This is a continuation of an earlier post wherein I described problems with converting a large LIDAR dataset to DTM in Power Civil. I have the points file clipped to the area of interest and have sucessfully built a DTM. Now I want to thin the data to improve performance. The clipped data file contains almost 5 million points and is about 166 MB in size. When I run the Filter XYZ tool in LIDAR tools I get a memory allocation error. I have tried both the TILE and TIN filter options and have used the recommended settings in the help system.
I have 2 GB of memory on my Dual Core 1.6 GHZ XEON machine. Any suggestions on alternatives or troubleshooting techniques?
Power Civil XM ver. 08.09.06.30
I decided to have another try at generating contours from the TIN file and am now able to load them sucessfully. I suspect that during my efforts to find a solution to my task I was loading and unloading other applications that may have left residue in my main memory and caused the allocation problems. Today I sucessfully generated 2' and 10' contours from my UN-thinned TIN consisting of 4.9 million points...impressive!
Neil Wilson (aka Neilw)
Power Civil v8i 08.11.07.245
AutoCAD Civil 3D 2018
In the mean time, here is a suggestion that perhaps will help get you underway:
Not the most efficient method, sure, but if you are facing a deadline.
Note: Pay special attention to the edges where the multiple small TINs merge back together to check for oddities after they are merged.
Robert GarrettSenior Product EngineerBentley Systems Inc.
Good Morning Neil
The 30 million number apprently comes from our help file. (Thanks Mike). My experience still indicates that your computer is choking on the size of the data. But I am investigating. I have a LIDAR test set, but if you can send me a link to your your LIDAR file then I will look at yours. Or post it at ftp://ftp.bentley.com/pub/incoming and send me an email when it arrives.
I did thin the data as far as I dared to. Z filter was 1.0 which gave marginal results. Actually Power Civil is handling it pretty well considering the size. The 30 million points info came from various inside sources. Not to say that is an official claim, but I believe it is in a blog or forum somewhere as well.
I managed to get the data into my other applications by using the same thinned XYZ file so the XML export is no longer necessary at this time.
It looks like we are going to need to move to a 64 bit OS soon with data like this becoming more and more common.