NJIT
Stereo matching algorithm by propagation of correspondences and stereo vision instrumentation
Abstract
dc:description.abstractA new image processing method is described for measuring the 3-D coordinates of a complex, biological surface. One of the problems in stereo vision is known as the accuracy-precision tradeoff problem. This thesis proposes a new method that promises to solve this problem. To do so, two issues are addressed. First, stereo vision instrumentation methods are described. This instrumentation includes a camera system as well as camera calibration, rectification, matching and triangulation. Second, the approach employs an array of cameras that allow accurate computation of the depth map of a surface by propagation of correspondences through pair-wise camera views. The new method proposed in this thesis employs an array of cameras, and preserves the small baseline advantage by finding accurate correspondences in pairs of adjacent cameras. These correspondences are then propagated along the consecutive pairs of cameras in the array until a large baseline is accomplished. The resulting large baseline disparities are then used for triangulation to achieve advantage of precision in depth measurement. The matching is done by an area-based intensity correlation function called Sum of Squared Differences (SSD). In this thesis, the feasibility of using these data for further processing to achieve surface or volume measurements in the future is discussed.
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Electrical Engineering - (M.S.)
- Discipline thesis:degree_discipline
- Electrical and Computer Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Witoonchart, Peerajak
- Contributors dc:contributor
-
- Stanley S. Reisman
- Yun Q. Shi
- Richard A. Foulds
Subjects
dc:subject × 4Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.njit.edu/theses/701