Share Email Print

Optical Engineering

Local combination histogram based segmentation of Landsat Thematic Mapper/Enhanced Thematic Mapper Plus images using clustering aggregation
Author(s): Wei Yang; Kun Hou; Fanhua Yu; Tieli Sun
Format Member Price Non-Member Price
PDF $20.00 $25.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

We present a new segmentation method for multispectral remote sensing imagery using the K-means clustering aggregation coupled with local combination histogram (LCH). First of all, the bands are partitioned into several nearly uncorrelated subsets. Then, some band combinations of the multispectral images are generated from the subsets. After that, the LCHs are computed from quantized images in each band combination. The LCH represents a pixel in the sense of both spectral information and neighborhood spatial information implicitly and serves as the input feature. Identical K-means procedures are employed to get several relatively coarse segmentation maps. The final K-means procedure refines these intermediate segmentation maps to achieve the final results. The segmentation results are evaluated and compared with other multispectral image segmentation methods by visual inspection and object-based image classification. Experimental results show that the proposed method can achieve more accurate segmentation maps and higher classification accuracy.

Paper Details

Date Published: 23 April 2012
PDF: 11 pages
Opt. Eng. 51(4) 046201 doi: 10.1117/1.OE.51.4.046201
Published in: Optical Engineering Volume 51, Issue 4
Show Author Affiliations
Wei Yang, Northeast Normal Univ. (China)
Kun Hou, Northeast Normal Univ. (China)
Fanhua Yu, Changchun Normal Univ. (China)
Tieli Sun, Northeast Normal Univ. (China)

© SPIE. Terms of Use
Back to Top