What kind of strategy and material should I use to georeference a point cloud. I use FARO M70 as scanner Lazer terrestre and SCENE 7.0 as software?
I always used Leica Cyclone for that job, because I had at disposal topographic sparse cloud that I imported into the application. The main issue is the one to give the same ID to homologous points (on sparse and dense clouds) I'm sure you can find similar tools inside Scene
Hi,
I also use Leica Cyclone, but if you don't have this software at your disposal, you can consider using CloudCompare. CloudCompare is free software that works really well. You can download the software from http://www.danielgm.net/cc/release/
You can find tutorials and videos on this website:
http://www.cloudcompare.org/doc/wiki/index.php?title=Main_Page
Hi Brahim,
personally I feel very comfortable with Faro Scene, and if you have at your disposal the SW with your Faro M70 I highly recommend it to you.
If you don't have it, or if you prefere open source platforms, yes, you could use CC that is very user friendly! What do you mean exactly with " material should I use"? If you have topographicaly measured some targets/spheres etc.., but you also performed a good laser acquisition with a suitable coverage, as well, I recommend to you to firstly align the scans each others with ICP-solution, and then, if good, use target-based registration to check the registration quality as control points and to georeference the block. In this sense, SCENE offers effective tools for it, and also Cloud Compare offerts ICP algorithm implementation together picking points-based clouds aligment!
Have a good job!
Thank you @Giulia for the advices. What I meant by saying materials or strategy, so I've done the topographic measures with a GPS station and I have the XYZ points, but still can't align the blocks I don't know why. I would know if there is some other strategies or material to work with, I'll try what you said and see the result. The project area is about 5 hectares, it's so big.
Regards.
Hi Brahim,
In TLS processing, there are two (quasi-separate) steps: the registration of the stations and the georeferencing of the said registered block towards a certain coordinate system (e.g. your GPS measurements). When you said that you didn't manage to align the blocks, did you mean that the registration step failed? To perform good registration, well distributed tie points between stations are required. This can be in the form of artificial spheres, printed targets, or natural points that you can click manually in SCENE.
In terms of georeferencing, again a well distributed network of control points are required throughout the project area in order to ensure a homogeneous quality.
In any case, even if your registration failed (e.g. if you have multiple blocks), as long as you have at least 3 GPS control points for each block, you should still be able to georeference your blocks separately. Since the blocks are georeferenced to the same GPS coordinate system, by the time you export them and visualise them (for example in CloudCompare), you should have all the stations where they are supposed to be (i.e. georeferenced). I actually use this method sometimes to integrate my TLS data with those from photogrammetry and even other sensors.
Hope this helps.
Best,
Arnadi
Let me answer this question from the underlying complexity of point could. In general, automated 3D ponts collection by lidar or the like technologies is a random sampling process. In other words, we cannot guarantee that we could obtain identical point sets from the scanning of the same object at different epochs. In particular, we cannot easily find a one-to-one relationship between the point cloud in two different sets. Yes, this process can be approximated under dense points measurement.
In general and under rigid body assumption the relationship for georeferencing between two sets or more of point cloud can be realized by a 3D similarity transformation (3 angles, 3 translations, and a global scale factor). Indeed, this model can be replaced by a more comprehensive one to account for articulated objects.
One of the most general approach for georeferencing is to segment one data set, for example, into planar object and leave the second set of point cloud in their original format as points. Then build a mathematical relationship between the planar objects in the first data set and their corresponding points in the second data set. As such, the relationship in the georeferencing process will be transformed into one-to-many.
- Badges
- Science topic