A single shot automatic optical inspection system for evaluating the projection distance, astigmatism, field curvature and geometric distortions of a virtual image created by a Head-Up Display (HUD) prototype is presented. The system uses plenoptic or light-field imaging. Although results are only presented for a Augmented Reality (AR)/Mixed Reality (MR) HUD application, the inspection workflow is also applicable to a plurality of other optical systems for AR, MR, and Virtual Reality (VR) applications. The evaluation of the quality of a latent image (in most cases a virtual image) is necessary in all these systems, both for product development as well as for end-of-line production quality control. Therefore the results here presented should have an appeal for a broad readership.
For the HUD virtual image, the projected distance distribution functions are presented, for two selected patches in the FOV (Area 1 and 2). These are well described by Gaussians. As representative values, we obtained a mean projected distance of 0.15 ± 0.02 (horizontal details) and 0.03 ± 0.02 (vertical) Diopters, for Area 1, and 0.33 ± 0.04 (horizontal) and 0.03 ± 0.02 (vertical) Diopters, for Area 2. The estimated astigmatism is 0.12 and 0.30 Diopters, for Areas 1 and 2, respectively. When expressed in meters, the projected distances are 7 1 (horizontal) and 28 9 (vertical), for Area 1, and 3.1 ± 0.4 (horizontal) and 30 ± 10 (vertical), for Area 2. The estimated astigmatic difference is 21 and 23 meters, for Areas 1 and 2, respectively.