Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model
Improvements in hyperspectral image technology, diversification methods, and cost reductions have increased the convenience of hyperspectral data acquisitions. However, because of their multiband and multiredundant characteristics, hyperspectral data processing is still complex. Two feature extracti...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-01-01
|
| Series: | Discrete Dynamics in Nature and Society |
| Online Access: | http://dx.doi.org/10.1155/2023/9150482 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850221946206158848 |
|---|---|
| author | Xiaoai Dai Junying Cheng Shouheng Guo Chengchen Wang Ge Qu Wenxin Liu Weile Li Heng Lu Youlin Wang Binyang Zeng Yunjie Peng Shuneng Liang |
| author_facet | Xiaoai Dai Junying Cheng Shouheng Guo Chengchen Wang Ge Qu Wenxin Liu Weile Li Heng Lu Youlin Wang Binyang Zeng Yunjie Peng Shuneng Liang |
| author_sort | Xiaoai Dai |
| collection | DOAJ |
| description | Improvements in hyperspectral image technology, diversification methods, and cost reductions have increased the convenience of hyperspectral data acquisitions. However, because of their multiband and multiredundant characteristics, hyperspectral data processing is still complex. Two feature extraction algorithms, the autoencoder (AE) and restricted Boltzmann machine (RBM), were used to optimize the classification model parameters. The optimal classification model was obtained by comparing a stacked autoencoder (SAE) and a deep belief network (DBN). Finally, the SAE was further optimized by adding sparse representation constraints and GPU parallel computation to improve classification accuracy and speed. The research results show that the SAE enhanced by deep learning is superior to the traditional feature extraction algorithm. The optimal classification model based on deep learning, namely, the stacked sparse autoencoder, achieved 93.41% and 94.92% classification accuracy using two experimental datasets. The use of parallel computing increased the model’s training speed by more than seven times, solving the model’s lengthy training time limitation. |
| format | Article |
| id | doaj-art-71c0bb7c7e5e435bbe38d9f526122872 |
| institution | OA Journals |
| issn | 1607-887X |
| language | English |
| publishDate | 2023-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Discrete Dynamics in Nature and Society |
| spelling | doaj-art-71c0bb7c7e5e435bbe38d9f5261228722025-08-20T02:06:31ZengWileyDiscrete Dynamics in Nature and Society1607-887X2023-01-01202310.1155/2023/9150482Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification ModelXiaoai Dai0Junying Cheng1Shouheng Guo2Chengchen Wang3Ge Qu4Wenxin Liu5Weile Li6Heng Lu7Youlin Wang8Binyang Zeng9Yunjie Peng10Shuneng Liang11School of Earth Science Chengdu University of TechnologySchool of Earth Science Chengdu University of TechnologySchool of Earth Science Chengdu University of TechnologySchool of Earth Science Chengdu University of TechnologySchool of Earth Science Chengdu University of TechnologySchool of Earth Science Chengdu University of TechnologyState Key Laboratory of Geohazard Prevention and Geoenvironment ProtectionCollege of Hydraulic and Hydroelectric EngineeringNorthwest Engineering Corporation LimitedSouthwest Branch of China Petroleum Engineering Construction Co. LtdGEOVIS Wisdom Technology Co. LtdLand Satellite Remote Sensing Application CenterImprovements in hyperspectral image technology, diversification methods, and cost reductions have increased the convenience of hyperspectral data acquisitions. However, because of their multiband and multiredundant characteristics, hyperspectral data processing is still complex. Two feature extraction algorithms, the autoencoder (AE) and restricted Boltzmann machine (RBM), were used to optimize the classification model parameters. The optimal classification model was obtained by comparing a stacked autoencoder (SAE) and a deep belief network (DBN). Finally, the SAE was further optimized by adding sparse representation constraints and GPU parallel computation to improve classification accuracy and speed. The research results show that the SAE enhanced by deep learning is superior to the traditional feature extraction algorithm. The optimal classification model based on deep learning, namely, the stacked sparse autoencoder, achieved 93.41% and 94.92% classification accuracy using two experimental datasets. The use of parallel computing increased the model’s training speed by more than seven times, solving the model’s lengthy training time limitation.http://dx.doi.org/10.1155/2023/9150482 |
| spellingShingle | Xiaoai Dai Junying Cheng Shouheng Guo Chengchen Wang Ge Qu Wenxin Liu Weile Li Heng Lu Youlin Wang Binyang Zeng Yunjie Peng Shuneng Liang Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model Discrete Dynamics in Nature and Society |
| title | Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model |
| title_full | Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model |
| title_fullStr | Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model |
| title_full_unstemmed | Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model |
| title_short | Optimization Strategy of a Stacked Autoencoder and Deep Belief Network in a Hyperspectral Remote-Sensing Image Classification Model |
| title_sort | optimization strategy of a stacked autoencoder and deep belief network in a hyperspectral remote sensing image classification model |
| url | http://dx.doi.org/10.1155/2023/9150482 |
| work_keys_str_mv | AT xiaoaidai optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT junyingcheng optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT shouhengguo optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT chengchenwang optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT gequ optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT wenxinliu optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT weileli optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT henglu optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT youlinwang optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT binyangzeng optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT yunjiepeng optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel AT shunengliang optimizationstrategyofastackedautoencoderanddeepbeliefnetworkinahyperspectralremotesensingimageclassificationmodel |