Back to beginning, shoot 20 images on 360 angle, then load to PhotoScan JUST 3 of them, try do mesh = OK im getting something
Phase two, try the same photos BUT add 2 more, - OK im getting more of what i need
phase three ad another 3 photos - OK im getting preciously what i nedd..
Is that simple, start with litle and the add more IF eventely needed, is from object to object.....
im testing iot on buildings and im getting wery useful output with around 8-10 photos, later on will add eventuely more, but every time is better to start with lower number of photos.....