Messages

1

General / Re: Agisoft Metashape 1.6.3 pre-release

« on: May 17, 2020, 11:49:55 PM »

The 1.6.3 seems to be not properly signed/notarized:
❯ spctl -a -vv MetashapePro.app                                                 
MetashapePro.app: invalid signature (code or signature have been modified)

2

General / Re: Phantom 4 rtk + d-RTK 2 base workflow. Z errors. Any suggestions please?

« on: April 13, 2020, 11:20:23 PM »

Hello Ian_riverscience,

Yes I encountered similar problems using P4RTK with an NTRIP-Caster, but the RTK-datasource is not the problem. The main issue is the wrong f-value in the camera calibration. The auto-calibration is not capable to determine the +/- appriopriate value, all other values are meaningful (!).
Four our P4RTK, instead of 3666.666 (8.8mm/0.0024) we got 3678. For low flights (e.g. 40m) the error is not so severe, but flying at 120m amplifies the error.
How to get the correct f-value. Take an already flown campaign with 200 images and iterate the f-value (only this value!) by reprocessing, until you find our best f-value, actually to estimate  the true focal length for your device. A bit tedious, but it works. It is not clear for me how stable the estimated value is over longer periods (months), concerning temperature and mechanical movements.
Try it.

3

Bug Reports / Re: building tiled model with 1.5.0 7492 crashes computer hard

« on: January 16, 2019, 10:16:12 PM »

OK - I will try out 8 or 4 - takes some time...

4

Bug Reports / building tiled model with 1.5.0 7492 crashes computer hard

« on: January 14, 2019, 01:18:39 PM »

Hello,
I tried out several times building tiled models using 100 DJI-images without any problems.
Using Ultracam aerial images (17004x26460) did crash the computer really hard after ca. 50% of the GPU work (goto line 652 of the log).
There is no even a system-crash log...

The preceding image alignement worked without any errors.
Is it related to large images ? FirePro D700 ?
I assumed the tiling approach avoids the "image is too large" problem

Ruedi

5

Bug Reports / Re: Python with PhotoScan.RasterTransformValue

« on: January 04, 2018, 03:26:26 PM »

Hi Alexey,

variable "saveOrtho" contains: /Volumes/FastWork/PGResults/WSL/eBee/Sequoia/Sequoia_2017_11_01_v1.4.0_hMatch_proj_v1.4.0_hMatch_hDens_moFilt_ortho.tif


I used the same scripting classes for hundreds of successful process results.
When I use "raster_transform = ...", the export always fails,  without  "raster_transform = ...", everything works without RunTimeError and I get a correct exported orthophoto-product, exported report, exported points.
Using the saved project and afterwards manual export of the ortho with NDVI colors (via raster calculator) works!

??
Ruedi

6

Bug Reports / Python with PhotoScan.RasterTransformValue

« on: January 03, 2018, 03:14:53 PM »

Hi there,
I try to use raster-transforms, actually to export NDVI-Orthos.
When I run my script, a RunTimeError occurs, when exporting the ortho
Without "raster_transform = PhotoScan.RasterTransformValue", everything works fine.
Do I use the wrong syntax ?

Thanks, Ruedi


...
               self.chunk.raster_transform.formula = ["(B4-B2)/(B4+B2)"]
               self.chunk.raster_transform.calibrateRange()
               self.chunk.raster_transform.enabled = True
               self.chunk.exportOrthomosaic(saveOrtho, image_format = PhotoScan.ImageFormatTIFF, raster_transform = PhotoScan.RasterTransformValue)
...
Log:
2018-01-03 10:46:00 Agisoft PhotoScan Professional Version: 1.4.0 build 5650 (64 bit)
2018-01-03 10:46:00 Platform: Mac OS
2018-01-03 10:46:00 CPU: Intel(R) Xeon(R) CPU           X5670  @ 2.93GHz (server)
...
2018-01-03 12:43:33     self.chunk.exportOrthomosaic(saveOrtho, image_format = PhotoScan.ImageFormatTIFF, raster_transform = PhotoScan.RasterTransformValue)
2018-01-03 12:43:33 RuntimeError: Invalid data type
2018-01-03 12:43:33 Error: Invalid data type

7

General / Re: Agisoft PhotoScan 1.4.0 pre-release

« on: December 19, 2017, 05:11:52 PM »

1.4.0 API (chunk.buildDepthMaps)
Hi Agisoft,
Please, put significant changes also in the python-api, otherwise you get something like

Agisoft PhotoScan Professional Version: 1.4.0 build 5532 (64 bit)
Platform: Mac OS
...
self.chunk.buildDenseCloud(quality=DCQuality,filter=DCFilter)
TypeError: 'quality' is an invalid keyword argument for this function
Error: 'quality' is an invalid keyword argument for this function
...

and you have no idea why (seems to be new since 1.4.0 build 5432)...
Thanks, Ruedi

8

Feature Requests / Sequoia import (work around for missing channel files)

« on: September 12, 2017, 02:00:04 PM »

Hello Agisoft,
Recently I had a couple of Sequoia-datasets (using v1.4.1 Sequoia firmware...) with partially missing channel-files, e.g. a single xxxGRE.TIF or xxxRED.TIF was missing. Typically only 1 single file  from 400 is missing.
Unfortunately when Photoscan v1.3.x detects a missing channel, it treats all images as normal dataset and the multispectral-import does not appear.

I would suggest the following approach:
Please only check if the single multispectral image-tuple is complete (_GRE.TIF,_NIR.TIF,_RED.TIF,_REG.TIF (don't care about the rgb.JPG)). If one of the images in a 4-tuple is missing, skip only this tuple and continue searching.

9

Python and Java API / Null elevation

« on: March 08, 2017, 04:39:49 PM »

Hello Agisoft,

Using 113 UAV images (with X,Y,Z) and 3 GCP with proper X,Y,Z-values

Everything like build dense cloud, etc. runs fine with my Python script
With Build 3772 the scripts stops ending up in the GUI-Console with:
...
2017-03-08 14:27:50     self.chunk.exportDem(saveDEM)
2017-03-08 14:27:50 RuntimeError: Null elevation

With Build-version 2834 and older versions it runs without any complaints and generates a proper report, DEM,etc.
There is an PhotoScan.Elevation-class, but how to handle it ? in previous builds, the reference height was obviously tkane from the GCP-data  !(?)

10

Bug Reports / Orthomosaic Export Problems with v1.2.5 (2735) and Float32 image data

« on: August 10, 2016, 03:35:56 PM »

Hello,
we are using a FLIR-camera and the same set of images works fine concerning aligning,dense cloud,etc. with v116 and v125.
Each input image is a 1-channel Float32 TIFF, containing temperature values, typically 10..40.
With v116 I get a descent temperature  Orthomosaic in channel 1.
With the current v125 the scaling of the Orthomosaic is now somehow strangely normalized (I tried also with  PhotoScan.BlendingMode=DisabledBlending).
How to get the channel layout (and disabled scaling ) of v116 ?
Ruedi

Float32-images with v1.1.6
 
Band 1 Block=2647x32 Type=Float32, ColorInterp=Gray
  Minimum=-17.021, Maximum=49.017, Mean=10.409, StdDev=9.912
  Metadata:
    STATISTICS_MAXIMUM=49.017379760742
    STATISTICS_MEAN=10.409090188344
    STATISTICS_MINIMUM=-17.020740509033
    STATISTICS_STDDEV=9.9120056741228
Band 2 Block=2647x32 Type=Float32, ColorInterp=Alpha
  Minimum=0.000, Maximum=1.000, Mean=0.541, StdDev=0.498
  Metadata:
    STATISTICS_MAXIMUM=1
    STATISTICS_MEAN=0.5414550280809
    STATISTICS_MINIMUM=0
    STATISTICS_STDDEV=0.49827851714359


Float32-images with v1.2.5 (2735)

Band 1 Block=256x256 Type=Float32, ColorInterp=Gray
  Min=0.000 Max=0.000
  Minimum=0.000, Maximum=0.000, Mean=0.000, StdDev=0.000
  Overviews: 1376x836, 688x418, 344x209, 172x105
  Metadata:
    STATISTICS_MAXIMUM=0
    STATISTICS_MEAN=0
    STATISTICS_MINIMUM=0
    STATISTICS_STDDEV=0
Band 2 Block=256x256 Type=Float32, ColorInterp=Alpha
  Min=0.000 Max=1.000
  Minimum=0.000, Maximum=1.000, Mean=0.663, StdDev=0.473
  Overviews: 1376x836, 688x418, 344x209, 172x105
  Metadata:
    STATISTICS_MAXIMUM=1
    STATISTICS_MEAN=0.66312668646378
    STATISTICS_MINIMUM=0
    STATISTICS_STDDEV=0.47264117908128

11

Feature Requests / Open Project and run python script from command line

« on: June 13, 2016, 03:59:25 PM »

Hi,
Opening projects and launching Photoscan works fine, the world would be even better if as additional parameters a related python script with its argument could be added on command line:
in MacOS-unix:

open -a PhotoScanPro.app /Volumes/BigData/project09638.psx runPHScanner.py projectParameters09638.txt


preferably also for Windows/Linux...

Ruedi

12

Bug Reports / Re: can't export point cloud

« on: December 20, 2015, 12:34:30 AM »

Hi,
Using the Python API, I also found, exported point clouds in laz,ply are everytime corrupt due to the "cant write".
Ruedi

13

General / Re: Apple's graphics "Metal" API

« on: November 30, 2015, 04:50:52 PM »

Hi Jerry,
The metal API for OSX is a very low-level API which does not make sense to use within Agisoft (Qt is controlling the GUI), it is +/- a game engine API to render graphic context.
As OSX or Radeon-Hardware users, we can be quite happy that Agisoft  uses the generic OpenCL  for any meaningful numeric acceleration, so we are not forced to use specific hardware.

Ruedi

14

General / Re: Agisoft PhotoScan 1.2.0 pre-release

« on: November 25, 2015, 11:01:00 AM »

I found some minor confusing documentation issues in the (amazing !) Python API build 2198:

1. (page 15,16)
buildDem(source=DensePoints,...
...
source (PhotoScan.DataSource)
...
   
2.  (page 16)
buildOrthomosaic(surface=OrthoSurfaceDem,....

...
surface (PhotoScan.DataSource)
...

declaration is:

class PhotoScan.DataSource
   PointCloudData, DenseCloudData, ModelData, ElevationData

=> what is "densePoints"
=> surface points to the wrong enum
=> I could not find "SparseCloudData": is "PointCloudData" == "SparseCloudData" ??

Thanks, Ruedi

15

Bug Reports / Re: Different pixelsize in exportDem and exportOrthophoto in v1.1.0

« on: January 12, 2015, 10:59:09 PM »

I used

buildDenseCloud(quality=PhotoScan.HighQuality)
buildModel(surface=PhotoScan.HeightField, source=PhotoScan.DensePoints, interpolation=PhotoScan.EnabledInterpolation, face_count=PhotoScan.HighFaceCount)
buildUV(mapping=PhotoScan.OrthophotoMapping)
buildTexture(blending=PhotoScan.MosaicBlending, size=4096)