1

I'm currently performing an engineering honours thesis on terrestrial photogrammetry and laser scanning in Archaeology. Both methods will be used to document a single location first and analysis will be performed into which technique is best suited to different styles of documentation.

Geo-referencing and scaling will be performed with reasonably permanent targets (e.g. stick on or glued). The initial area of measurement is quite small (5m x 5m) so I am looking at placing 4-5 geo-referenced targets.

It would be great if i could use targets that would be auto-recognized by most software packages. I have looked for a solution but am coming up empty handed.

Are most targets proprietary and linked to a specific software or is there a open-source style of target recognition?

It would be amazing if i could use the same targets for both laser scanning and photogrammetry. However, I am open to the idea of using two styles of target, one for photogrammetry and the other for laser scanning. For instance, using the photogrammetry targets to accurately scale and georeference the photogrammetry model but then registering the laser scanning targets manually, so there are still points of comparison. It would be great to use the same set of data rather than multiple sets with different targets for each software. This would cut my data management and time down immensely.

In sum, I am asking is there a standard coded target that works across multiple software solutions, and could also be used to manually register a point cloud.

The software I will be experimenting with are:

  • COLMAP
  • Meshroom
  • Pix4D
  • 3DF Zephyr
  • Agisoft Photoscan
  • RealityCapture
  • 3DVEM
  • CloudCompare
  • Leica Cyclone
  • AutoDesk ReCap
Improve this question
7
  • How are you acquiring your LiDAR? Your airborne or ground based scanner should have manufacturer recommended software with it which will convert the raw laser data with inertial navigation information to 'georeference' the pulses.. you can create an intensity image (raster) from the LAS to locate your targets to confirm your accuracy but it is likely your elevations are ellipsoidal and need to be converted to a local height datum to compare with your photography... essentially you're comparing apples and oranges.
    – Michael Stimson
    Commented Nov 8, 2018 at 2:32
  • Terrestrial laser scanning. I would love to use a drone but there are legal requirements. I'm using a total station and known coordinates (MGA in Australia) to create a control network from the targets. I'll perform some least squares and hopefully obtain a pretty tight georeferenced network from which i should be able to compare the two models in CloudCompare. As they will be located in the same space, the differences in the models should become clear if overlayed onto each other.
    – William Andrews
    Commented Nov 8, 2018 at 4:07
  • I don't know much about terrestrial laser scanning (or drones), I specialize in airborne laser scanning. Your laser scanner should come with software recommendations to convert the pulses into coordinates, reading the output of the total station, further adjustment may be unnecessary.
    – Michael Stimson
    Commented Nov 8, 2018 at 4:27
  • I was just going to register the point clouds using the placed targets. The laser scanner is calibrated to measure in meters so everything should be coordinated once the targets have been registered, but the photogrammetry will have to be scaled and rotated as it does not have a inherent measurement standard. If all the targets have coordinates the model should be converted to my coordinate system (hopefully) and hence should be theoretically almost identical. There will be non-linear errors, and i want to see if they are minimal or not, and how cheap can you go before the data isn't useful.
    – William Andrews
    Commented Nov 8, 2018 at 5:15
  • We may be talking about different technologies. A total station, although it uses a laser for ranging and direction, is not a LiDAR scanner. A terrestrial LiDAR scanner sits on a base (tripod), rapidly rotates and scans many thousands of points from which you can generate a wireframe and stretch your pictures onto - like osirisnet.net/3d-tours/e_3d-tours.htm the pictures need be no better than the latest iPhone camera to generate these results but a better camera needs less frames. There are terrestrial scanners that capture RGB points though I have no experience in these.
    – Michael Stimson
    Commented Nov 8, 2018 at 5:55

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.