Hello, I feel like the tracking with small marker dots doesn’t work as well or has gotten worse with the new software update.
I’m using the RevoPoint Mini, and I printed hexagon towers where I place my markers. This setup means there’s a 60° angle difference, and the tracking markers are clearly visible. The brands do not have the same pattern, each one is individual. You can see this in the video.
Unfortunately, when I scan with reference points, the tracking fails completely. I’m not sure if this is due to the new update, but here are my questions in general:
What is the maximum angle for a reference point to the camera lens for the scanner to detect it?
Why am I able to achieve better scans with feature tracking than with reference points, given the same setup?
How are the markers detected by the RGB camera or the laser sensor?
Will it be the same with the new Metro x scanner?
Here is the video (double speed): I scan the first half without Markermode and the second half with .
I think it would be important to know technical details like How are markers detected (camera or sensor), how many markers need to be detected and what is the maximum angle (to the the Camera/Sensor).
I just had a discussion with the product software manager.
The conclusion is: the object you are scanning doesn’t have enough features, and by placing markers next to the object, because of the different angles, the scanner may not be able to recognize 5 markers in each frame, which will lead to frame splicing failure.
We recommend that you use this maker pad or place more markers next to the object to ensure that 5 markers are recognized in each frame.
Could you try again after recalibrating? Also, selecting points during measurement that are not exactly the same as the actual position of the measured object can lead to some errors.
I’ll try a recalibration tomorrow and then do another scan. As far as I can tell, these issues didn’t occur in previous versions. However, I still have the following questions:
How many markers need to be visible at the same time? You mentioned five markers; is that correct?
How are the markers detected by the RGB camera or depth camera?
What is the maximum angle of the markers to the scanner for proper recognition?
Is the tracking done solely by the markers, or does it also use the object itself when I use the markermode?
I sometimes need the high markers for tall objects. Unfortunately, the markers on the turntable don’t help me much in that case.
In the scan itself, as seen in the video, at least 6 or 7 markers are always detected.
So a bug that I’ve discovered now is that when the exposure is set to automatic, it sometimes jumps to the maximum and then doesn’t go down anymore, which leads to severe overexposure.
And even after calibrating the scanner, I still have the issues.
Although enough reference points are detected, definitely more than 5.
You can clearly see here how the frame is incorrectly assembled. It should be blocked on the software side to prevent such abnormal angles from being combined in the first place. After all, I can’t rotate my scanner 180° in less than a second.
Thank you Manuel , some of the bugs are already submitted to the Dev Teams by myself .
Another thing, pay attention that you can’t have markers in a row, they can’t be in a symmetrical pattern and so close to each other .
The software works like face recognition, if the markers are too close it don’t recognize it correctly .
Keep proper distance and keep unique pattern .
For the rest of the new bugs report I will tag @Revopoint-Jane
