Hello!
I just downloaded ReCap Pro and have imported 23 raw scan files (taken with a Faro Laser 3D scanner) of a Terraced home in London. I'm testing Recap as an alternative to Faro Scene software and I was just wondering if someone could shed some light on the scan registration accuracy?
- I have imported the files
- Clicked Register Scans
And have come to the point where I get two screens, one of Scan001 and the other with Scan002 where it asks me to select points in the same region (so I pick a plane in Scan001 and the same plane in Scan002 and repeat 2 more times to select 3 corresponding points.)
Once I select my three points I get the "refining" screen where it aligns the scans, and the diagram showing Overlap, Balance and Points <6mm. Can you tell me if 6mm is the most accurate ReCap can align these points or if thats just a general figure Autodesk has put in?
Any help/guidance would be much appreciated!
Nayeem
Greetings Nayeem,
The <6mm is just a "bucket" showing the percent of overlapping points below the 6mm threshold. If you mouse over it, you will see a complete histogram. In the "<6mm" bucket, most overlaping points will be much closer. Keep in mind, this is a statistical way of representing accuracy. If you want a more rigourous evaluation of accurracy, survey control is the standard approach. You will need to establish three or more benchmarks (typically targets) and survey them. Once surveyed, you can locate the targets in ReCap Pro and assign survey values to them. ReCap Pro can then produce a report that states how the survey compares to registration.
In normal building interior scenarios, we see between 2mm to 6mm accuracy. This should be comparable to other methods of registration.
Aaron
Aaron, ive attached two images below, the bottom one is the result of registering two scans which i wrote about in my other post, the top image highlights the markers I used. In Scene it registers properly but in ReCap the scan is rotated. How can i fix this (manually?)?
Cheers
It appears the overlap in this situation is less than ideal.
First, you will have to remove the scan from the registered data set.
Second, given the low overlap, register any scans visible from this scan before this scan.
Third, when you go to register this scan, use the panoramic views and select features located on flat features. You might have to select a few times if the first couple of matches do not alighn well.
If you try this and are still unable to register, please contact me directly to discuss.
Best,
Aaron