An unsupervised marker image generation method for watershed segmentation of multispectral imagery

Watershed transformation in mathematical morphology is a powerful tool for image segmentation. Watershed transformation based segmentation is generally a marker-controlled segmentation. This paper proposes a novel method of maker image generation based on unsupervised ISODATA classification. The met...

Full description

Saved in:
Bibliographic Details
Published in:Geosciences journal (Seoul, Korea) Korea), 2004-09, Vol.8 (3), p.325-331
Main Authors: Li, Peijun, Xiao, Xiaobai
Format: Article
Language:English
Subjects:
Classification
Genetic transformation
Image processing
Image segmentation
Markers
Mathematical analysis
Mathematical morphology
Methods
Segmentation
Spectra
Transformations
Watersheds
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Watershed transformation in mathematical morphology is a powerful tool for image segmentation. Watershed transformation based segmentation is generally a marker-controlled segmentation. This paper proposes a novel method of maker image generation based on unsupervised ISODATA classification. The method incorporates spectral information contained in multispectral data into watershed transformation for image segmentation. The method is evaluated and compared to existing unsupervised method. The results show that the proposed method could effectively reduce the oversegmentation effect and achieve more accurate segmentation results, compared to existing method. The combination of segmentation result and pixed-based classification is found to improve the overall classification accuracy over pixel-based classification.
ISSN:1226-4806
1598-7477
DOI:10.1007/BF02910252