A Band Grouping Based Hyperspectral Imagery Classification Method with Analysis Dictionary Learning

Mengting Ma, Cong Wang, Lei Zhang, Yanning Zhang, Wei Wei

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Dictionary learning (DL) has became popular in image classification tasks. Due to the discriminative analysis dictionary learning (DADL) model can supply richer feature representations and discriminability, it gradually received extensive attention. Inspired by this, we propose a band grouping based hyperspectral imagery classification method with analysis dictionary learning framework in this study. First, we segment all spectra into several segments according to the spectral correlation. Second, DADL is utilized to represent the data and obtain the sparse representation coefficients. Finally, we employ the k-nearest neighbor (KNN) algorithm to classify the sparse representation coefficients, and the final classification label is obtained by voting the KNN results. Experimental results on hyperspectral imagery classification validated the effectiveness of the proposed method.

Original languageEnglish
Title of host publication2018 IEEE 4th International Conference on Multimedia Big Data, BigMM 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538653210
DOIs
StatePublished - 18 Oct 2018
Event4th IEEE International Conference on Multimedia Big Data, BigMM 2018 - Xi'an, China
Duration: 13 Sep 201816 Sep 2018

Publication series

Name2018 IEEE 4th International Conference on Multimedia Big Data, BigMM 2018

Conference

Conference4th IEEE International Conference on Multimedia Big Data, BigMM 2018
Country/TerritoryChina
CityXi'an
Period13/09/1816/09/18

Keywords

  • Discriminative analysis dictionary learning
  • hyperspectral imagery classification
  • spectra segmentation

Fingerprint

Dive into the research topics of 'A Band Grouping Based Hyperspectral Imagery Classification Method with Analysis Dictionary Learning'. Together they form a unique fingerprint.

Cite this