Topics

1

General / projection file - AGD66 EPSG20255 + AHD heights

« on: March 26, 2019, 07:13:59 AM »

hi team,

we have some equipment that exports either WGS84 lat/long and ellipsoid heights, or WGS84 lat/long and AHD heights.

we are having alot of trouble coordinating the Z value of the projects due to incorrect height transformation.

is it possible to be supplied with a WGS84 projection file with AHD as the vertical datum?

Using the standard wgs84 to convert the ellipsoid heights isnt accurate due to a local datum being used.

thanks!

2

General / Separate areas in the one chunk - works MOST of the time

« on: February 19, 2019, 03:00:55 AM »

morning Alexey,

we generally process discrete areas from a single RPA flight in the one chunk (with no overlap between areas) and the alignment process works just fine for all cameras (see attachment report mulit-job.jpg)

however lately, with one particular RGB camera, it only aligns one of the areas-rather than all areas. We have tried disabling one section of cameras, which allows the other section to align, but they wont process together. (refer photoscan attachment separate.jpg)

after aligning photos, we have tried resizing the region to encompass all images, however a re-alignment still only aligns the first section and the region is resized to the extents of the first area. I found a forum post from a few years ago, discussing why it SHOULDNT work (mulitple areas without overlap), but this doesn't explain why it DOES work with one of our RGB cameras, but not the other.

i have also tried selecting the unaligned photos and 'aligning selected cameras', but this doesnt do anything.

your help here would be appreciated. we are using photoscan 1.3.4

thanks!

3

Python and Java API / Boundary export option

« on: January 31, 2019, 08:35:41 AM »

hi Alexey

i notice since V1.4.3, an outer boundary is now used to clip the exported models.  Previously this was only applicable to ortho/dem.

whilst i can appreciate the convenience of this feature, is it possible to implement (if it doesn't exist already, as i cant see it in the API) a flag we can set as to whether a particular export should honour the boundary or not?

for now we are re-importing our shape as 'noboundary' and running the exports. it works - but hoping it can be neater?

thanks!

4

Bug Reports / cudastreamdestroy failed error

« on: October 24, 2017, 11:34:05 AM »

hello

running agisoft 1.3.4  build 5067 on a xeon 2687w v4, 3x1080.

I am getting the attached error as soon as I started 'align photos'

2017-10-24 18:24:39 Warning: cudaStreamDestroy failed: all CUDA-capable devices are busy or unavailable (46)
2017-10-24 18:24:45 Finished processing in 12.54 sec (exit code 0)
2017-10-24 18:24:45 Error: unspecified launch failure (4) at line 119

and a windows 10 notification of "application has been blocked from accessing Graphics hardware"

I have enabled the 3 x gpu's in preferences and disabled the CPU.

disabling GPU's and enabling CPU's allows the align photos to start. It is still processing, so no comments yet on the dense cloud generation.

I have tried multiple versions of agisoft and multiple version of Nvidia driver between 385.28 to 388.0

we have another system with different specs that doesn't have any issues.
can you help please Alexey!?

thanks

5

General / System configuration check please

« on: September 07, 2017, 11:00:14 PM »

Time has come for us to invest in a few more processing nodes.  I'm still tied up as to whether to go for high GHz i9 or stick with dual 2687v4.  We intend to equip with 4x1080. As you can tell Pryce isn't really an issue, but any decrease in processing time is key.

Would really appreciate some suggestions on cpu, as I have read so much here now I don't know which way to go!

Dual 2687v3 ( or i9 7900x , i7 7820x, i7 7740x,)
128gb ram
4x1080

I am concerned that not having enough CPU cores will affect when we enable 4xgpu.

We process with a custom script with the below job settings

Photos: 200-1000
Res,: 20mp
Alignment: high
Dense cloud: medium

Any help you could give would be great please

6

Python and Java API / 'null' value on exportOrthomosaic

« on: March 07, 2017, 09:13:05 AM »

good afternoon,

i thought i remember seeing somewhere that the null or background value for an exported Orthomosaic changed in the latest 1.3.0

previous release 1.2.3 had a background value of 0,0,0, but version 1.3.0 has background value 255,255,255

is this easily changed somewhere? i cant see it as an option in the API.

we would like to change it because we have alot of downstream processes setup to use 0,0,0

thanks

7

Python and Java API / Bug? chunk.exportModel & vertex colors

« on: February 25, 2016, 12:35:01 AM »

Hello

i have some code that exports the 3dmodel as an OBJ format.

when exporting the model through the GUI, with no colors or vertex normals selected, i get an OBJ model that is (for eg) 160mb in size, and exports quite quick.

with the below snippet, the software continues to calculate vertex colors, even though it appears i have set it to false.
the exported model is 320mb and takes a few minutes longer.

is there something wrong with the code, or is this a python bug?

chunk.exportModel(path, precision=6, format='obj', projection = proj, texture=False, colors=False)

thanks
Luke

2nd q: is it possible to export the 3dmodel clipped to the shapes?

8

General / Customising Open_CL max group size settings

« on: August 26, 2015, 02:55:43 AM »

morning guys.

we have setup a workstation, along the lines of my previous post with the following:

Xeon E5-2687w v2 (3.4ghz) 8 cores
64gb ram
12gb Titan X
4gb Quadro K5000

Our current workflow is processing orthophotos and 3d models from an RTK UAV with a GSD of around 4cm.

We currently process with:

Photo Alignment: High
Dense point cloud: Medium
Mesh: Medium

Orthophoto: 50cm

The above settings are more than adequate for our output purpose and provide a very fast turnaround on the data.

One issue we are seeing with this upgraded workstation is the 'max group size' setting (which i believe to be the amount of data sent to the GPU's at once) appears to be too low.

The Titan X processes each piece in less than 1 second and doesnt get to max performance, and only reaches 45% utilisation. (reported performance 800 million with 0/16 CPU enabled, or 625 million with 12/16 CPU enabled)
The K5000 takes a little bit longer and hits 92% utilisation. (330 million)

I understand this is due to the resampled photo size (set by dense point cloud on medium) because when it is set to ultra, the Titan X hits full performance. (1200 million)

My question is: is it possible to edit the max_group size settings in Open_CL to send more data to the GPU's (it seemingly only sends 1 picture at a time.  my thoughts are if this could happen, the GPU would spend longer processing and there hit full performance?

is this the case? or are we 'doomed' to a reduced performance score on the Titan  X due to the processing settings?

thanks!

9

General / Help with new gpu configuration

« on: July 06, 2015, 10:05:32 AM »

Hi all, first up, there is so much shared knowledge here it is fantastic.  Thanks alot.

So I need to configure a system for max performance. At this stage bang for buck isn't a consideration, but over price could be an issue, and there are some hardware limitations.

The system is based on a hp z820.

CPU is an xeon e5-2687w

Gpu is where it gets interesting. I have read that the mobo can only support sli for quaddro cards and it comes with a K5000. I need the best performance I can. Whilst I said $ is not an issue 2 x K5000 is too much.

I was Looking at dual 980gtx but assumed this wouldn't work in agisoft,  is this the case if the board doesn't support sli? Failing that a titan x was the backup option.

Can anyone advise whether 2 x 980 ti will work in agisoft if not sli and still score higher than single titan x.The comparison sheet seems to score them higher than a single titan x.

Thanks