Super-resolution land cover mapping based on the convolutional neural network

Yuanxin Jia, Yong Ge*, Yuehong Chen, Sanping Li, Gerard B.M. Heuvelink, Feng Ling

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Super-resolution mapping (SRM) is used to obtain fine-scale land cover maps from coarse remote sensing images. Spatial attraction, geostatistics, and using prior geographic information are conventional approaches used to derive fine-scale land cover maps. As the convolutional neural network (CNN) has been shown to be effective in capturing the spatial characteristics of geographic objects and extrapolating calibrated methods to other study areas, it may be a useful approach to overcome limitations of current SRM methods. In this paper, a new SRM method based on the CNN (SRMCNN) is proposed and tested. Specifically, an encoder-decoder CNN is used to model the nonlinear relationship between coarse remote sensing images and fine-scale land cover maps. Two real-image experiments were conducted to analyze the effectiveness of the proposed method. The results demonstrate that the overall accuracy of the proposed SRMCNN method was 3% to 5% higher than that of two existing SRM methods. Moreover, the proposed SRMCNN method was validated by visualizing output features and analyzing the performance of different geographic objects.

Original languageEnglish
Article number1815
JournalRemote Sensing
Volume11
Issue number15
DOIs
Publication statusPublished - 1 Aug 2019

Fingerprint

land cover
mapping method
remote sensing
geostatistics
method
experiment

Keywords

  • Convolutional neural network
  • Land cover
  • Remote sensing imagery
  • Super-resolution mapping

Cite this

Jia, Yuanxin ; Ge, Yong ; Chen, Yuehong ; Li, Sanping ; Heuvelink, Gerard B.M. ; Ling, Feng. / Super-resolution land cover mapping based on the convolutional neural network. In: Remote Sensing. 2019 ; Vol. 11, No. 15.
@article{0721474fa8764cb49cdb8cff146ffeb2,
title = "Super-resolution land cover mapping based on the convolutional neural network",
abstract = "Super-resolution mapping (SRM) is used to obtain fine-scale land cover maps from coarse remote sensing images. Spatial attraction, geostatistics, and using prior geographic information are conventional approaches used to derive fine-scale land cover maps. As the convolutional neural network (CNN) has been shown to be effective in capturing the spatial characteristics of geographic objects and extrapolating calibrated methods to other study areas, it may be a useful approach to overcome limitations of current SRM methods. In this paper, a new SRM method based on the CNN (SRMCNN) is proposed and tested. Specifically, an encoder-decoder CNN is used to model the nonlinear relationship between coarse remote sensing images and fine-scale land cover maps. Two real-image experiments were conducted to analyze the effectiveness of the proposed method. The results demonstrate that the overall accuracy of the proposed SRMCNN method was 3{\%} to 5{\%} higher than that of two existing SRM methods. Moreover, the proposed SRMCNN method was validated by visualizing output features and analyzing the performance of different geographic objects.",
keywords = "Convolutional neural network, Land cover, Remote sensing imagery, Super-resolution mapping",
author = "Yuanxin Jia and Yong Ge and Yuehong Chen and Sanping Li and Heuvelink, {Gerard B.M.} and Feng Ling",
year = "2019",
month = "8",
day = "1",
doi = "10.3390/rs11151815",
language = "English",
volume = "11",
journal = "Remote Sensing",
issn = "2072-4292",
publisher = "MDPI",
number = "15",

}

Super-resolution land cover mapping based on the convolutional neural network. / Jia, Yuanxin; Ge, Yong; Chen, Yuehong; Li, Sanping; Heuvelink, Gerard B.M.; Ling, Feng.

In: Remote Sensing, Vol. 11, No. 15, 1815, 01.08.2019.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Super-resolution land cover mapping based on the convolutional neural network

AU - Jia, Yuanxin

AU - Ge, Yong

AU - Chen, Yuehong

AU - Li, Sanping

AU - Heuvelink, Gerard B.M.

AU - Ling, Feng

PY - 2019/8/1

Y1 - 2019/8/1

N2 - Super-resolution mapping (SRM) is used to obtain fine-scale land cover maps from coarse remote sensing images. Spatial attraction, geostatistics, and using prior geographic information are conventional approaches used to derive fine-scale land cover maps. As the convolutional neural network (CNN) has been shown to be effective in capturing the spatial characteristics of geographic objects and extrapolating calibrated methods to other study areas, it may be a useful approach to overcome limitations of current SRM methods. In this paper, a new SRM method based on the CNN (SRMCNN) is proposed and tested. Specifically, an encoder-decoder CNN is used to model the nonlinear relationship between coarse remote sensing images and fine-scale land cover maps. Two real-image experiments were conducted to analyze the effectiveness of the proposed method. The results demonstrate that the overall accuracy of the proposed SRMCNN method was 3% to 5% higher than that of two existing SRM methods. Moreover, the proposed SRMCNN method was validated by visualizing output features and analyzing the performance of different geographic objects.

AB - Super-resolution mapping (SRM) is used to obtain fine-scale land cover maps from coarse remote sensing images. Spatial attraction, geostatistics, and using prior geographic information are conventional approaches used to derive fine-scale land cover maps. As the convolutional neural network (CNN) has been shown to be effective in capturing the spatial characteristics of geographic objects and extrapolating calibrated methods to other study areas, it may be a useful approach to overcome limitations of current SRM methods. In this paper, a new SRM method based on the CNN (SRMCNN) is proposed and tested. Specifically, an encoder-decoder CNN is used to model the nonlinear relationship between coarse remote sensing images and fine-scale land cover maps. Two real-image experiments were conducted to analyze the effectiveness of the proposed method. The results demonstrate that the overall accuracy of the proposed SRMCNN method was 3% to 5% higher than that of two existing SRM methods. Moreover, the proposed SRMCNN method was validated by visualizing output features and analyzing the performance of different geographic objects.

KW - Convolutional neural network

KW - Land cover

KW - Remote sensing imagery

KW - Super-resolution mapping

U2 - 10.3390/rs11151815

DO - 10.3390/rs11151815

M3 - Article

VL - 11

JO - Remote Sensing

JF - Remote Sensing

SN - 2072-4292

IS - 15

M1 - 1815

ER -