I’ve got a project with 265 images from a Zenmuse P1 on M350. GCPs are Aeropoints. RTK enabled and processed with Propeller PPK as usual. Everything looks great. Overlap was set to 80%.
This is my first project with Matic 2.0. The calibration is excellent. 100% of images calibrated. The reported checkpoint position error is less than 0.01 feet. The imagery appears to be very good.
Problem: my dense point cloud is almost non existent. I have more ATPs than I have point cloud points. The .laz export is less than 5mb when I’d expect it to be more like half a gigabyte.
In the screenshot you can see my pointcloud processing settings and yes, that is the point cloud you can kindof see. This isn’t a tough site or anything. I expected excellent results and this is what I get.
Am I missing a new setting or something in 2.0?
Hello,
It appears the point cloud is not currently visible in the 3D view. Could you please check if the point size is set too small? You can verify this in Settings → Sizes → Point size.
Additionally, please ensure that the dense point cloud visibility hasn’t been toggled off. If the issue persists, could you share a screenshot of what is displayed in the Content → Point clouds panel?
Looking forward to hearing from you.
Kind regards, Alexey
Alexey, thank you for the response.
The point cloud is displayed, which is the problem.
I’ve got the point size turned all the way up to 200.
I’ve been going back and forth with support on this. I’m currently reprocessing it from scratch. I am able to get a good point cloud without the GCPs and PPK image tags. I’m about to reintroduce the GCPs and reprocess. If that’s successful, I’m going to reintroduce the PPK image tags.
I’ve completed many projects with both mapper and matic with the exact workflow I’m using now. This is frustrating.
Hello,
Thank you for your patience. We have finished investigating your project with our support team.
It looks like the issue was triggered by a slight mismatch in the coordinate reference system (CRS) during the image import (specifically using EPSG:6598 instead of 6599). While this was just a small typo, it caused the software to shift the “center” of your project an enormous distance away.
Although you corrected the CRS immediately, the software unfortunately didn’t “re-center” the internal scene origin after the fix was applied. Because the project was still trying to calculate data at a massive distance from the origin, it ran into mathematical precision errors. This caused software to incorrectly identify valid points as duplicates and discard them, resulting in the sparse point cloud you saw.
I am filing a bug report with the development team to ensure that correcting a CRS typo automatically triggers a re-centering of the scene to prevent this from happening in the future.
In the meantime, to get a clean result, I recommend recreating the project from scratch with the correct CRS and reprocessing it.
Thanks for your patience while we tracked this down!
Best regards,
Alexey
1 Like
