Assessing Binocular Advantage in Aided Vision

摘要

Advances in microsensors, microprocessors and microdisplays are creating new opportunities for improving vision in degraded environments through the use of head-mounted displays. Initially, the cutting-edge technology used in these new displays will be expensive. Inevitably the cost of providing the additional sensor and processing required to support binocularity brings the value of binocularity into question. The value of binocularity in human vision is well documented, yet the value of binocularity in sensor/display systems for aided vision is controversial. Several assessments comparing binocular, biocular and monocular head-mounted displays for aided vision have concluded that the additional performance, if any, provided by binocular head-mounted displays does not justify the cost. It is possible that the human binocularity advantage doesn't carry over to the aided vision application, but more likely the experimental approaches used in the past have been too coarse to measure its subtle but important benefits. Evaluating the value of binocularity in aided vision applications requires an understanding of the characteristics of both human vision and head-mounted displays. With this understanding, the value of binocularity in aided vision can be estimated and experimental evidence can be collected to confirm or reject the presumed binocular advantage, enabling improved decisions in aided vision system design. This paper describes four models; stereopsis, modulation transfer function area for stereopsis, probability summation and binocular summation, that may be useful in assessing the advantage of binocularity in aided vision.