Need better support for repositioning objects/turntables while scanning

[MiloMindbender]

I started doing photogrammetry using both Meshroom and Apple Object Capture at the same time, both produce really good results but Meshroom is missing one feature Apple has that makes scanning objects like toys/action figures MUCH easier. They support repositioning the object during a scan to make it easy to get photos of "hard to reach" areas like under the arms and legs, and the bottoms of things. Apple's solution supports flipping an object on it's back or side to get photos from angles that are hard to get if the object is sitting still on a table. This works in a situation where the object is on a table and you are walking around it to take pictures, but they also support taking pictures of an object on a turntable (including letting you flip the object on it's side or back too) without requiring any special setup or background.

Describe the solution you'd like I would like to see support for repositioning the object. For getting a good capture of something like a 12" human action figure it is much easier if you can take a series of pictures with the object standing, then flip it on it's back or front to get hard to see areas. Similarly, support for taking pictures on a moving turntable also greatly speeds up workflow. The apple system does this without requiring any special background setup. I believe what the software is doing is solving the camera based on the object being scanned and ignoring background objects

Describe alternatives you've considered The only alternative solution I have seen is to suspend the object being scanned in mid air using some kind of clamp. This allows you to get pictures from all sides, top and bottom but is usually much harder to setup, properly light. and photograph The object must be held securely to prevent motion and the clamping solution is likely to obscure parts of the object you want to photograph. If the object is delicate, suspending it could damage it. Also this requires more post processing of the resulting mesh to remove the clamping rig from the output and patch up the things it covers up.

The approach of using a textured base on a turntable with a flat black or white background appears to be hit-or-miss at present, and it still doesn't handle taking low angle shots of the bottom sides of things.

Additional context Consider the basic workflow with the Apple system. You place the action-figure on a turntable inside a light tent. Start the turntable rotating. Snap pictures using a stationary camera from high, low and middle angles. Then flip the object on it's back and do another series of photos, repeat with the object face down for good measure.

With this workflow you can get a large set of photos in just a couple of minutes, much easier than walking around an object on a table or suspending it in mid air.

Using a light tent gives much more uniform lighting to the photos, it's not really possible to use a light tent if the object has to remain stationary with you moving around, above and below it.