Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images
Autor(a) principal: | |
---|---|
Data de Publicação: | 2020 |
Outros Autores: | , , , , , , , , , , |
Tipo de documento: | Artigo de conferência |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UNESP |
Texto Completo: | http://dx.doi.org/10.5194/isprs-Annals-V-3-2020-193-2020 http://hdl.handle.net/11449/228844 |
Resumo: | The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user's knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and <i>Kappa</i> of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN. |
id |
UNSP_07fe3aa638161c0b1803f07251261c8c |
---|---|
oai_identifier_str |
oai:repositorio.unesp.br:11449/228844 |
network_acronym_str |
UNSP |
network_name_str |
Repositório Institucional da UNESP |
repository_id_str |
2946 |
spelling |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral imagesConvolutional neural networkData augmentationDeep learningFeature extractionRandom forestSupport vector machineTropical diversityUnmanned aerial vehiclesThe classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user's knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and <i>Kappa</i> of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN.School of Geography and Earth Sciences M CMaster UniversityDepartment of Electrical Engineering Pontifical Catholic University of Rio de JaneiroDivision of Remote Sensing National Institute for Space ResearchDepartment of Forest Engineering Santa Catarina State UniversityDepartment of Sustainable Agro Agro-Ecosystems and Bioresources Fondazione Edmund MachDepartment of Geography Santa Catarina State UniversityDepartment of Cartography São Paulo State UniversityDepartment of Cartography São Paulo State UniversityM CMaster UniversityPontifical Catholic University of Rio de JaneiroNational Institute for Space ResearchSanta Catarina State UniversityFondazione Edmund MachUniversidade Estadual Paulista (UNESP)Sothe, C.La Rosa, L. E.C.De Almeida, C. M.Gonsamo, A.Schimalski, M. B.Castro, J. D.B.Feitosa, R. Q.Dalponte, M.Lima, C. L.Liesenberg, V.Miyoshi, G. T. [UNESP]Tommaselli, A. M.G. [UNESP]2022-04-29T08:28:58Z2022-04-29T08:28:58Z2020-08-03info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/conferenceObject193-199http://dx.doi.org/10.5194/isprs-Annals-V-3-2020-193-2020ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, v. 5, n. 3, p. 193-199, 2020.2194-90502194-9042http://hdl.handle.net/11449/22884410.5194/isprs-Annals-V-3-2020-193-20202-s2.0-85090355806Scopusreponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciencesinfo:eu-repo/semantics/openAccess2024-06-18T15:02:08Zoai:repositorio.unesp.br:11449/228844Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462024-08-05T16:19:56.036044Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false |
dc.title.none.fl_str_mv |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
title |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
spellingShingle |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images Sothe, C. Convolutional neural network Data augmentation Deep learning Feature extraction Random forest Support vector machine Tropical diversity Unmanned aerial vehicles |
title_short |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
title_full |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
title_fullStr |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
title_full_unstemmed |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
title_sort |
Evaluating a convolutional neural network for feature extraction and tree species classification using uav-hyperspectral images |
author |
Sothe, C. |
author_facet |
Sothe, C. La Rosa, L. E.C. De Almeida, C. M. Gonsamo, A. Schimalski, M. B. Castro, J. D.B. Feitosa, R. Q. Dalponte, M. Lima, C. L. Liesenberg, V. Miyoshi, G. T. [UNESP] Tommaselli, A. M.G. [UNESP] |
author_role |
author |
author2 |
La Rosa, L. E.C. De Almeida, C. M. Gonsamo, A. Schimalski, M. B. Castro, J. D.B. Feitosa, R. Q. Dalponte, M. Lima, C. L. Liesenberg, V. Miyoshi, G. T. [UNESP] Tommaselli, A. M.G. [UNESP] |
author2_role |
author author author author author author author author author author author |
dc.contributor.none.fl_str_mv |
M CMaster University Pontifical Catholic University of Rio de Janeiro National Institute for Space Research Santa Catarina State University Fondazione Edmund Mach Universidade Estadual Paulista (UNESP) |
dc.contributor.author.fl_str_mv |
Sothe, C. La Rosa, L. E.C. De Almeida, C. M. Gonsamo, A. Schimalski, M. B. Castro, J. D.B. Feitosa, R. Q. Dalponte, M. Lima, C. L. Liesenberg, V. Miyoshi, G. T. [UNESP] Tommaselli, A. M.G. [UNESP] |
dc.subject.por.fl_str_mv |
Convolutional neural network Data augmentation Deep learning Feature extraction Random forest Support vector machine Tropical diversity Unmanned aerial vehicles |
topic |
Convolutional neural network Data augmentation Deep learning Feature extraction Random forest Support vector machine Tropical diversity Unmanned aerial vehicles |
description |
The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user's knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and <i>Kappa</i> of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN. |
publishDate |
2020 |
dc.date.none.fl_str_mv |
2020-08-03 2022-04-29T08:28:58Z 2022-04-29T08:28:58Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/conferenceObject |
format |
conferenceObject |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://dx.doi.org/10.5194/isprs-Annals-V-3-2020-193-2020 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, v. 5, n. 3, p. 193-199, 2020. 2194-9050 2194-9042 http://hdl.handle.net/11449/228844 10.5194/isprs-Annals-V-3-2020-193-2020 2-s2.0-85090355806 |
url |
http://dx.doi.org/10.5194/isprs-Annals-V-3-2020-193-2020 http://hdl.handle.net/11449/228844 |
identifier_str_mv |
ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, v. 5, n. 3, p. 193-199, 2020. 2194-9050 2194-9042 10.5194/isprs-Annals-V-3-2020-193-2020 2-s2.0-85090355806 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
193-199 |
dc.source.none.fl_str_mv |
Scopus reponame:Repositório Institucional da UNESP instname:Universidade Estadual Paulista (UNESP) instacron:UNESP |
instname_str |
Universidade Estadual Paulista (UNESP) |
instacron_str |
UNESP |
institution |
UNESP |
reponame_str |
Repositório Institucional da UNESP |
collection |
Repositório Institucional da UNESP |
repository.name.fl_str_mv |
Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP) |
repository.mail.fl_str_mv |
|
_version_ |
1808128634522697728 |