I'd like to know more about how the motorcycle scan was done. So far I've found the best results come from using the turn table which allows the unit to seamlessly stitch the scans together correctly, but clearly that doesn't work for larger items.
I've had some difficulties with a car fender I'm trying to scan in getting the scans aligned. I'd like to understand the approach used by the folks that built the thing and others who've been successful using the THREE to scan large objects.
Hi Andrew,
Larger scans can be a bit tough for sure. The Scooter was scanned by setting everything to low and scanning from a distance to get as much coverage as possible. Then, I manually aligned each scan to get a low quality but complete scan. Afterwards, I brought the scanner in closer and set the quality to medium to scan the details. After aligning the higher details to the lower, I simply deleted the low quality parts and finished the mesh. It was difficult and time consuming for sure, but I wanted to see how far I could push the scanner.
The Scooter had lots of detail and geometry to match to, which means markerless tracking worked great for the most part. However, with your fender this might be a bigger issue because it's smooth and lacks details, which are important parts to getting the alignment to work well.
In the future, we will be working on a large object workflow, that will also support markers for tracking. But for now, my suggestion is to put some really large objects around the fender to help in the scanning (something like paint cans, boxes, or tools. Make sure these objects are part of the scan and don't move with respect to the fender. This should help with alignment.
If you need to scan both sides of the fender, then we have to talk more about how the alignment process works. Let me know if you'd like me to go into detail with that.
Thanks Drew. The lack of features was definitely an issue. I spent some time over the weekend chipping away at the process and got a reasonable degree of success. I used some high-tech adhesive markers aka balled up bits of masking tape on the fender itself which proved to be quite effective, though will need some editing once all the scanning is complete and aligned.
When I merged the aligned scans, the end result looked like, in the words of one of my kids, Han Solo in carbonite. The fender was surrounded by a large square of extraneous filler, like being shrink-wrapped to a backer. I don't believe I had any random points in space - I'd deleted all of the background data. I didn't work this process though so ot might be something that just requires a bit of experience and experimentation on my end.
Will upload some scan images later for reference.
