Hi Olena,
To my knowledge, the area targets are basically a point cloud. Usually a device with a LiDAR sensor is used to create these. The outcome of a scan is the interpretation of the laser beams being sent and returned to the sensor. The reflection is dependent of a few factors like material, light transmission, distance, etc. Most of these factors will not be included, when trying to interpret CAD data.
The accuracy is at first dependent off the scanner’s resolution. For this I would check the technical documentation of the scanner. Second will be the data management within the import process, it might be compressed. Third is the treshold of the device, the viewer app will run on. There must be a huge generosity, otherwise even the focal length of the camera lens will prevent the app from tracking.
More information can be found here:
https://library.vuforia.com/features/environments/area-targets/best-practices-for-preparing-and-scanning-an-environment.html
Please note: This is the documentation for the Vuforia Engine and will not match 100% to Vuforia Studio
