A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification
This paper assesses the performance of DoTRules—a dictionary of trusted rules—as a supervised rule-based ensemble framework based on the mean-shift segmentation for hyperspectral image classification. The proposed ensemble framework consists of multiple rule sets with rules constructed based on diff...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2019-09-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/11/17/2057 |
id |
doaj-44bb886977874555ac356c4047f9c53d |
---|---|
record_format |
Article |
spelling |
doaj-44bb886977874555ac356c4047f9c53d2020-11-24T20:42:55ZengMDPI AGRemote Sensing2072-42922019-09-011117205710.3390/rs11172057rs11172057A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image ClassificationMajid Shadman Roodposhti0Arko Lucieer1Asim Anees2Brett A. Bryan3Discipline of Geography and Spatial Sciences, School of Technology, Environments and Design, University of Tasmania, Churchill Ave, Hobart, TAS 7005, AustraliaDiscipline of Geography and Spatial Sciences, School of Technology, Environments and Design, University of Tasmania, Churchill Ave, Hobart, TAS 7005, AustraliaSchool of Engineering, University of Tasmania, Churchill Ave, Hobart, TAS 7005, AustraliaCentre for Integrative Ecology, School of Life and Environmental Sciences, Deakin University, 221 Burwood Hwy, Burwood, VIC 3125, AustraliaThis paper assesses the performance of DoTRules—a dictionary of trusted rules—as a supervised rule-based ensemble framework based on the mean-shift segmentation for hyperspectral image classification. The proposed ensemble framework consists of multiple rule sets with rules constructed based on different class frequencies and sequences of occurrences. Shannon entropy was derived for assessing the uncertainty of every rule and the subsequent filtering of unreliable rules. DoTRules is not only a transparent approach for image classification but also a tool to map rule uncertainty, where rule uncertainty assessment can be applied as an estimate of classification accuracy prior to image classification. In this research, the proposed image classification framework is implemented using three world reference hyperspectral image datasets. We found that the overall accuracy of classification using the proposed ensemble framework was superior to state-of-the-art ensemble algorithms, as well as two non-ensemble algorithms, at multiple training sample sizes. We believe DoTRules can be applied more generally to the classification of discrete data such as hyperspectral satellite imagery products.https://www.mdpi.com/2072-4292/11/17/2057image classificationensemblemean-shiftentropyuncertainty map |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Majid Shadman Roodposhti Arko Lucieer Asim Anees Brett A. Bryan |
spellingShingle |
Majid Shadman Roodposhti Arko Lucieer Asim Anees Brett A. Bryan A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification Remote Sensing image classification ensemble mean-shift entropy uncertainty map |
author_facet |
Majid Shadman Roodposhti Arko Lucieer Asim Anees Brett A. Bryan |
author_sort |
Majid Shadman Roodposhti |
title |
A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification |
title_short |
A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification |
title_full |
A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification |
title_fullStr |
A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification |
title_full_unstemmed |
A Robust Rule-Based Ensemble Framework Using Mean-Shift Segmentation for Hyperspectral Image Classification |
title_sort |
robust rule-based ensemble framework using mean-shift segmentation for hyperspectral image classification |
publisher |
MDPI AG |
series |
Remote Sensing |
issn |
2072-4292 |
publishDate |
2019-09-01 |
description |
This paper assesses the performance of DoTRules—a dictionary of trusted rules—as a supervised rule-based ensemble framework based on the mean-shift segmentation for hyperspectral image classification. The proposed ensemble framework consists of multiple rule sets with rules constructed based on different class frequencies and sequences of occurrences. Shannon entropy was derived for assessing the uncertainty of every rule and the subsequent filtering of unreliable rules. DoTRules is not only a transparent approach for image classification but also a tool to map rule uncertainty, where rule uncertainty assessment can be applied as an estimate of classification accuracy prior to image classification. In this research, the proposed image classification framework is implemented using three world reference hyperspectral image datasets. We found that the overall accuracy of classification using the proposed ensemble framework was superior to state-of-the-art ensemble algorithms, as well as two non-ensemble algorithms, at multiple training sample sizes. We believe DoTRules can be applied more generally to the classification of discrete data such as hyperspectral satellite imagery products. |
topic |
image classification ensemble mean-shift entropy uncertainty map |
url |
https://www.mdpi.com/2072-4292/11/17/2057 |
work_keys_str_mv |
AT majidshadmanroodposhti arobustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT arkolucieer arobustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT asimanees arobustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT brettabryan arobustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT majidshadmanroodposhti robustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT arkolucieer robustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT asimanees robustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification AT brettabryan robustrulebasedensembleframeworkusingmeanshiftsegmentationforhyperspectralimageclassification |
_version_ |
1716821232925540352 |