Dimensionality Reduction and Hierarchical Clustering in Framework for Hyperspectral Image Segmentation



The first attempts to analyse hyperspectral images were based on techniques that were developed for multispectral images which only have a few spectral channels, usually less than seven. As a matter of fact, with a limited number of available samples, the performance of any segmentation algorithm on hyperspectral data in terms of accuracies will dramatically be downgraded when the number of data channels increases. In this paper, a new framework is designed for the analysis of hyperspectral image by taking all the data channels. This framework consists of three stages namely- dimensionality reduction, hierarchical image fusion and segmentation. This paper presents a dimensionality reduction method using subset selection and hierarchical clustering in framework for hyperspectral image segmentation. A methodology based on subset construction is used for selecting k informative bands from d bands dataset. In this selection, similarity metrics such as Average Pixel Intensity [API], Histogram Similarity [HS], Mutual Information [MI] and Correlation Similarity [CS] are used to create k distinct subsets and from each subset, a single band is selected.  The informative bands which are selected are merged into a single image using hierarchical fusion technique. After getting fused image, Hierarchical clustering algorithm is used for segmentation of image. The qualitative and quantitative analysis shows that CS similarity metric in dimensionality reduction algorithm gets high quality segmented image.



Image Processing, Hyperpectral images, Hierarchical clustering, Image Segmentation


  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

Bulletin of EEI Stats