Hi
since there is no out-of-core stragety for RCPointBuffer, one can only write rcs files in the size of the points that can be hold in memory using createScanFromPoints().
Is this affecting the rcp point cloud in any way regarding performance in ReCap or other Autodesk products?
What I mean: I guess there is an internal octree structure in the *.rcs files, so if we split the point cloud in many big rcs files the octree structure is not very efficient.
Our idea was to write a rcs file everytime the RAM is nearly full (using createScanFromPoints()) and call createProjectFromScans() at the end.
In addition: If we call importFiles() at the end instead of using createProjectFromScans() will that create one big rcs file and will this be faster than using importFiles() on a las (since we call it with rcs files)?
Solved! Go to Solution.
Solved by yan.fu. Go to Solution.
Hi @martin.graner,
Importing multiple .rcs files won't create a big .rcs file for now. We will include this feature in future releases.
createScanFromPoints() is targeted for quick .rcs creation for points that can fit in memory.
If your workflow is to import points that cannot fit in memory, are the points stored in some files on the disk? If this is the case, you can implement your customized Codec to feed points from your files into ReCap engine. You can refer to .e57 or .las codec on how to feed structured /unstructured raw points to ReCap importing engine.
Regards,
Yan
Can't find what you're looking for? Ask the community or share your knowledge.
