Semi-supervised graph fusion of hyperspectral and LiDAR data for classification

Wenzhi Liao; Junshi Xia; Peijun Du; Wilfried Philips; Vito Pascazio; Sebastiano B Serpico

doi:10.1109/IGARSS.2015.7325695

Semi-supervised graph fusion of hyperspectral and LiDAR data for classification

Wenzhi Liao, Junshi Xia, Peijun Du, Wilfried Philips, Vito Pascazio (Editor), Sebastiano B Serpico (Editor)

Electronic And Electrical Engineering

Research output: Contribution to conference › Paper › peer-review

11 Citations (Scopus)

Abstract

This paper proposes a semi-supervised graph-based fusion framework to couple dimensionality reduction and the fusion of multi-sensor data for classification. First, morphological features are used to model the elevation and spatial information contained in both LiDAR data and on the first few principal components (PCs) of the original hyperspectral (HS) image. Then, we fuse the features by projecting the spectral, spatial and elevation features onto a lower subspace through our proposed semi-supervised fusion graph. Experimental results on fusion of HS and LiDAR data from the 2013 IEEE GRSS Data Fusion Contest demonstrate effectiveness of the proposed method. Compared to the methods using single data source or unsupervised graph fusion, with the proposed method, overall classification accuracies were improved by 9% and 4%, respectively.

Original language	English
Pages	53-56
Number of pages	4
DOIs	https://doi.org/10.1109/IGARSS.2015.7325695
Publication status	Published - 12 Nov 2015
Event	IEEE International Geoscience and Remote Sensing Sympium (IGARSS 2015) - Convention Center, Milan, Italy Duration: 26 Jul 2015 → 31 Jul 2015

Conference

Conference	IEEE International Geoscience and Remote Sensing Sympium (IGARSS 2015)
Abbreviated title	IGARSS 2015
Country/Territory	Italy
City	Milan
Period	26/07/15 → 31/07/15

Keywords

graph-based
data fusion
remote sensing
hyperspectral image
LiDAR data
areas
images
feature extraction
directional morphological profiles

Access to Document

10.1109/IGARSS.2015.7325695

Cite this

@conference{c926604e669b4179adeb754f7b9371e9,

title = "Semi-supervised graph fusion of hyperspectral and LiDAR data for classification",

abstract = "This paper proposes a semi-supervised graph-based fusion framework to couple dimensionality reduction and the fusion of multi-sensor data for classification. First, morphological features are used to model the elevation and spatial information contained in both LiDAR data and on the first few principal components (PCs) of the original hyperspectral (HS) image. Then, we fuse the features by projecting the spectral, spatial and elevation features onto a lower subspace through our proposed semi-supervised fusion graph. Experimental results on fusion of HS and LiDAR data from the 2013 IEEE GRSS Data Fusion Contest demonstrate effectiveness of the proposed method. Compared to the methods using single data source or unsupervised graph fusion, with the proposed method, overall classification accuracies were improved by 9% and 4%, respectively.",

keywords = "graph-based, data fusion, remote sensing, hyperspectral image, LiDAR data, areas, images, feature extraction, directional morphological profiles",

author = "Wenzhi Liao and Junshi Xia and Peijun Du and Wilfried Philips and Vito Pascazio and {B Serpico}, Sebastiano",

year = "2015",

month = nov,

day = "12",

doi = "10.1109/IGARSS.2015.7325695",

language = "English",

pages = "53--56",

note = "IEEE International Geoscience and Remote Sensing Sympium (IGARSS 2015), IGARSS 2015 ; Conference date: 26-07-2015 Through 31-07-2015",

}

TY - CONF

T1 - Semi-supervised graph fusion of hyperspectral and LiDAR data for classification

AU - Liao, Wenzhi

AU - Xia, Junshi

AU - Du, Peijun

AU - Philips, Wilfried

A2 - Pascazio, Vito

A2 - B Serpico, Sebastiano

PY - 2015/11/12

Y1 - 2015/11/12

N2 - This paper proposes a semi-supervised graph-based fusion framework to couple dimensionality reduction and the fusion of multi-sensor data for classification. First, morphological features are used to model the elevation and spatial information contained in both LiDAR data and on the first few principal components (PCs) of the original hyperspectral (HS) image. Then, we fuse the features by projecting the spectral, spatial and elevation features onto a lower subspace through our proposed semi-supervised fusion graph. Experimental results on fusion of HS and LiDAR data from the 2013 IEEE GRSS Data Fusion Contest demonstrate effectiveness of the proposed method. Compared to the methods using single data source or unsupervised graph fusion, with the proposed method, overall classification accuracies were improved by 9% and 4%, respectively.

AB - This paper proposes a semi-supervised graph-based fusion framework to couple dimensionality reduction and the fusion of multi-sensor data for classification. First, morphological features are used to model the elevation and spatial information contained in both LiDAR data and on the first few principal components (PCs) of the original hyperspectral (HS) image. Then, we fuse the features by projecting the spectral, spatial and elevation features onto a lower subspace through our proposed semi-supervised fusion graph. Experimental results on fusion of HS and LiDAR data from the 2013 IEEE GRSS Data Fusion Contest demonstrate effectiveness of the proposed method. Compared to the methods using single data source or unsupervised graph fusion, with the proposed method, overall classification accuracies were improved by 9% and 4%, respectively.

KW - graph-based

KW - data fusion

KW - remote sensing

KW - hyperspectral image

KW - LiDAR data

KW - areas

KW - images

KW - feature extraction

KW - directional morphological profiles

UR - http://hdl.handle.net/1854/LU-6908425

U2 - 10.1109/IGARSS.2015.7325695

DO - 10.1109/IGARSS.2015.7325695

M3 - Paper

SP - 53

EP - 56

T2 - IEEE International Geoscience and Remote Sensing Sympium (IGARSS 2015)

Y2 - 26 July 2015 through 31 July 2015

ER -

Semi-supervised graph fusion of hyperspectral and LiDAR data for classification

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this