Cargando…

Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network

In a traditional convolutional neural network structure, pooling layers generally use an average pooling method: a non-overlapping pooling. However, this condition results in similarities in the extracted image features, especially for the hyperspectral images of a continuous spectrum, which makes i...

Descripción completa

Detalles Bibliográficos
Autores principales: Li, Chenming, Yang, Simon X., Yang, Yao, Gao, Hongmin, Zhao, Jia, Qu, Xiaoyu, Wang, Yongchang, Yao, Dan, Gao, Jianbing
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2018
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6210679/
https://www.ncbi.nlm.nih.gov/pubmed/30360445
http://dx.doi.org/10.3390/s18103587
_version_ 1783367171250847744
author Li, Chenming
Yang, Simon X.
Yang, Yao
Gao, Hongmin
Zhao, Jia
Qu, Xiaoyu
Wang, Yongchang
Yao, Dan
Gao, Jianbing
author_facet Li, Chenming
Yang, Simon X.
Yang, Yao
Gao, Hongmin
Zhao, Jia
Qu, Xiaoyu
Wang, Yongchang
Yao, Dan
Gao, Jianbing
author_sort Li, Chenming
collection PubMed
description In a traditional convolutional neural network structure, pooling layers generally use an average pooling method: a non-overlapping pooling. However, this condition results in similarities in the extracted image features, especially for the hyperspectral images of a continuous spectrum, which makes it more difficult to extract image features with differences, and image detail features are easily lost. This result seriously affects the accuracy of image classification. Thus, a new overlapping pooling method is proposed, where maximum pooling is used in an improved convolutional neural network to avoid the fuzziness of average pooling. The step size used is smaller than the size of the pooling kernel to achieve overlapping and coverage between the outputs of the pooling layer. The dataset selected for this experiment was the Indian Pines dataset, collected by the airborne visible/infrared imaging spectrometer (AVIRIS) sensor. Experimental results show that using the improved convolutional neural network for remote sensing image classification can effectively improve the details of the image and obtain a high classification accuracy.
format Online
Article
Text
id pubmed-6210679
institution National Center for Biotechnology Information
language English
publishDate 2018
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-62106792018-11-02 Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network Li, Chenming Yang, Simon X. Yang, Yao Gao, Hongmin Zhao, Jia Qu, Xiaoyu Wang, Yongchang Yao, Dan Gao, Jianbing Sensors (Basel) Article In a traditional convolutional neural network structure, pooling layers generally use an average pooling method: a non-overlapping pooling. However, this condition results in similarities in the extracted image features, especially for the hyperspectral images of a continuous spectrum, which makes it more difficult to extract image features with differences, and image detail features are easily lost. This result seriously affects the accuracy of image classification. Thus, a new overlapping pooling method is proposed, where maximum pooling is used in an improved convolutional neural network to avoid the fuzziness of average pooling. The step size used is smaller than the size of the pooling kernel to achieve overlapping and coverage between the outputs of the pooling layer. The dataset selected for this experiment was the Indian Pines dataset, collected by the airborne visible/infrared imaging spectrometer (AVIRIS) sensor. Experimental results show that using the improved convolutional neural network for remote sensing image classification can effectively improve the details of the image and obtain a high classification accuracy. MDPI 2018-10-22 /pmc/articles/PMC6210679/ /pubmed/30360445 http://dx.doi.org/10.3390/s18103587 Text en © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Li, Chenming
Yang, Simon X.
Yang, Yao
Gao, Hongmin
Zhao, Jia
Qu, Xiaoyu
Wang, Yongchang
Yao, Dan
Gao, Jianbing
Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title_full Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title_fullStr Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title_full_unstemmed Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title_short Hyperspectral Remote Sensing Image Classification Based on Maximum Overlap Pooling Convolutional Neural Network
title_sort hyperspectral remote sensing image classification based on maximum overlap pooling convolutional neural network
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6210679/
https://www.ncbi.nlm.nih.gov/pubmed/30360445
http://dx.doi.org/10.3390/s18103587
work_keys_str_mv AT lichenming hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT yangsimonx hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT yangyao hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT gaohongmin hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT zhaojia hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT quxiaoyu hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT wangyongchang hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT yaodan hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork
AT gaojianbing hyperspectralremotesensingimageclassificationbasedonmaximumoverlappoolingconvolutionalneuralnetwork