I came across this video the other day in which the guy made a complete model of a crocs in one single scan :
I tried the method on a cube which I have made for checking the accuracy of my POP3 and I was delighted to see that it does produce better results.
To get the complete model of the cube, I need to do at least two scans. One with the top facing up and the other facing down. Merging the two scans always produces ugly , scar-like seam lines unless the rough edges around the shadows in the fused models are trimmed away. Currently there is a known software bug making it very difficult to do so ( Revo Scan 5.4.3 unintended cut through) .
This is the result of merging two seperate scans :
Normally I always recommend to trim the edges before merging and eliminate too great overlaps , but the edges are not right the way the scan produce it from the start and the seam on the merged scans is a definitely a bug that should not take please while merging .
It was fixed once but resurfaced again .
Scanning objects in one run is of course possible and works best with organic objects with lots of features .
No matter it is full body scan or organic figurine ,I always scan it in one run , however I prefer to scan none organic objects with different angles and merge it , until the bug is fixed it is rather problematic job so cutting the edges is nesesery what in some situation will cause issues with overlapped areas.
Regarding your original topic question ? Professionally you should process individual scans at different angles and merge them for none organic models , but until the bug is fixed it is rather a task .
the seam on the merged scans is a bug ? do you mean without this bug, the rough edges of individual scans will not show up on the merged model ? no manual trimming required ?
Normally no edges trimming was necessary in the past beside lose points or overlapped points cleaning .
There was no seams either or cleaning nesesery after merging , the current merging algorithms are not the best , not exactly a bug , but bad choice in algorithms.
With Revoscan MetroX, the merging algorithm is still worse. While merging PCs seems to produce nice results, once you mesh you get ugly seam lines and rough orange peel surfaces. PUTV, can you say something about this?
Noticed this today. Spent so long testing the scanning portion, but the merge function appears broken. The merge will appear good enough when aligning the point clouds, but upon merging it looks like one of them shifts.
EDIT: After some experimentation, it seems like it’s Global Marker scans that aren’t merging correctly. Hm.
I did not merged yet anything with the new MetroX software, thank you for the tip, I will investigate tonight and forward to the team for more investigation.
The latest beta was ok, but another public version was released so maybe something was overlooked since I did not tested the public version .
Hi Daniel , the meshes results makes no sense vs fused cloud point
Was that after merge of 2 different scans ?
Regarding the edges, if you don’t capture the edges fully and missing data is there there is nothing you can actually do to make it better , that’s the capturing algorithms .
Sometimes when I have very fine edges I put it flat on s white strip paper so the edge angle is better captured then after I cut off the paper scan portion living me smooth edge but it depends of the object of course.
The trick here is to make the edge scanned fully without missing data .
Early algorithms in RS 4.2 did not had that issues but we discussed it for so many time and I reported it endless times already.
Thanks for your opinion. I should maybe schedule an online meeting with revopoint staff to discuss this issue. The only workarund for now is scan all in one with paus and resume, but that can be tricky, too, depending on the object.
