Hi there,
We are using the 3-month Promo version to do a full 3D reconstruction of biological specimens (i.e. including all sides of the specimen). During image acquisition, each object lies on a turntable so that the base of the objects is not visible. Therefore, to obtain images of the object’s base, we have to flip the object over between different sets of images. To merge the different image sets into one model, we use control points (after masking the images to remove the background).
Problem 1: Despite a good level of overlap between the sets different, components are created for the different sets of images. With some models this works very well, but with others it doesn’t and we have noticed a difference between the demo version and the 3-month Promo version, e.g. a model we created successfully in the demo version in no time, whereas using the same masks in the licensed version does not lead to an appropriately merged model. Could you advise how to use control points more effectively (e.g. how many, where to place them,…) or what else might be the issue?
Our workflow in Capturing Reality: import folders, align images, set 3-7 control points (until model merges or if nothing happens on the last ones abort project and start over; using approximately 20 images per control point), reconstruct in normal detail, unwrap, texturise, export as obj.
Furthermore we use the models to create animations in 3DS Max.
Problem 2: The quality of the models decreases significantly when opened with 3DS Max (some areas full of tiny spikes in the texture creating shadows all over the model, other areas overly smooth). Are there settings in Capturing reality that could help maintain the quality of the exported model? Neither the Simplifying nor the Smoothing tool appear to help.
Thanks in advance.