3D Imaging Techniques

by

TABLE OF CONTENTS

CHAPTER 2: LITERATURE REVIEW1

Technological Transition: from 2D to 3D1

3D Imaging6

Evolution of 3D imaging technology7

Types of 3D imaging techniques10

Stereoscopy10

Autostereoscopic Viewing14

Holography15

Volumetric 3D displays16

3D Holoscopic Imaging18

REFERENCES22

BIBLIOGRAPHY25

CHAPTER 2: LITERATURE REVIEW

Technological Transition: from 2D to 3D

When some analysis needs to be performed on real 3D objects and their environments, a 3D reading needs to be acquired. 3D model reconstruction is a way to achieve this. 3D model data can be used for multiple things from simulations, virtual environments and accurate mapping of environments in robot navigation. There arises a need for innovative 3D model reconstruction systems since existing systems can be relatively slow, expensive and cannot easily adapt to multiple environments. A complete system that can perform 3D model reconstruction of objects requires multiple modules including gathering data, calibration, reconstruction, data verification and data correction. Each module may have multiple possible implementations. These implementations will vary in effectiveness, performance, resolution and accuracy [1]. An important common trade-off that must be considered for each module is the one between the quality of results and the time required to achieve them. Furthermore, some techniques are limited by their surrounding environment and the location of the objects of interest with respect to the reconstruction system.

The exponential increase in the computation power of computers and modern workstation has enabled the proliferation of three dimensional interfaces for a wide spectrum of application domains including entertainment and gaming, education and training, data visualization, teleportation, among others. However, one of the key challenges in 3D-based interface development is the ability to present a 3D object in a 2D display (a flat computer screen). That is, the ability of users to depict spatial relationship among objects in 3D space with respect to the depth dimensions (so called depth perception). Depth perception is defined as the ability of humans to perceive the world in three dimensions. Depth perception arises from a variety of depth cues that are traditionally acquired visually. In real environments, depth cues such as perspective, size difference, occlusion, shadow, accommodation, convergence and disparity, provide higher comprehension of the environment which results in a significant increase in the quality of performance. Visual depth cues can be classified into either monocular (requires input from one eye) or binocular (cuts that require input from both eyes).

Many techniques can be implemented to estimate the shape and position of objects in a scene. This paper aims at researching and developing strategies that will be best suited for 3D modelling over different objects or scenes. The solution needs to be able to gather data with varying levels of density while only sacrificing processing time when it is deemed necessary or when needed. Many techniques will arguably produce superior results but will either require a higher amount of processing or more expensive setups and equipment to achieve them. The resulting 3D sensing technique should give results that are accurate and can be interpreted readily, arid are achieved in an acceptable amount of time while operating from affordable off-the-shelf equipment, unlike ultra specialized 3D ...