Lidar to DXF ?

[owenransen]

I'm considering extending one of my applications, Pointor, to read LiDar files (LAS and LAZ). Pointor takes pointlists and turns them into DXF files.

But after a few experiments I've found that AutoCAD cannot really handle large collections of points, as native AutoCAD points anyway. 3D viewing is so slow I've had to stop AutoCAD running.

So I'm wondering if an application which created DXF files from lidar point data would have any realworld application?

Or maybe I should allow the user to input only a fracrion of the points from large files?

Or...?
 

[mark312]

Maybe you looking for Meshlab.  Meshlab is able to save many pointcloud formats to the "dxf".

 

http://meshlab.sourceforge.net/

[owenransen]

I've just had a look at the import file options, and neither LAS  nor LAZ nor LIDAR is there. Maybe in the future...

 

But the lack of responses seem to indicate that no-one here cares about doing LAS to DXF, which is interesting.

[Smurfer222]

Hi owenransen, I'll bite on your question, I don't think it's a matter of people not caring about being able to convert LAS to DXF, more that DXF is a CAD/drawing format and not a true remote sensing/GIS format. CAD is used more for drawings and dafting, not data creation or analysis, I mean that the DXF format and associated programs were never designed to view massive datasets all at once. I would never look at a LAS file in AutoCAD and expect it to draw quickly anymore than I would expect ArcGIS/MapInfo/ERDAS... to be capable to draw arcs, lines or polygons with the precision and survey grade accuracy of a CAD, they are designed with two seperate paradigms, one is fine scale hyper accuracy, the other is more regional...

Hope that clears things up a bit.

 

S222

[owenransen]

Hi owenransen, I'll bite on your question,...ArcGIS/MapInfo/ERDAS... to be capable to draw arcs, lines or polygons with the precision and survey grade accuracy of a CAD, they are designed with two seperate paradigms, one is fine scale hyper accuracy, the other is more regional...

Hope that clears things up a bit.

 

S222

 

I think you're right. I asked because a few of my Pointor customers have asked for this and I wondered if it could be useful. But then I saw the size of the files and quantities of points and thought I'd better ask here first.

 

Thanks for the feedback.

[ferrellsl]

There are a couple approaches for creating a DXF from LIDAR points.  As you noted earlier, TINS created from LIDAR are so large as to make AutoCAD virtually unusable.  The other option is to create a quadrilateral mesh from the points which will yield a much smaller but just as accurate model.  I've written two applications to convert LAS files to an AutoCAD DXF.  One outputs a TIN while the other outputs a mesh.  The mesh generation app can process over 50 million points in less than 3 minutes.  The TIN'ing app is slightly slower.

 

 

See my website at http://lidarwidgets.com

for more info or you an email me at:  [email protected]

Edited April 12, 2016 by ferrellsl

[miladahmad]

You can use FME to export all lidar point to DXF or within FME you can also choose the density of point you want or have a  separate files, or create 3D surface, create TIN, etc.

[ferrellsl]

I've also just completed writing an application the converts a LAS point cloud into a DXF point cloud.  See: http://lidarwidgets.com

[ferrellsl]

I've have several requests for an application that extracts the ground returns from a LAS/LAZ file so that the ground could be modeled separately without any vegetation or structures on it.  In response, I've written a new app called LAStoGround-XYZ that it will extract all the ground returns from a LAS/LAZ file and place them in an XYZ text file suitable for meshing or TIN'ing.  I'm also giving away for free another app called LAStoDXF-TIN.  It will create a TIN from a LAS/LAZ and generate a DXF as output.  You can get copies for OS X and Windows here:  http://lidarwidgets.com/#!page-freestuff

[ferrellsl]

I just made a minor update to the LAStoDXF-TIN app.  On some systems it would fail during DXF creation.  This has now been fixed.  You can still grab a copy here:

http://lidarwidgets.com/#!page-freestuff

[ferrellsl]

I've just updated a few tools for working with LAS/LAZ files and added a couple new tools.  My LAS to DXF point cloud tool now converts a LAS point cloud to a DXF point cloud and preserves the classification of the laser returns.  I.e. LIDAR ground returns are classified as a 2 so all ground returns are placed on layer 2 during the conversion process.  Vegetation is classified as 3, 4, or 5 for Low, Medium or High Vegetation and are placed on layers, 3, 4, and 5 respectively.  The batch conversion tool has also been updated to support LIDAR classification.

 

I've added a LAS Feature Extraction Tool.  This tool will allow the user to extract/export any or all 18 LAS classifications from a LAS point cloud so that they can be modeled/mapped independently of each other.  For example, all vegetation can be extracted for modeling or study, or all roads, railways and bridges can be extracted for traffic studies.  Of course it also allows the user to extract just the ground returns for terrain modeling or mapping as well.

 

See the feature extraction tool here:  http://lidarwidgets.com/#!page-feature-extraction

 

Another tool that I've added is called Mesh2Contour.  It allows the user to create contour maps in DXF format from a mesh file that was created by my LAS to Mesh generation tools.

 

See an output example here:  http://lidarwidgets.com/#!page-2dcontours

[ferrellsl]

Many of my clients have been supplied with LAS and LAZ files that haven't had any ground point return classifications performed on them which makes it nearly impossible generate a contour map or surface model.  There's a great open source tool called MCC-LIDAR which performs this function very well but it is a command line app and only compiled for 32-bit systems so the number of points it can handle is somewhat limited.  There are a couple commercial products that perform ground classification but they are extremely costly and some even require annual subscriptions.

I've written a GUI version of MCC-LIDAR and compiled it for 64-bit editions of Windows and released it as freeware.  It has successfully classified the ground returns on a file with 83+ million points and should have no trouble on larger point clouds/point sets.

 

You can download a copy here:  MCC-LIDAR GUI and see my other apps here:  LIDAR Widgets

Edited May 31, 2017 by ferrellsl
typo

[ferrellsl]

I've found that MCC-LIDAR does a very good job at classifying ground returns but it's very slow and doesn't support compressed LAS/LAZ.  To that end, I've written a GUI version of the excellent PDAL automated ground classification tool.  It supports compressed LAS/LAZ files and is significantly faster than MCC-LIDAR.  You can download the PDAL automated ground classification tool here:  PDAL Automated Ground Classification

 

There's also a version for MacOS/OSX.  If any Mac users would like a copy, please contact me via my web page here:  Contact

[on_da_mike]

@ferrellsl

no longer freeware?