Adaptive Spatial-Spectral Feature Learning for Hyperspectral Image Classification

Simin Li; Xueyu Zhu; Yang Liu; Jie Bao

doi:10.1109/ACCESS.2019.2916095

Back

Adaptive Spatial-Spectral Feature Learning for Hyperspectral Image Classification

Journal article

Open access

Peer reviewed

Adaptive Spatial-Spectral Feature Learning for Hyperspectral Image Classification

Simin Li, Xueyu Zhu, Yang Liu and Jie Bao

IEEE access, Vol.7, pp.61534-61547

2019

DOI: 10.1109/ACCESS.2019.2916095

Files and links (1)

url

https://doi.org/10.1109/ACCESS.2019.2916095View

Published (Version of record) Open Access

Abstract

The combination of spectral and spatial information provides an effective way to improve the hyperspectral image (HSI) classification. However, the local spatial contexture changes with different neighborhood regions over the HSI image plane, and methods with fixed weights to integrate spatial information for all neighborhood regions could result in inaccurate spatial features, leading to adverse effects on classification performance. To address this issue, a novel adaptive spatial-spectral feature learning network (ASSFL) has been proposed to reflect spatial contexture changes and learn robust adaptive features in this paper. In the implementation of the proposed method, a convolution neural network (CNN) is first applied to learn weight features for each pixel within a hyperspectral patch and adaptive weights can be obtained based on a softmax normalization. Then, the shallow joint adaptive features can be acquired according to these weights. After that, a stacked auto-encoder (SAE) is proposed to further extract deeper hierarchical features for the final classification. The experimental results on four benchmark HSI data sets demonstrate that the proposed method can achieve competitive classification results compared with other existing classifiers.

Adaptive spatial-spectral features

Adaptive systems

Convolution

convolutional neural network (CNN)

Deep learning

Feature extraction

hyperspectral image (HSI) classification

Hyperspectral imaging

stacked auto-encoder (SAE)

Support vector machines

Details

Title: Subtitle: Adaptive Spatial-Spectral Feature Learning for Hyperspectral Image Classification
Creators: Simin Li - Tsinghua University
Xueyu Zhu - University of Iowa
Yang Liu - QuantaEye (Beijing) Technologies Co., Ltd., Beijing, China
Jie Bao - Tsinghua University
Resource Type: Journal article
Publication Details: IEEE access, Vol.7, pp.61534-61547
Publisher: IEEE
DOI: 10.1109/ACCESS.2019.2916095
ISSN: 2169-3536
eISSN: 2169-3536
Grant note: Tsinghua University (10.13039/501100004147) Beijing National Research Center for Information Science and Technology
Language: English
Date published: 2019
Academic Unit: Mathematics
Record Identifier: 9984240861502771

Metrics

55 Record Views

19 Times Cited - Web of Science