Messages

46

Bug Reports / Re: Speed of 1.3.2 onwards vs 1.3.0

« on: December 30, 2017, 07:26:05 PM »

Hi Alexey,
So sorry seems I have a typo!
My readings should have been as follows:

The dense cloud build comparative timings and resultant number of points is as follows:
1.3.0 - Depth Map 00:32:50, Dense Cloud 00:28:54, Points 56,043,481
1.3.4 - Depth Map 00:19:24, Dense Cloud 07:39:00, Points 26,606,884

I understand the improvement in time for the depth map (better GPU utilization)...however reduction in speed for the dense cloud is many orders of magnitude, around 15x.  Maybe I am not understanding any benefits resulting from the change....

Thanks

John

47

Bug Reports / Speed of 1.3.2 onwards vs 1.3.0

« on: December 30, 2017, 06:08:27 PM »

I'm currently doing a benchmarking exercise to compare speeds of different machines, including Azure based VMs, and in doing so I seem to be experiencing a massive difference in performance between v 1.3.0 versus v 1.3.2 & v1.3.4 (the versions I have tested) for the dense cloud build.

I have a fully aligned data set of 1509 photographs, which i am running through an Azure NV24 VM machine.  This is quite a powerful machine with 24 cores and 2 x Tesla M60 GPUs, see screen grab.

The dense cloud build comparative timings and resultant number of points is as follows:
1.3.0 - Depth Map 00:32:50, Dense Cloud 07:39:00, Points 56,043,481
1.3.4 - Depth Map 00:19:24, Dense Cloud 00:28:54, Points 26,606,884

So note the MASSIVE difference in time to build the dense cloud.

All processing parameters were of course identical.

Interesting also is that the depth map took substantially SHORTER time using 1.3.4, but also the number of resulting points is much less, around half, yet another strange thing. 

What I noticed during the processing of the depth maps on 1.3.0 was that the GPUs were maxing out at around 10-12% each...could explain the timing difference here.

But of course the most significant factor is the timing on the dense cloud part, and also the difference in point count.

Any ideas?  Is this a bug in 1.3.2 and later versions?

Thanks

John

 

48

General / Re: Dense Point Cloud calculation taking suspiciously long time on AWS EC2 Instance

« on: December 30, 2017, 10:39:48 AM »

Hi Farza,
By any chance are you running the identical version of Photoscan on the EC2 machine as on your other test machine, or are you running a different version, if so, what versions?
Thanks
John

49

General / Older versions

« on: December 30, 2017, 10:37:09 AM »

Hi
Is there any way to download older versions of Photoscan?
The reason I am asking is that together with a colleague i am currently conducting some benchmarks on different machines,including Azure based VMs, and I am suspecting a significant performance drop in the dense cloud generation phase (at least with our with our test dataset), after version 1.3.0. 
Thanks
John

50

General / Re: Agisoft PhotoScan 1.4.0 pre-release

« on: December 26, 2017, 07:11:29 PM »

Hello John,

In the build 5543 you can select the points that includes the problematic area (no need to have an accurate selection) while looking from top, then in the Tools Menu select Dense Cloud -> Invert Point Normals option and in the related dialog check on "Opposite normals" option only. In this case only incorrectly oriented normals will be inverted.

Works perfectly! Thanks very much.

51

General / Re: Agisoft PhotoScan 1.4.0 pre-release

« on: December 16, 2017, 05:50:53 PM »

Hello jinjamu,

It seems that the normals for the points in the middle are inverted. Can you try using Invert Normals option in the Tools menu for the selected points?

Hi Alexey,
Indeed you are correct!
Is there an easy way to select the points that need inversion?  Manually is not so easy.
Is this something that will be corrected in future releases?
Regards
John

52

General / Re: Agisoft PhotoScan 1.4.0 pre-release

« on: December 16, 2017, 12:41:28 PM »

Hi Dmitry,
Release 1.4 looks great!  The new "Import Points" feature has allowed me to test a workflow where I take a point cloud out of Photoscan into CloudCompare which provides great features for editing point clouds (I am working on underwater models which can get quite "noisy" and require some manual editing), and then import the clean point cloud back into PS in order to mesh and texture.
However, one particular model I am working on ended up with a chunk of points missing when I did the import, and this was the case for both PLY and LAS file formats.
Attached you can see the point cloud imported with the missing chunk (empty piece in the middle), and also shown in cloud compare completely whole
I'm thinking this could be a software issue?
Thanks!

53

General / Re: Photoscan Pro on cloud based virtual machine

« on: December 02, 2017, 06:52:30 PM »

There's some info here, but a google search will find you much more
https://azure.microsoft.com/en-us/blog/azure-n-series-general-availability-on-december-1/
I have used the NV machines extensively.  Same performance as the NC machines, not sure the actual technical differences.  Price varies depending on the size of VM you select from around $1 to $5 per hour.
They work great to process models, but not so great if you need to edit on screen, as you'll need a really good bandwidth to have a decent user experience.
Just remember to switch off when you are done, otherwise you keep paying.
I think you can open a trial account and get $70 free usage, to try it out.

54

General / Re: Bringing a dense cloud back into photoscan

« on: December 02, 2017, 06:42:08 PM »

Thanks Alexey
I need the process to work as though the imported dense cloud is generated from the point cloud in photoscan, so there needs to be a "connection" (sorry, don't know the right words!) with the original photographs, to be able to texture the generated mesh.
Before I start to test, is there anything in particular to look out for?

55

General / Bringing a dense cloud back into photoscan

« on: December 02, 2017, 10:12:57 AM »

Hi,
The workflow I am using involves doing the alignment with Photoscan, then exporting as a bundler.out format, and using other software to generate the dense cloud and to clean up that dense cloud.  This provides me with a really nice clean dense cloud.
My question is, is there a way to bring that clean dense cloud back into Photoscan in order to mesh and texture in Photoscan?
Many thanks!

56

General / Re: Azure NC24 & Nvidia K80 not working with Photoscan (edit: DOES work)

« on: July 09, 2017, 02:58:05 PM »

Hi
Rather than starting a new thread on Azure GPU machines I hope you don't mind continuing on this one.
I've started to carry out some benchmarking on Azure VMs in anticipation of of a project where I form part of a team which will be using photogrammetry to document deep historic wrecks in the seas around Malta.  I am working with a sample set of 607 stills captured from HD video (so 2 Megapixel each still), attempting to compare the performance of the NV24 and NC24 VMs.  (More info on the VMs here https://azure.microsoft.com/en-us/blog/azure-n-series-preview-availability/ )
I aligned the images, optimised the cameras and generated the dense cloud.  In both cases I reset the Photoscan parameters to default to ensure the same baseline.  I ran the alignment and the dense cloud both on "High", otherwise leaving the settings as per default.
What has really confused me is that the different machines, running the same software version, with the same settings, on the same set of photos, generated different outcomes!  For starters, the NV alignment was more successful than the NC - after the initial run, on the NV just some 6-7 adjacent photos were not aligned and had to be aligned using the Right Click..Align method.  On the NC there was an additional set of some 15 adjacent photos which needed "forced" alignment.  The final outcome was the same though - 603 out of 607 aligned
When I ran the dense cloud, the NC was a lot faster, however on examining the result, the NC only generated some 7 million points vs the NV's 17 million - which I guess accounts at least in part for the NC being faster.
I am attaching the 2 different "info"s....there are a number of differences also in other areas (way above my head) for example the RMS reprojection errors are totally different, etc
Would be really curious to know if anyone can provide any explanation.
The models can be made available if anyone wants a closer look
Thanks,
John

57

General / Saving as high quality TIFF

« on: April 18, 2017, 10:48:56 PM »

Hi
I am using Photoscan Standard to document shipwrecks, and I would like to find a way of generating a high quality, large TIFF from the textured model.  There does not seem to be an easy way to do this (Export....TIFF), so I was wondering if anybody had any advice for me.
Thanks!
John

58

General / Azure NV VMs for Photoscan

« on: January 19, 2017, 11:03:13 PM »

Hi
Does anybody have experience using the Azure NV range of VMs for Photoscan?  The spec seems pretty impressive, as can be seen here
https://azure.microsoft.com/en-us/blog/azure-n-series-general-availability-on-december-1/
I have done a couple of quick tests (with the GPU enabled on a NV6 VM) and I expected a lightning performance compared to my i7 laptop with a Geforce 745M, but I was a little disappointed.  For example, the identical dense cloud generation took 22.3 secs on the VM vs 31.9 secs on the laptop. The biggest speed improvement I experienced was with the Align Photos part of the workflow, 17 secs vs 40 secs.
I have only tried this with a small set of 28 photos, so maybe bigger savings come with more photos, but I am curious to know other people's experience, especially with the Nvidia M60.
Thanks!