An entropy and MRF model-based CNN for large-scale Landsat image classification

Xuemei Zhao, Lianru Gao, Zhengchao Chen, Bing Zhang, Wenzhi Liao

Research output: Contribution to journalArticle

Abstract

Large-scale Landsat image classification is essential for the production of land cover maps. The rise of convolutional neural networks (CNNs) provides a new idea for the implementation of Landsat image classification. However, pixels in Landsat images have higher uncertainty compared with high-resolution images due to its 30-m spatial resolution. In addition, the current deep learning methods tend to lose detailed information such as boundaries along with the stacking of convolutional and pooling layers. To solve these problems, we propose a new method called entropy and MRF model (EMM)-CNN based on Pyramid Scene Parsing Network. The EMM-CNN uses entropy to decrease the uncertainty of pixels. Then, the Markov random filed (MRF) model is employed to construct the connections between neighboring pixels and defined a prior distribution to prevent the cross entropy from sacrificing detailed information for the overall accuracy. Finally, transfer learning based on the pretrained ImageNet is introduced to overcome the shortage of training samples and boost the speed of the training process. Experimental results demonstrate that the proposed EMM-CNN is able to obtain classification results with fine structure by decreasing the uncertainty and retaining detailed information of the detected image.
LanguageEnglish
Pages1145-1149
Number of pages5
JournalIEEE Geoscience and Remote Sensing Letters
Volume16
Issue number7
Early online date24 Jan 2019
DOIs
Publication statusPublished - 31 Jul 2019

Fingerprint

Image classification
image classification
Landsat
entropy
Entropy
Neural networks
pixel
Pixels
learning
Optical resolving power
image resolution
Image resolution
stacking
land cover
spatial resolution
Uncertainty
method

Keywords

  • convolutional neural network (CNN)
  • entropy
  • Landsat image classification
  • Markov random field (MRF) model
  • transfer learning

Cite this

Zhao, Xuemei ; Gao, Lianru ; Chen, Zhengchao ; Zhang, Bing ; Liao, Wenzhi. / An entropy and MRF model-based CNN for large-scale Landsat image classification. In: IEEE Geoscience and Remote Sensing Letters. 2019 ; Vol. 16, No. 7. pp. 1145-1149.
@article{8ab4a4e7b0c7427e99748edf5eebda96,
title = "An entropy and MRF model-based CNN for large-scale Landsat image classification",
abstract = "Large-scale Landsat image classification is essential for the production of land cover maps. The rise of convolutional neural networks (CNNs) provides a new idea for the implementation of Landsat image classification. However, pixels in Landsat images have higher uncertainty compared with high-resolution images due to its 30-m spatial resolution. In addition, the current deep learning methods tend to lose detailed information such as boundaries along with the stacking of convolutional and pooling layers. To solve these problems, we propose a new method called entropy and MRF model (EMM)-CNN based on Pyramid Scene Parsing Network. The EMM-CNN uses entropy to decrease the uncertainty of pixels. Then, the Markov random filed (MRF) model is employed to construct the connections between neighboring pixels and defined a prior distribution to prevent the cross entropy from sacrificing detailed information for the overall accuracy. Finally, transfer learning based on the pretrained ImageNet is introduced to overcome the shortage of training samples and boost the speed of the training process. Experimental results demonstrate that the proposed EMM-CNN is able to obtain classification results with fine structure by decreasing the uncertainty and retaining detailed information of the detected image.",
keywords = "convolutional neural network (CNN), entropy, Landsat image classification, Markov random field (MRF) model, transfer learning",
author = "Xuemei Zhao and Lianru Gao and Zhengchao Chen and Bing Zhang and Wenzhi Liao",
note = "{\circledC} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",
year = "2019",
month = "7",
day = "31",
doi = "10.1109/LGRS.2019.2890996",
language = "English",
volume = "16",
pages = "1145--1149",
journal = "IEEE Geoscience and Remote Sensing Letters",
issn = "1545-598X",
publisher = "IEEE",
number = "7",

}

An entropy and MRF model-based CNN for large-scale Landsat image classification. / Zhao, Xuemei; Gao, Lianru; Chen, Zhengchao; Zhang, Bing; Liao, Wenzhi.

In: IEEE Geoscience and Remote Sensing Letters, Vol. 16, No. 7, 31.07.2019, p. 1145-1149.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An entropy and MRF model-based CNN for large-scale Landsat image classification

AU - Zhao, Xuemei

AU - Gao, Lianru

AU - Chen, Zhengchao

AU - Zhang, Bing

AU - Liao, Wenzhi

N1 - © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2019/7/31

Y1 - 2019/7/31

N2 - Large-scale Landsat image classification is essential for the production of land cover maps. The rise of convolutional neural networks (CNNs) provides a new idea for the implementation of Landsat image classification. However, pixels in Landsat images have higher uncertainty compared with high-resolution images due to its 30-m spatial resolution. In addition, the current deep learning methods tend to lose detailed information such as boundaries along with the stacking of convolutional and pooling layers. To solve these problems, we propose a new method called entropy and MRF model (EMM)-CNN based on Pyramid Scene Parsing Network. The EMM-CNN uses entropy to decrease the uncertainty of pixels. Then, the Markov random filed (MRF) model is employed to construct the connections between neighboring pixels and defined a prior distribution to prevent the cross entropy from sacrificing detailed information for the overall accuracy. Finally, transfer learning based on the pretrained ImageNet is introduced to overcome the shortage of training samples and boost the speed of the training process. Experimental results demonstrate that the proposed EMM-CNN is able to obtain classification results with fine structure by decreasing the uncertainty and retaining detailed information of the detected image.

AB - Large-scale Landsat image classification is essential for the production of land cover maps. The rise of convolutional neural networks (CNNs) provides a new idea for the implementation of Landsat image classification. However, pixels in Landsat images have higher uncertainty compared with high-resolution images due to its 30-m spatial resolution. In addition, the current deep learning methods tend to lose detailed information such as boundaries along with the stacking of convolutional and pooling layers. To solve these problems, we propose a new method called entropy and MRF model (EMM)-CNN based on Pyramid Scene Parsing Network. The EMM-CNN uses entropy to decrease the uncertainty of pixels. Then, the Markov random filed (MRF) model is employed to construct the connections between neighboring pixels and defined a prior distribution to prevent the cross entropy from sacrificing detailed information for the overall accuracy. Finally, transfer learning based on the pretrained ImageNet is introduced to overcome the shortage of training samples and boost the speed of the training process. Experimental results demonstrate that the proposed EMM-CNN is able to obtain classification results with fine structure by decreasing the uncertainty and retaining detailed information of the detected image.

KW - convolutional neural network (CNN)

KW - entropy

KW - Landsat image classification

KW - Markov random field (MRF) model

KW - transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85068238461&partnerID=8YFLogxK

UR - http://hdl.handle.net/1854/LU-8604474

U2 - 10.1109/LGRS.2019.2890996

DO - 10.1109/LGRS.2019.2890996

M3 - Article

VL - 16

SP - 1145

EP - 1149

JO - IEEE Geoscience and Remote Sensing Letters

T2 - IEEE Geoscience and Remote Sensing Letters

JF - IEEE Geoscience and Remote Sensing Letters

SN - 1545-598X

IS - 7

ER -