Computer Graphics is now a huge industry. Computer Graphics (CG) applications are now used across a wide range of industries, including the more obvious, such as film and TV, manufacturing and the defence sector, through to the less well known such as specialist medical units.
CG has dominated the film industry in recent years, with a greater proportion of movies being made that use computer generated special effects compared with those that contain none at all. Even movies that look as if they have no special effects are likely to have been digitally manipulated or enhanced in post-production. From a director’s point of view, the use of computer graphics gives an endless creative power, in what would normally be an impossible shot to make. The downside is that the cg elements have to be added after the scenes have been shot, a process that can take many months to complete. It is not until the cg and live action footage are merged together that the final result can be viewed and is hopefully how the director envisaged it.
The ability of computers to handle complex tasks has grown in parallel with their processing power and, in particular, their ability to perform graphically intensive operations in real-time has improved so much that previously impossible cg scenarios are now both feasible and commonplace.
An exciting development within the computer graphics arena is the field of Computer Vision. This is the ability of a computer to be able to recognise key points or extract specific features from both still images and video footage. The use of Computer Vision for reading still images is used in optical character recognition (ocr), where a computer is able to translate handwritten characters into computer text. Facial recognition is another commonly used area of computer vision with still images. However, recognition of moving images is a much less developed field. One such example is a police pedestrian monitor, used to identify potential criminals by their unlawful activities [DAVIES 2005].
Augmented Reality is a fast growing field within Computer Vision, where the use of computer graphics and live footage has taken advantage of the advances in processing power. Augmented Reality is a process in which computer graphics are placed over live footage so that the cg elements look as if they are integral to the scene. The use of Augmented Reality could, in the not too distant future, help directors to visualise how the computer graphic elements will interact in the scene during the actually shooting.
This Masters Thesis describes the process of developing an Augmented Reality C++ development kit called ARLib. The ARLib toolkit enables developers to create Augmented Applications quickly and easily. The applications work by looking for predefined black and white markers in the video footage and then placing computer graphics over the top of them. The entire process has been broken down into separate modules, each one handling a specific task.
As a means of placing this AR toolkit into context within the AR marketplace, Chapter 2 details how AR applications are being used today, and by whom.
Chapters 3-6 cover how the ARLib was developed, and what models and methods were used in its implementation.
Chapter 3 describes the first stage of performing Augmented Reality by taking a colour image and converting it to black and white so that features within it can be detected.
Chapter 4 details the process of extracting useful features, ie; black and white markers, from the black and white image.
Potential markers in an image are then further analysed to determine whether it is a valid marker. This process is described in Chapter 5.
The final stage of the Augmented Reality process involves determining where the marker is in relation to the camera and is detailed in Chapter 6.
Chapter 7 describes how to create applications using ARLib. The C++ code for two working examples is detailed, which are included as part of ARLib.
Finally, the conclusion and recommendations for further developments are discussed in Chapter 8.
| 2. Previous and Current Work >> |