Quality report for aligning laserscans

[Dieter]

Now that the output in E57 format works then we should turn to the next -very important function for a surveyor- a proof of accuracy in alligning laser scans.

It would be very nice to have a graphical preparation with the corresponding errors of the individual points analogous to the output mask "Show matches" for the TiePoints.

In addition, the following information should be presented:
-Distribution and number of identical points found
-Accuracy of the identical points found
-And from this calculates a key figure for quality estimation (0=bad 100=super)

Other suggestions are welcome.


Dieter

[Tas]

Dieter - with your suggestions above in mind, do you know of anyone using Metashape for professional laser scan registrations? I haven't even considered using it to replace something like Leica Cyclone or Register 360

[Dieter]

As long as the scanning support is as poorly implemented as it is at the moment, I wouldn't do that. I do 95% of my scans in Faro Scene. I only calculate in Metashape if we have a combination of images and scans. And that's usually more trial and error. A lot still needs to change in Metashape before I can use it more often.

[jedfrechette]

I’d second Dieter’s assessment of the lidar support. At the moment, it’s not really adequate for laser scan registration. When we combine scans and photos it feels like we need to jump through quite a few hoops to make it work and prevent Metashape from messing  up our high quality existing scan alignments.

I don’t actually think it is a problem that Metashape isn’t good at lidar alignments, there are other good software packages for doing that. Rather than invest time on lidar alignment, I’d much rather have the Agisoft team improve the process for conforming the photo alignment to the lidar.

[Tas]

Jed, I entirely agree. I've invested quite a bit of time into Agisoft workflows, but I'm wondering if I need to switch to another software for jobs that include laser scans. Is that your current approach?