Calculating lidar Statistics

rmart

I'm updating some lidar classification and is getting really slow just in the "calculating lidar statistics" the more I work, the more it takes, sometimes taking up to 3 minutes.

It doesn't looks like much but when I have to make more than 100 changes I spend more time watching my pc calculating statistics that don't need.

I have the 01/06/23 daily build that in another topic said that it has addressed this issue, am I doing anything wrong?

bmg_mike

Do you have any Lidar classes filtered out for display on the Lidar tab of the Configuration dialog? In the daily build that you have the statistics shouldn't be calculated when modifying class unless you happen to have some types filtered out for display.

Thanks,

Mike

Global Mapper Guru

rmart

I do have the ground layer filter out so I can add some points that didn't get classified correctly.

I'll try without the filters

XionicFire

This is still happening on the 24.1 version, even with all filters off, its extremely frustrating

Is there a way to just turn it off manually from settings? or maybe a script command? I do not need lidar statistics at all, literally all I do is open the point cloud, clean the "flying" points that the automatic clean never gets, and save for further processing down the line, I've probably never even opened the lidar statistics window.

XionicFire

For reference:

its a 250 million point point cloud, being worked on this:

overclocked to 5Ghz with 16 physical cores, and a 4090 RTX card, and it still takes forever

Even if I delete ONE single point its 4-7 minutes of wait for calculating lidar statistics for every point or group of points I delete, it takes me several hours to clean some 20-30 clusters of points on the point cloud.

bmg_mike

Can you provide your point cloud to geohelp@bluemarblegeo.com ? The statistics should not be taking anywhere near that long to calculate, so something unusual may be going on.

The statistics are used internally to make all kinds of operations faster, so we do need them, but for very large clouds we may be able to improve this so that we can re-calculate the statistics much faster by storing statistics for groups of points, so then we only need to re-calculate the stats for the actual groups that are modified and just add in the other, so it would basically be instantaneous unless editing huge numbers of points at a time.

Thanks,

Mike

Global Mapper Guru