الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
This research presents an FPGA-based disparity map computation. The overall objective of this thesis is to provide information about the geometry of the surrounding environment, by knowing the 3D information. Accomplishing this task easily implies that one of the functions of the human visual system is to reconstruct a 3-D representation of the world from its 2-D projection onto our eyes.
Stereo vision deals with images acquired by a stereo camera setup, where the disparity between the stereo images allows depth estimation within a scene. 3D information, hence, is retrieved which is essential in many machine vision applications; 3D reconstruction, positioning, mobile robot navigation, obstacle avoidance, etc. Disparity map extraction of an image is a computationaliy demanding task.
Previous work on disparity map computation IS mainly limited to software based techniques on general-purpose architectures which are not good candidates for real time applications. In this thesis, the proposed algorithm has been implemented to be a good candidate for real time applications and a good trade off in terms of accuracy computational efficiency.
To conduct this objective, ID optimization based on dynamic programming stereo has been implemented and developed to obtain the proposed algorithm because it can reconstruct an observed 3D optical surface very quickly and thus has potential for real-time applications. While being efficient, its performance is far from the state of the art because the vertical consistency between the scanlines is not enforced.ID optimization based on dynamic programming for stereo correspondence is re-examined by applying it to the vertical consistency between the scanlines as opposed to the individual scanlines.
The proposed disparity map usmg suboptimal cost with dynamic programming algorithm has been evaluated on the benchmark Middlebury database. Enhanced performance is achieved by using the proposed algorithms in comparison to the previously presented algorithms.
Finally, FPGA-based implementations of the sum of absolute difference algorithm are presented at a very high frame rate to be adequate for most real-time vision applications.