Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures
Hyperspectral image classification has been increasingly used in the field of remote sensing. In this study, a new clustering framework for large-scale hyperspectral image (HSI) classification is proposed. The proposed four-step classification scheme explores how to effectively use the global spectr...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-12-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/13/12/330 |
| _version_ | 1797545112138416128 |
|---|---|
| author | Mohamed Ismail Milica Orlandić |
| author_facet | Mohamed Ismail Milica Orlandić |
| author_sort | Mohamed Ismail |
| collection | DOAJ |
| description | Hyperspectral image classification has been increasingly used in the field of remote sensing. In this study, a new clustering framework for large-scale hyperspectral image (HSI) classification is proposed. The proposed four-step classification scheme explores how to effectively use the global spectral information and local spatial structure of hyperspectral data for HSI classification. Initially, multidimensional Watershed is used for pre-segmentation. Region-based hierarchical hyperspectral image segmentation is based on the construction of Binary partition trees (BPT). Each segmented region is modeled while using first-order parametric modelling, which is then followed by a region merging stage using HSI regional spectral properties in order to obtain a BPT representation. The tree is then pruned to obtain a more compact representation. In addition, principal component analysis (PCA) is utilized for HSI feature extraction, so that the extracted features are further incorporated into the BPT. Finally, an efficient variant of k-means clustering algorithm, called filtering algorithm, is deployed on the created BPT structure, producing the final cluster map. The proposed method is tested over eight publicly available hyperspectral scenes with ground truth data and it is further compared with other clustering frameworks. The extensive experimental analysis demonstrates the efficacy of the proposed method. |
| first_indexed | 2024-03-10T14:10:49Z |
| format | Article |
| id | doaj.art-88dd896a029a4105af9ef393cf3697a6 |
| institution | Directory Open Access Journal |
| issn | 1999-4893 |
| language | English |
| last_indexed | 2024-03-10T14:10:49Z |
| publishDate | 2020-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj.art-88dd896a029a4105af9ef393cf3697a62023-11-21T00:12:21ZengMDPI AGAlgorithms1999-48932020-12-01131233010.3390/a13120330Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning StructuresMohamed Ismail0Milica Orlandić1Department of Electronic Systems, NTNU: Norwegian University of Science and Technology (NTNU), 7491 Trondheim, NorwayDepartment of Electronic Systems, NTNU: Norwegian University of Science and Technology (NTNU), 7491 Trondheim, NorwayHyperspectral image classification has been increasingly used in the field of remote sensing. In this study, a new clustering framework for large-scale hyperspectral image (HSI) classification is proposed. The proposed four-step classification scheme explores how to effectively use the global spectral information and local spatial structure of hyperspectral data for HSI classification. Initially, multidimensional Watershed is used for pre-segmentation. Region-based hierarchical hyperspectral image segmentation is based on the construction of Binary partition trees (BPT). Each segmented region is modeled while using first-order parametric modelling, which is then followed by a region merging stage using HSI regional spectral properties in order to obtain a BPT representation. The tree is then pruned to obtain a more compact representation. In addition, principal component analysis (PCA) is utilized for HSI feature extraction, so that the extracted features are further incorporated into the BPT. Finally, an efficient variant of k-means clustering algorithm, called filtering algorithm, is deployed on the created BPT structure, producing the final cluster map. The proposed method is tested over eight publicly available hyperspectral scenes with ground truth data and it is further compared with other clustering frameworks. The extensive experimental analysis demonstrates the efficacy of the proposed method.https://www.mdpi.com/1999-4893/13/12/330hyperspectral image (HSI)HSI clusteringHSI segmentationk-meanswatershedPrinciple Component Analysis (PCA) |
| spellingShingle | Mohamed Ismail Milica Orlandić Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures Algorithms hyperspectral image (HSI) HSI clustering HSI segmentation k-means watershed Principle Component Analysis (PCA) |
| title | Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures |
| title_full | Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures |
| title_fullStr | Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures |
| title_full_unstemmed | Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures |
| title_short | Segment-Based Clustering of Hyperspectral Images Using Tree-Based Data Partitioning Structures |
| title_sort | segment based clustering of hyperspectral images using tree based data partitioning structures |
| topic | hyperspectral image (HSI) HSI clustering HSI segmentation k-means watershed Principle Component Analysis (PCA) |
| url | https://www.mdpi.com/1999-4893/13/12/330 |
| work_keys_str_mv | AT mohamedismail segmentbasedclusteringofhyperspectralimagesusingtreebaseddatapartitioningstructures AT milicaorlandic segmentbasedclusteringofhyperspectralimagesusingtreebaseddatapartitioningstructures |