Natural Visual Perception for (Augmented Reality Workstations

点击翻译按钮获取中文摘要

Objective of this research is a method to enable natural visual perception for Augmented Reality (AR) workstations that facilitates intuitive assembly of virtual mechanical systems. Common AR workstations that are in use in industry present a monoscopic visualization of virtual parts, superimposed on video of physical fixtures or parts, on a large 2D screen or monitor. Although relatively inexpensive and easy to deploy, the monoscopic visualization afforded by these displays cannot support the user¿s natural vision, which is binocular and thus perceived stereoscopically. Despite their benefits, these first-generation industrial AR systems do not provide a natural visual interface. It is well known that natural interaction enhances assembly training tasks. The research proposed here addresses these deficiencies of first generation industrial AR systems by developing a method to simulate visual cues in a 2D workstation. Thus, the user will get a realistic three-dimensional perception of virtual parts. Two new methods will be introduced: 1) view-dependent rendering for AR applications, and 2) simulation of depth cues, to substantially enhance visual perception in desktop rendering in AR applications. As an application field, virtual assembly has been chosen. The hypothesis is that the simulation of depth cues and the addition of the view-depended artifacts will enhance the assembly process, including its accuracy and speed.