Messages

1

General / Re: Sparse cloud mesh - always extrapolated in 1.2?

« on: December 29, 2015, 07:29:42 PM »

Arbitrary for these models. Did not test on height field. I did do a test with a small part of a sense cloud, seems to be behaving normally in that case.

2

General / Sparse cloud mesh - always extrapolated in 1.2?

« on: December 29, 2015, 05:22:47 PM »

Just getting up to speed with V1.2. It seems like identical results building mesh from sparse cloud with the interpolated and extrapolated options; both seem to be extrapolated. Is this a bug? In past versions the interpolated option typically avoided the bulbous over-meshing that the extrapolated approach often created. Note that I am working on oblique views of steep terrain, so the meshes shouldn't be closed.

Thanks!
Dave

3

Bug Reports / Re: Large coordinates truncated in PLY

« on: June 08, 2015, 05:08:20 PM »

Hi Alexey
Thanks for your reply. I have tried the text PLY export, and it works well. But, my models are often in the 50-75 million face range, in which case the text PLY is a huge file and takes a really long time to write. For now the best workaround has been to export points as .las in UTM, and model in .ply local coords, then align these in Cloud Compare to get the model in the UTM coordinates for comparing to other data. For some reason I've had less success using the transformation at export, although this is probably just errors on my part. Its also a bit less convenient in Cloud Compare,
Cheers
Dave

4

Bug Reports / Re: Large coordinates truncated in PLY

« on: June 04, 2015, 06:07:55 PM »

I know this is an old thread, but I've been struggling with this issue lately. My workflow involves making fine co-registration between models in Cloud Compare, so exporting in real-world coordinates is best. It would save me a bunch of time and potential for errors to be able to export direct to .ply in UTM, with no shift required. Point clouds work great, of course, both on export and import. And, Cloud Compare can handle the big coordinates, and even proposes the global shift. (Even this would be a very useful feature in Photoscan!)

Is there any thought to work on this issue in coming versions?