Forum

Author Topic: LAS files with a shift  (Read 6246 times)

Herman

  • Newbie
  • *
  • Posts: 22
    • View Profile
LAS files with a shift
« on: December 28, 2012, 09:09:49 PM »
Hi
I am a surveyor working mostly in Namibia and we have a rather standard projection.
A Transverse Mercator , two degrees wide per panel and based on the the odd longitudes
lines e.g. 13, 15, 17 Deg. East .
Our Northing's  is based are base on the 22 deg. South latitude and we use the Besels 1841
Geoid
The system is called Schwarzeck and for example the the Lo panel 15 is the standard
EPSG : 29375 projection - catered for and working perfectly in PS.
The problem I do have is that the LAS files generated are positioned roughly 1 deg. ,
85km to 110km wrong when viewed in Global Mapper and overlaid with the TIFF
images or the DEM that all plot perfectly.

Can anyone offer a possible explanation?

Regards
Herman

Alexey Pasumansky

  • Agisoft Technical Support
  • Hero Member
  • *****
  • Posts: 14927
    • View Profile
Re: LAS files with a shift
« Reply #1 on: December 29, 2012, 05:58:35 PM »
Hello Herman,

Thank you for reporting. We will check this problem.

Do you have this problem only with Schwarzeck projections or with others too, if you have exported point cloud in another projections, of course?
Best regards,
Alexey Pasumansky,
Agisoft LLC

Herman

  • Newbie
  • *
  • Posts: 22
    • View Profile
Re: LAS files with a shift
« Reply #2 on: December 29, 2012, 07:22:02 PM »
Hello Alexey

Thank for the quick reply - I have to note again that between yourself & Dmitry support is very professional and your software is great!!!!!!

I unfortunately I only work in Schwarzeck , I will try with a smaller project and convert the GCP to
WGS 84 and test that.

Regards and the best to you all for 2013

Regards
Herman

mwillis

  • Full Member
  • ***
  • Posts: 140
    • View Profile
Re: LAS files with a shift
« Reply #3 on: December 29, 2012, 09:27:09 PM »
For what it is worth, I tested this several months back using UTM and it seemed to align correctly.  I did not get great results from LAStools but worth more investigation.

Matt

  • Full Member
  • ***
  • Posts: 104
    • View Profile
Re: LAS files with a shift
« Reply #4 on: December 30, 2012, 08:20:44 AM »
LASTools also has very good support as long as you are prepared to pay for a software license. Martin has a great network of highly skilled users and a good forum to draw experience from.
It takes some time to learn the correct method and values for filtering and exporting but the toolbox has given me very good results in may different landscapes with very large datasets.

Herman

  • Newbie
  • *
  • Posts: 22
    • View Profile
Re: LAS files with a shift
« Reply #5 on: December 30, 2012, 09:47:32 AM »
Hi Matt & Willis
Thanks for the reply. I will try with a different projection. But I would eventually have to revert back to my local projection.
I think the problem is not with LAStools.
The PS file and the LAStools file (after classification) plot in exactly the same position, so LAStools does not intorduce any positional error
I think the problem is partly a Globalmapper problem and a PS problem - the DEM and TIFF files are correctly located in Ver.8 of GM but the LAS file is offset by 80km's.
But since ver.11 you have to multiply with -1 to have data plotting the correct way up , TIFF images plot 180deg. rotated etc.
I agree Matt the bit I have seen and the attitude shown by Martin makes me very keen to invest in this software and I will gladly pay to use it.
Similarly the PS guys support is also great and hopefully this is only a minor projection offset error.
Have any you tried VRmesh , also looks very promising.
Regards
Herman

George

  • Full Member
  • ***
  • Posts: 193
    • View Profile
Re: LAS files with a shift
« Reply #6 on: December 31, 2012, 01:22:46 AM »
Probably that is off topic ... Just as a comment ... we roughly discussed the issue of coordinate systems with Alexey but it seems like PS does its job perfectly (atleast regarding coordinate sys's)
... recently ArcGIS had a problem with Finnish coordinate system with EPSG: 3067 so that when you had a .tiff file produced at PS imported into ArcGIS a huge shift occured of approx. several km to the south ...
In that case that was a bug of ArcGIS ... In other words, PS did it's job correctly ...