CloudCompare forum

Good Evening!

I am a master student from Germany and i have a problem in importing Dense Point Clouds to CC which i created from UAV-Data in Agisoft Photoscan.
In Photoscan, the Cloud seems quite dense, almost like an Orthophoto. But imported in CC, it transforms itself in a kind of sparse raster-shaped cloud which is far more sparse than the original cloud seemed. This also influences the Quality of an processed Mesh for example, which shows a very bad quality, in contrast to the mesh from Photoscan. It seems like only a certain range of points is visible and used for processes- i already got through all display options and didn't find anything fitting.
Also, after re-importing a segment of the point cloud back to Photoscan, the rough raster-like shape stays.

Did somebody had a similar problem?

Best regards!
Geronimo

Has the point cloud large coordinates maybe? If yes have you accepted to apply the 'Global Shift' as suggested by CloudCompare?

What you describe looks a lot like accuracy loss because of large coordinates... Applying the 'Global shift' is only temporary and the original coordinates will be restored on output.

Hi Geronimo & Daniel,

I agree with Daniel, it sounds like some decimal places might have been cut.

So coordinates will be restored after global shift and re-safe, but what I have been dying to ask Daniel for ages is, if you generate some new processed point cloud from the original, e. g. by applying a tool like subsampling, will this new point cloud also get the restored (shifted) coordinates, and if not, how can I simply achive this?

cheers

mat

Well, normally most of the algorithms should preserve the Global Shift information (if not then it's a bug - or there's a good reason).