Decision tree-based classification in coastal area integrating polarimetric SAR and optical data
| DOI | https://doi.org/10.1108/DTA-08-2019-0149 |
| Published date | 14 October 2021 |
| Date | 14 October 2021 |
| Pages | 342-357 |
| Subject Matter | Library & information science,Librarianship/library management,Library technology,Information behaviour & retrieval,Metadata,Information & knowledge management,Information & communications technology,Internet |
| Author | Yuanyuan Chen,Xiufeng He,Jia Xu,Lin Guo,Yanyan Lu,Rongchun Zhang |
Decision tree-based classification
in coastal area integrating
polarimetric SAR and optical data
Yuanyuan Chen
College of Civil Engineering, Nanjing Forestry University, Nanjing, China
Xiufeng He and Jia Xu
School of Earth Sciences and Engineering, Hohai University, Nanjing, China
Lin Guo
Beijing Laboratory of Water Resources Security, Capital Normal University, Beijing,
China and
Key Laboratory of Mechanism, Prevention and Mitigation of Land Subsidence, MOE,
Capital Normal University, Beijing, China
Yanyan Lu
Research Academy of Natural Resources and Environment Audit,
Nanjing Audit University, Nanjing, China, and
Rongchun Zhang
School of Geographic and Biologic Information,
Nanjing University of Posts and Telecommunications, Nanjing, China
Abstract
Purpose –As one of the world’s most productive ecosystems, ecological land plays an important role in
regional and global environments. Utilizing advanced optical and synthetic aperture radar (SAR) data for land
cover/land use research becomes increasingly popular. This research aims to investigate the complementarity
of fully polarimetric SAR and optical imaging for ecological land classification in the eastern coastal area
of China.
Design/methodology/approach –Four polarimetric decomposition methods, namely, H/Alpha,
Yamaguchi3, VanZyl3 and Krogager, were applied to Advanced Land Observing Satellite (ALOS) SAR
image for scattering parameter extraction. These parameters were merged with ALOS optical parameters for
subsequent classification using the object-based quick, unbiased, efficient statistical tree decision tree method.
Findings –The experimental results indicate that an improved classification performance was obtained in the
decision level when merging the two data sources. In fact, unlike classification using only optical images, the
proposed approach allowed to distinguish ecological land with similar spectrum but different scattering.
Moreover, unlike classification using only polarimetric information, the integration of polarimetric and optical
data allows to accurately distinguish reed from artemisia and sand from salt field and therefore achieve a
detailed classification of the coastal area characteristics.
Originality/value –This research proposed an integrated classification method for coastal ecological land
with polarimetric SAR and optical data. The object-based and decision-level fusion enables effective ecological
land classification in coastal area was verified.
Keywords Decision tree, Polarimetric SAR, Optical data, Coastal area, Ecological land, Classification
Paper type Research paper
DTA
56,3
342
This research was supported by the Natural Science Foundation of Jiangsu Province (Grant No.
BK20180779), Natural Science Foundation of China (41830110, 41901401), and the Youth Science and
Technology Innovation Fund Project of Nanjing Forestry University (Grant No. CX2018015) from China.
The ALOS PALSAR imagery was provided by the Japan Aerospace Exploration Agency.
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2514-9288.htm
Received 7 May 2020
Revised 6 October 2020
8 February 2021
Accepted 3 April 2021
Data Technologies and
Applications
Vol. 56 No. 3, 2022
pp. 342-357
© Emerald Publishing Limited
2514-9288
DOI 10.1108/DTA-08-2019-0149
1. Introduction
As an important land type in the biosphere, ecological land plays an important role in water
conservation, soil protection, windbreaking for sand fixation, climate regulation,
environmental purification, biodiversity conservation and other aspects (Cramb, 1998;Liu
et al., 2017a). In coastal areas, assessing the ecological land is important for measuring the
environmental quality and essential for developing coastal ecological security patterns.
Specifically, to protect the ecological land in the eastern coastal area of Jiangsu, China, it is
necessary to regularly monitor land resources, especially the ecological land, and accurately
determine the distribution and availability of land resources.
As a fast and large-scale measurement technology, remote sensing has been widely used
for land cover monitoring and investigation (Fuller et al., 2002). Among the available
techniques, optical remote sensing (e.g. high-resolution satellite imagery) allows obtaining
high geometrical resolution, multispectral information and detailed characteristics, thus
having been widely used for land cover classification (Haack and English, 1996;Li et al., 2014).
However, optical imaging is severely undermined by conditions such as weather and
illumination, especially in coastal areas with high vegetation coverage and water vapor
content. On the other hand, microwave remote sensing (e.g. synthetic aperture radar [SAR])
can overcome the limitations of optical remote sensing (Zhang et al., 2014;Samadi et al., 2019;
Sharifzadeh et al., 2019;Akbarizadeh, 2012). Moreover, the development of polarimetric SAR
systems enables to exploit SAR polarization for monitoring and identifying land cover types
(van Beijma et al., 2014;Chen et al., 2015;Niu and Ban, 2013;Fang et al., 2018), especially for
coastal region land use classifications (de Almeida Furtado et al., 2016;Pereira et al., 2018;
Chen et al., 2020). As optical and microwave remote sensing exhibit specific advantages and
drawbacks, their combination can improve the accuracy of ecological land classification. In
fact, various studies have demonstrated the complementarity of these two technologies in
land cover classification and mapping (Xiao et al., 2016;Luo et al., 2014), but scarce research
has been devoted to the application of fully polarimetric Advanced Land Observing Satellite
(ALOS) SAR (phased array type L-band SAR, PALSAR) and ALOS optical images (PRIMS
and AVNIR-2) for coastal ecological land classification.
The selection of classification method determines the accuracy of land cover mapping and
identification. Most existing studies, which apply pixel-based fusion and classification
methods, only consider the spectral information of images and use the individual pixel value
to determine the class of that pixel (Harris et al., 2018;Wei et al., 2017). To improve
classification accuracy, geometric, statistical and textural information was considered in
object-based analysis of remote sensing data (Xiao et al., 2016;Chen et al., 2014;Qin et al.,
2017). By focusing on spatially related objects, object-based methods can retrieve more
information and mitigate the adverse effects of noise compared to pixel-based methods.
Alternatively, classification accuracy and data fusion can be improved using decision trees,
which allow to integrate different information, such as polarimetric scattering information
and optical information, at the decision level (Chen et al., 2014;Zhang et al., 2015;Yang
et al., 2018).
In this study, fully polarimetric ALOS SAR image was integrated with ALOS optical data
at the decision level. Then, the object-based quick, unbiased, efficient statistical tree (QUEST)
decision tree method was applied for ecological land classification in the eastern coastal area
of China. Experiments aimed to verify the effectiveness of combining polarimetric scattering
information and spectral information and that of the object-based QUEST method.
2. Study area and satellite image
The study area considers the eastern coastal region of Jiangsu, China. As an important part of
the Yangtze River Delta, it includes a part of land, intertidal zone and shallow sea. This region
Decision tree-
based
classification
343
To continue reading
Request your trial