Abstract
Head-mounted displays (HMDs) generally exhibit significant image distortion, which must be reduced/eliminated prior to effective use. Additionally, biocular or binocular near-eye displays must be carefully aligned to enable overlapping two- or three-dimensional image synthesis without causing eye strain, fatigue, or performance loss. Typically, HMDs include distortion correction maps supplied by the manufacturer that are often generated by theoretical calculations that do not precisely match the as-built optical system or account for manufacturing variance. However, HMD users often assert that manufacturer-supplied distortion maps are not accurate enough for some alignment-critical applications. In this work we present the design and validation of a relatively low cost alignment and distortion characterization toolset (hardware and software) for characterization of biocular HMDs. This toolset is able to replicate the ocular alignment of most human observers by emulating a user’s ocular position to examine both on- and off-axis distortion and alignment over a wide range of viewing angles and eye positions. This enables accurate characterization of distortion changes experienced as a user’s eyes move to view different regions of the display (e.g., viewing off-boresight symbols in a well-aligned HMD or viewing a new alignment after an HMD has “slipped” to a slightly different position). The toolset characterizes distortion through image registration of distorted patterns displayed in the HMD to undistorted reference patterns. This work is intended to be of interest to HMD manufacturers, vision scientists, and operators of biocular HMDs for use in precision-critical applications.
