|
Type of Document Thesis Author Yu, Hyun Geun URN etd-11152004-120254 Title Morphological Image Segmentation for Co-Aligned Multiple Images Using Watersheds Transformation Degree Master of Science Department Electrical and Computer Engineering, Department of Advisory Committee
Advisor Name Title Rodney G. Roberts Committee Chair Anke Meyer-Baese Committee Member Simon Y. Foo Committee Member Keywords
- Region Merging
- Image Segmentation
- Morphology
- Watersheds Transformation
- Co-Aligned Images
Date of Defense 2004-11-05 Availability unrestricted Abstract Image segmentation is one of the most important categories of image processing. The purpose of image segmentation is to divide an original image into homogeneous regions. It can be applied as a pre-processing stage for other image processing methods. There exist several approaches for image segmentation methods for image processing. The watersheds transformation is studied in this thesis as a particular method of a region-based approach to the segmentation of an image. The complete transformation incorporates a pre-processing and post-processing stage that deals with embedded problems such as edge ambiguity and the output of a large number of regions. Multiscale Morphological Gradient (MMG) and Region Adjacency Graph (RAG) are two methods that are pre-processing and post-processing stages, respectively. RAG incorporates dissimilarity criteria to merge adjacent homogeneous regions.In this thesis, the proposed system has been applied to a set of co-aligned images, which include a pair of intensity and range images. It is expected that the hidden edges within the intensity image can be detected by observing range data or vice versa. Also it is expected that the contribution of the range image in region merging can compensate for the dominance of shadows within the intensity image regardless of the original intensity of the object.
Files
Filename Size Approximate Download Time (Hours:Minutes:Seconds)
28.8 Modem 56K Modem ISDN (64 Kb) ISDN (128 Kb) Higher-speed Access HyunThesis.pdf 3.06 Mb 00:14:09 00:07:16 00:06:22 00:03:11 00:00:16