Hyperspectral Image Analysis using Principal Component Analysis and Siamese Network

The process of collecting information in hundreds of images having bands with different wavelengths using remotely sensed devices is called Hyperspectral Imaging Technology. So many methods are used for analysis of this hyperspectral information of which Classification is a popular one, which is the...

Full description

Saved in:
Bibliographic Details
Published in:Turkish journal of computer and mathematics education 2021-04, Vol.12 (7), p.1191-1198
Main Authors: Gowtham, B, Reddy, I Ajay Kumar, Reddy, T Subba, Harikiran, J, Chandana, B Sai
Format: Article
Language:English
Subjects:
Artificial neural networks
Classification
Data augmentation
Datasets
Hyperspectral imaging
Image analysis
Pixels
Principal components analysis
Principals
Remote sensing
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The process of collecting information in hundreds of images having bands with different wavelengths using remotely sensed devices is called Hyperspectral Imaging Technology. So many methods are used for analysis of this hyperspectral information of which Classification is a popular one, which is the process of assigning label to each pixel. For classification of hyperspectral data, many supervised methods are proposed in literature with excellent performance, but the quality the classifier depends strongly on the number of training samples used to construct the model. Data Augmentation (DA) is a strategy that can increase the quantity of training data and effective to overcome the limited training samples problem. The hyperspectral data set suffers with curse of dimensionality with redundant information, influencing the classification accuracy. In this paper, a supervised model for hyperspectral data set is proposed, with dimensionality reduction using Principal Component Analysis and data augmentation using mixed pixels is applied to increase training samples and tested using Siamese network and finally classification is done using CNN.
ISSN:1309-4653