Missaligned photogrammetry marker point clouds

Discussions MIRACO Series
1

I’m having trouble with the photogrammetry step. I’m trying to scan a car chassis and after calculation of the photos I’ve taken with the chassis covered in the coded markers and with the scale bars in place I am getting a very large error number (~280mm!) and in the 3d point cloud i’m seeing a small scale cloud inside the large scale. I can find numbered points that exist in both.

I don’t really understand why the software isn’t able to line these up, it knows that there can only be one of each number?

…

I just had another go on a fresh file, Initially I took about 70 pictures and stopped and processed it. All good, and low error number except for some sparse areas where I clearly hadn’t gotten enough angles.

I resumed and took another ~50pictures and after processing that, a second set of points at a different orientation was showing overlapping with the first (although the scale appeared similar this time), the error was somehow still low despite the obvious missmatch.

I resumed and took another 10-20pictures, hoping that maybe it might improve. Instead I got a third orientation of the points on top of the first two, again similar scale and low error number.

Given I scanned in 3 blocks, and ended up with 3 missmatched sets of points, I’m now wondering if it’s not resuming how it used to (there was a software update this morning so maybe that broke it?).

My next attempt will be to try and do it in one, by taking ~270pictures before calculating.

….

Okay, doing it in one go has given me a point cloud that looks pretty good, only one copy of the chassis. scale bar error was reported at 108mm though.

…

Had a go at scanning with that last one and it wont track at all, similar to my first attempt yesterday. I guess because while it looks okay, the scale is drastically off

2

ok having fought with this all day, I’ve come to the conclusion that the overlapping points is probably a bug.

As long as I get a low error number, even if when I’m in the photogrammetry stage it’s showing 3 copies of my chassis (or as many as times I’ve stopped and calculated then continued taking pictures) if I then progress to actual scanning, the marker points look normal.

Not sure what was causing the very high scale bar errors I was seeing, but I got there in the end.

1 Like
3

Hi @luken I usually always calibrate the scanner after the firmware was updated to reflect the new adjusting .

Usually you should not use any stores global markers that was captured before any calibrations as it will not work properly, so if you captured half things then updated and continued, it may create error in the process , not saying that was exactly that . Just giving you some tip.

4

I did try calibrating far mode with the calibration boards between attempts 4 & 5 but it told me it was already calibrated recently (the scanner is relatively new) and it was unnecessary.

Other than my first attempt which was before the update (I couldn’t have predicted it was going to update the next morning), all of these were done on the same version number.

5

In this case , I would suggest to contact the technical support directly as they can find the exactly solution in this case .

For me from here is just guessing and really waste of our time talking about trying.

Please contact customer@revopoint3d.com

You can include the link to this thread as reference to avoid typing all over again .