Hello All,
I received a large (~45GB) .e57 file that contains a lot of panorama images (over 300 scan points). The images are a huge problem because my computer only has 40GB RAM. I discovered that I can manually stop CC from loading the panoramas and somehow succeeded to subsample the point cloud.
But I'd like to know if there's a safer method that allows me to work with a large e57 file with limited RAM.
Is there a command line or a procedure that would allow me to open the .e57 file only partially (a limited subset of scans without panoramas) and subsample them with Random 25% method? I'm assuming I could merge them later once I reduce the file size.
Note: I cannot request a different .e57 file, so I have to handle the data myself.
Thank you!
Extract and subsample cloud data from .e57
-
- Posts: 2
- Joined: Wed Feb 19, 2025 12:25 am
Re: Extract and subsample cloud data from .e57
Ah, interesting concept, but we currently don't have any such option.
We could have an small interface that displays the list of scans and the list of images, and let the user load which ones need to be loaded. I'll add that to the TODO list. But it might take a very long time before it's implemented anyway ;)
We could have an small interface that displays the list of scans and the list of images, and let the user load which ones need to be loaded. I'll add that to the TODO list. But it might take a very long time before it's implemented anyway ;)
Daniel, CloudCompare admin
-
- Posts: 2
- Joined: Wed Feb 19, 2025 12:25 am
Re: Extract and subsample cloud data from .e57
Merci bien, Daniel!
So, what's the best approach to such large scans with what's currently available in CC?
Is there a better method than hitting "cancel" when the program begins to load the images?
Note: as of February 2025, the most current CC version on my Mac (Intel iMac, 40GB RAM, 16GB VRAM) allows me to cancel loading the images from the 45GB e57, but when I do the same on my Debian-based Linux machine (Intel, 64GB RAM, UHD 760 integrated graphics, CC from Flatpak) it allows me to cancel, but then CC crashes upon reaching RAM limit (every time). Typically, I see FOSS apps perform better on Linux, surprisingly CC works better on the Mac. Maybe it's the lack of dedicated GPU on my Linux rig, which is not for 3D work.
So, what's the best approach to such large scans with what's currently available in CC?
Is there a better method than hitting "cancel" when the program begins to load the images?
Note: as of February 2025, the most current CC version on my Mac (Intel iMac, 40GB RAM, 16GB VRAM) allows me to cancel loading the images from the 45GB e57, but when I do the same on my Debian-based Linux machine (Intel, 64GB RAM, UHD 760 integrated graphics, CC from Flatpak) it allows me to cancel, but then CC crashes upon reaching RAM limit (every time). Typically, I see FOSS apps perform better on Linux, surprisingly CC works better on the Mac. Maybe it's the lack of dedicated GPU on my Linux rig, which is not for 3D work.
Re: Extract and subsample cloud data from .e57
Usually the panorama photo data is smaller than the pointcloud data in an e57 file.blackcloud wrote: ↑Wed Feb 19, 2025 12:55 am Hello All,
I received a large (~45GB) .e57 file that contains a lot of panorama images (over 300 scan points). The images are a huge problem because my computer only has 40GB RAM. I discovered that I can manually stop CC from loading the panoramas and somehow succeeded to subsample the point cloud.
But I'd like to know if there's a safer method that allows me to work with a large e57 file with limited RAM.
Is there a command line or a procedure that would allow me to open the .e57 file only partially (a limited subset of scans without panoramas) and subsample them with Random 25% method? I'm assuming I could merge them later once I reduce the file size.
Note: I cannot request a different .e57 file, so I have to handle the data myself.
Thank you!
You could use something like e572las to extract only the point data, then subsample it with other free lastools to make it more manageable in size .. then view that in CloudCompare.
You will want a reasonable midrange GPU to view pointclouds regardless .. that may be your main issue ?