Advances in Natural and Applied Sciences
Year: 2020
Type: paratext
Abstract: High dimensionality of the Hyperspectral data makes it more difficult to use it for classification. In this paper, unsupervised band selection method is used to reduce the dimensionality of Hyperspectral image. Linear Prediction Algorithm selects most informative bands from the Hyperspectral image band set, without changing the physical properties of the image. Spatial features are extracted using Morphological operators. Here selected bands are used as spectral features. The combination of spectral and spatial features is given as input to the classifier. To overcome the difficulty of high dimensionality of resulting features, it is a common practice that Morphological Profiles (MPs) are extracted from selected bands. It can improve classification because they contain more critical characteristics for classification. Support Vector Machine (SVM) used as a classifier to get better accuracy for all classes.
DOI: 10.22587/anas
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W3140372349
