Technique behind combining Laserscans with Images

Comments

7 comments

  • Avatar
    Götz Echtenacher

    Hi Julian,

    have you read around in the forum yet? There are plenty of threads about that.

    If you want sub-cm accuracy on a rooftop, then you need ground control points there.!   :-)

    I often use natural points for that, as in corners of tiles or such.

    0
    Comment actions Permalink
  • Avatar
    Julian Weber

    Thanks for the reply :)! Yes I've read a lot of threads already but i didn't found an answer yet :(.

    I want to know if RC uses the Laserscans as GCP's (so you would have millions of GCP's for the images to align) or does it just handles the panorama images of the Laserscanner as regular images and align it with the DSLR images?

    Do you use a total station and shoot the natural points reflectorless? :)

     

    0
    Comment actions Permalink
  • Avatar
    Götz Echtenacher

    No problem.

    It's been discussed quite a bit (I keep writing about it :-) but I admit the forum is not really good for searching.

    RC does convert the scan into 6 images arranged as a cube - you can see them in the 3D view. And hence the same rules as for photogrammetry apply.

    And yes, I often use a total station. I've done a thorough comparison once and if the object is ideally textured, the results are comparable to a laserscan, or better. And since you need a total station for laserscans, too, it's even quicker. Ther are cases where a laserscanner might still be better though, for example with unicolored surfaces or dark areas.

    0
    Comment actions Permalink
  • Avatar
    Julian Weber

    thats a good comparison :)!

    I do everything with cloud to cloud registration (with trimble realworks) and the accuracy is in most cases <2-5mm (much better than using paper targets for registration because all Points are considered in overlapping areas). I usually use total station to georeference the whole Pointcloud andn to verify registration results. We have a P4P but we didn't used it yet and the rooftops give me Headaches :/. My plan would be to generate markers in RC and put them on the facade. after that i would scan the rooftop also in 45 degree angles to capture the facade for overlapping and this should make the roof accurate because it is like a "loop". The markers on the facade which are captured by the Laserscanner also, are used to verify the Results of the Image-Alignment.

    So does the Software only use the 6 Images to Align the Images+Scans?

     

     

    0
    Comment actions Permalink
  • Avatar
    Julian Weber

    Götz how can I send you a private message?

    0
    Comment actions Permalink
  • Avatar
    Götz Echtenacher

    Not in this forum but you could use the email address that you used in january to contact me...   ;-)

    0
    Comment actions Permalink
  • Avatar
    Julian Weber

    I already wrote a message ;-)

    0
    Comment actions Permalink

Please sign in to leave a comment.