In trying to get a Simple scan done by steady camera and rotating platform. Its really wierd the preview shows Something all the time but still the scanner chooses to drop trackig At the midturn where the item is thinnest. Funny how it said scanner will pick up 0.1 mm when it fails to see 10mm and come on guys can’t you implement Something Simple into your software like recognicong you Are using a turn table would be quite easy to detect. Or maby a toc box (doh) why does the software after loosing tescking oviously done with steady cam assume i flipped the model around when the frikin camera is At the same position still.
This product can’t be this horrible.
Here is one example of this
The camera can’t see enough features or markers at that point to stay locked. You need to provide something else in the scene that it can stay locked to while passing the thin section.
Look in at fhis. I talked about the same problem and how I solved it.
One big problem is once the came looses trackin and gets it back it assumed i have filled scanning area magicly with another piece and it desegards the fact the original item is still At its original position and does another instance on top of it. I would assume when you have scanner like 4 instances on top of eachother you could tell hey im now scanning this item but according to my scanning there should be like 4 of them but i see only one. Hmm… should i just assume there is only one item i have scanned and connect the points of each 4 scans into one… oviously their realtive position to another is about the same as there is an instance of dots resenbling the same form just made on top of eachother.
I don’t understand how the software can be this stupid.
Though i have to say i don’t know how software works At these levels
You see you already doing stuff wrong .
You using markers but your choice feature mode for scanning and low resolution at 0.3mm
If you use markers on your model, you need to choice the Marker Mode for scanning so the software can track it proper .
0.1mm accuracy has nothing to do with the dimension of the scanned model as dimension is volume , the scanner can see volume of 20x20x20 mm with feature details at 0.1mm to 0.05mm and it works just fine .
If you have bigger object like yours , you should try to fit it as much as possible in the frame , and not use feature Mode as that is for very small objects , try Markers mode , or head mode for this type of model .
shure i tried marker mode. But i also tried feature mode but i let the markers be there.
The thing that´s really botching the scans that even as it seems the software keeps tracking still it keeps making new instances on top of ecahother.
because the object you scanning has zero features to be tracked , it like you try to scan an egg or orange what is actually impossible without markers , feature mode is for a very small detailed objects with lots of features and details . For the object from your video you have only one option what is using markers and marker mode as nothing else would scan it , not even $10K professional scanner .
Try to stick the markers on the object with a different pattern , like 2 very closer to each other , then next just 1 marker , creating unique pattern for best results . And if the object does not fit in the Depth camera frame try to go as far as you can from it between good and too far. All it takes is practice, practice, practice
Trust me I used so many 3D scanners but no one is the same and need different approach for best results . Keep learning , and start with something simpler with more features before you go into pro surface scanning , as I tell you the object you try to scan need some very good practice . Good luck Janne , I hope to see some good results soon !
