oFor AutoCAD users, RAW data, AutoCAD Civil 3D Features, general 3D constructs, Derived datasets, including contours etc.
>
> oFor Use in a Web based viewer (to be defined) and integrated with our 2D data distribution.
>
> §Our existing custom viewer (2d only): http://gis.ci.stpaul.mn.us/gis/gismo_public/html/
>
> §This is all available as WMS and WFS where possible.
>
> ·Hopefully some sort of standardized format (something OGC like, similar to WMS/WFS, etc.)
>
> This is all somewhat new to me from the LIDAR perspective, especially in the data management side of the equation. I
> have some ideas about automating the processing from a RAW LIDAR data store on the fly from Web calls. Some sort of 3D
> gridding/tiling, probably with generalized (pyramided?) layers for simplification and Web distribution). There are
> obvious derived products to be built from this type and density of data. We'll also need some method of adding to this
> "seamless" data store over time with additional flights as well as street level LIDAR collections.

Again, very interesting requirements. How can our community help?

> We currently have a Web based infrastructure in place that feeds our AutoCAD (90+ seats) users as well as our Web based
> Data visualization services. this is mostly 2D, but we've had some experience in 3D visualizing over the web in the past
> with Plugins. The new Browser capabilities are demonstrating that this is now possible without the plugins.

Agreed, though it's still painful that webGL is not a true standard. In our web viewer code we had to resort to
different technologies like webGL, Flash, and canvas for iOS, to get everything working on multiple platforms.

Cheers,
Radu.
_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: General question . . .

clockwise9
 Bob,

These are all great questions. I'm in the same boat looking to develop a lot of what you've described. Unfortunately, we don't have solutions for all of these things... yet (you can help contribute ;) I'm one of two current developers on the out-of-core project which should hit the streets sometime in the next month.  The viewer is written with a VTK backing, but I plan update the webgl stuff, possibly incorporating megatree (cool project over at ros). 

Getting pcl clouds in 3rd party software is possible. I've wrote a plugin for maya, andll eventually hit some of the other autodesk apps. Unfortunately again, there's no base OpenGL visualization classes to do this. 

Hope this helps. If you'd like to coordinate on these things and help make pcl better all the better!

On Thursday, June 14, 2012, Radu B. Rusu wrote:
Bob,


On 06/14/2012 09:26 AM, Basques, Bob (CI-StPaul) wrote:
All,

I've been a lurker on this list for a few months now, steeling myself for what's to come. I'm going to be receiving a
large point cloud dataset very soon covering a rather large area (for me/us). So, I'm asking for some ideas about how to
handle . . . .

The specifics of the collection:

·76+ square miles (Maybe triple this size if I can find the funding to process a neighboring collection as well)

·8pts per sq. meter (~8in post spacing?, using a simple 2D grid at 8in spacing starts to approach the 5 billion point mark)

·The collection utilized a side looking LIDAR rig (will include shots up the sides of buildings, hopefully enough to
produce good building volumes)

·Two separate datasets, bare earth (buildings removed) and the buildings.

·There will be additional collects for identified urban canyons, etc.

Sounds like a very exciting project!

Needs:

·Data storage and distribution ideas.

·Processing of the data, on the fly (feasible?) and/or as derived datasets (Contours, Slopes, drainage basins, 3d
structure, etc.)

·Over the Web distribution is the target end use.


All these are sort of on the PCL roadmap for out-of-core data structures and manipulation. The topic is hot, and our developers are doing an excellent work integrating various ideas into the new libpcl_outofcore library. With more usage scenarios and more minds behind it, I'm sure we'll have even better solutions faster.


oFor AutoCAD users, RAW data, AutoCAD Civil 3D Features, general 3D constructs, Derived datasets, including contours etc.

oFor Use in a Web based viewer (to be defined) and integrated with our 2D data distribution.

§Our existing custom viewer (2d only): http://gis.ci.stpaul.mn.us/gis/gismo_public/html/

§This is all available as WMS and WFS where possible.

·Hopefully some sort of standardized format (something OGC like, similar to WMS/WFS, etc.)

This is all somewhat new to me from the LIDAR perspective, especially in the data management side of the equation. I
have some ideas about automating the processing from a RAW LIDAR data store on the fly from Web calls. Some sort of 3D
gridding/tiling, probably with generalized (pyramided?) layers for simplification and Web distribution). There are
obvious derived products to be built from this type and density of data. We'll also need some method of adding to this
"seamless" data store over time with additional flights as well as street level LIDAR collections.

Again, very interesting requirements. How can our community help?

We currently have a Web based infrastructure in place that feeds our AutoCAD (90+ seats) users as well as our Web based
Data visualization services. this is mostly 2D, but we've had some experience in 3D visualizing over the web in the past
with Plugins. The new Browser capabilities are demonstrating that this is now possible without the plugins.

Agreed, though it's still painful that webGL is not a true standard. In our web viewer code we had to resort to different technologies like webGL, Flash, and canvas for iOS, to get everything working on multiple platforms.

Cheers,
Radu.
_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users

_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users