RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA
Autor(a) principal: | |
---|---|
Data de Publicação: | 2021 |
Tipo de documento: | Tese |
Idioma: | por |
Título da fonte: | Biblioteca Digital de Teses e Dissertações do UNICENTRO |
Texto Completo: | http://tede.unicentro.br:8080/jspui/handle/jspui/1725 |
Resumo: | Habitat loss and bioinvasion are the main causes of species extinctions associated with anthropic processes. One of the main tools for the Spatio temporal analysis of these processes is remote sensing. With the focus on recognizing patterns in the Mixed Ombrophylous Rainforest (MOF) at different spatial and temporal scales, three chapters were sought: (1) Classify and measure elements of the landscape of a hydrographic sub-basin in the Mixed Rainforest in Center-South Paraná state, associating the landscape components to the phytosociological parameters of the tree community; (2) Quantify the recognition rate of Hovenia dulcis Thunb. (Japanese raisin tree), invasive alien species and Araucaria angustifolia (Bertol.) Kuntze (Araucária), native conifer threatened with extinction, by machine learning algorithms; and (3) Improve the extraction of attributes for the classification of H. dulcis, in a pixel-by-pixel classification applied in a One-class-Classification (OCC) model. To achieve the objectives, the methods employed were: (1) For two WolrdVew2 images from the years 2012 and 2016, using only the multispectral bands of the sensor (2m spatial resolution), multiresolution segmentation, and Support Vector Machine (SVM). Nine landscape metrics were extracted and submitted to a Principal Component Analysis (PCA) from the most accurate classifications. These components were associated with regression models generated by Generalized Estimation Equation (GEE) for nine phytosociological parameters calculated from nine permanent forest inventory plots, with three measurements 2011, 2014, and 2017 distributed in the area of the basin in question; (2) For the recognition of A. angustifolia and H. dulcis, the visual interpretation and clipping of the treetops validated by RPA images (Remoted piloted aircraft) was performed in the 2016 WorldView-2 image fused (0.5 m spatial resolution). The cutouts had textural attributes extracted by Edge Filter and Pyramid Histogram of Oriented Gradients (PHOG), the textural attributes, and the images' pixels' spectral values were submitted to data mining using a genetic algorithm. Finally, they were submitted to two classifiers, Artificial Neural Networks (ANN) and Random Forest (RF); and (3) In two 2018 RPA images, random pixels from H. dulcis and the negative class (non-H. dulcis) were selected in a balanced way. Then, OCC models classified pixel-by-pixel were created by the Random Forest algorithm, using only the pixels' spectral attributes for the classification. In general, approximately 90% accuracy was found for all classifications used, with the different classification techniques applied to different images and targets. Concerning Chapter 1, there was a reduction in the area of MOF coverage of 6.4% and an agricultural expansion of 8.07% in the four years of analysis, mainly due to the increase in the landscape fragmentation process. It was still possible to find associations of fragmentation in the tree community, which showed changes mainly related to the increase in biomass, indicative of regeneration processes. Chapter 2, on the other hand, species classification was performed with high accuracy, reaching 95% of correct answers for Cross-Validation. RF surpassed the ANN classification rates and still proved to be more stable and faster for training and later species classification. For chapter 3, the classification of RPA images was also highly accurate, with around 95% correct answers. With only three spectral bands: Red (R), Green (G) and Blue (B), the spectral attributes of the red and green range were more relevant for the classification, and the pixel-by-pixel classification model of only one class proved efficient in the detection of H. dulcis. It was concluded that the WorldView-2 multispectral sensor and images of RPAs, associated or not, have a high potential to provide information for recognizing MOF forest patterns, ranging from landscape elements, e.g. forest remnants, to tree species canopies, such as H. dulcis and A. angustifolia. Thus, it can be effective in contributing to the conservation and sustainable management of the MOF. |
id |
UCEN_a6a7552f72cff245da018efa3b9376e7 |
---|---|
oai_identifier_str |
oai:localhost:jspui/1725 |
network_acronym_str |
UCEN |
network_name_str |
Biblioteca Digital de Teses e Dissertações do UNICENTRO |
repository_id_str |
|
spelling |
Figueiredo Filho, Afonsohttp://lattes.cnpq.br/4151544991447365Pesck, Vagner Alexhttp://lattes.cnpq.br/4670102252045134Lima, Vanderlei Aparecido dehttp://lattes.cnpq.br/9090461949264421069.450.439-40http://lattes.cnpq.br/3734724830416138Crisigiovanni, Enzo Luigi2021-11-05T12:46:25Z2021-06-10Crisigiovanni, Enzo Luigi. RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA. 2021. 147 f. Tese (Programa de Pós-Graduação em Ciências Florestais - Doutorado) - Universidade Estadual do Centro-Oeste, Irati-PR.http://tede.unicentro.br:8080/jspui/handle/jspui/1725Habitat loss and bioinvasion are the main causes of species extinctions associated with anthropic processes. One of the main tools for the Spatio temporal analysis of these processes is remote sensing. With the focus on recognizing patterns in the Mixed Ombrophylous Rainforest (MOF) at different spatial and temporal scales, three chapters were sought: (1) Classify and measure elements of the landscape of a hydrographic sub-basin in the Mixed Rainforest in Center-South Paraná state, associating the landscape components to the phytosociological parameters of the tree community; (2) Quantify the recognition rate of Hovenia dulcis Thunb. (Japanese raisin tree), invasive alien species and Araucaria angustifolia (Bertol.) Kuntze (Araucária), native conifer threatened with extinction, by machine learning algorithms; and (3) Improve the extraction of attributes for the classification of H. dulcis, in a pixel-by-pixel classification applied in a One-class-Classification (OCC) model. To achieve the objectives, the methods employed were: (1) For two WolrdVew2 images from the years 2012 and 2016, using only the multispectral bands of the sensor (2m spatial resolution), multiresolution segmentation, and Support Vector Machine (SVM). Nine landscape metrics were extracted and submitted to a Principal Component Analysis (PCA) from the most accurate classifications. These components were associated with regression models generated by Generalized Estimation Equation (GEE) for nine phytosociological parameters calculated from nine permanent forest inventory plots, with three measurements 2011, 2014, and 2017 distributed in the area of the basin in question; (2) For the recognition of A. angustifolia and H. dulcis, the visual interpretation and clipping of the treetops validated by RPA images (Remoted piloted aircraft) was performed in the 2016 WorldView-2 image fused (0.5 m spatial resolution). The cutouts had textural attributes extracted by Edge Filter and Pyramid Histogram of Oriented Gradients (PHOG), the textural attributes, and the images' pixels' spectral values were submitted to data mining using a genetic algorithm. Finally, they were submitted to two classifiers, Artificial Neural Networks (ANN) and Random Forest (RF); and (3) In two 2018 RPA images, random pixels from H. dulcis and the negative class (non-H. dulcis) were selected in a balanced way. Then, OCC models classified pixel-by-pixel were created by the Random Forest algorithm, using only the pixels' spectral attributes for the classification. In general, approximately 90% accuracy was found for all classifications used, with the different classification techniques applied to different images and targets. Concerning Chapter 1, there was a reduction in the area of MOF coverage of 6.4% and an agricultural expansion of 8.07% in the four years of analysis, mainly due to the increase in the landscape fragmentation process. It was still possible to find associations of fragmentation in the tree community, which showed changes mainly related to the increase in biomass, indicative of regeneration processes. Chapter 2, on the other hand, species classification was performed with high accuracy, reaching 95% of correct answers for Cross-Validation. RF surpassed the ANN classification rates and still proved to be more stable and faster for training and later species classification. For chapter 3, the classification of RPA images was also highly accurate, with around 95% correct answers. With only three spectral bands: Red (R), Green (G) and Blue (B), the spectral attributes of the red and green range were more relevant for the classification, and the pixel-by-pixel classification model of only one class proved efficient in the detection of H. dulcis. It was concluded that the WorldView-2 multispectral sensor and images of RPAs, associated or not, have a high potential to provide information for recognizing MOF forest patterns, ranging from landscape elements, e.g. forest remnants, to tree species canopies, such as H. dulcis and A. angustifolia. Thus, it can be effective in contributing to the conservation and sustainable management of the MOF.A perda de habitats e a bioinvasão, associadas a processos antrópicos, são as principais causas das extinções de espécies. Uma das principais ferramentas para a análise espaço-temporal desses processos é o sensoriamento remoto. Com o enfoque de reconhecer padrões na Floresta Ombrófila Mista (FOM) em diferentes escalas espaciais e temporais, em três capítulos buscou-se: (1) Classificar e mensurar elementos da paisagem de uma sub-bacia hidrográfica na Floresta Ombrófila Mista no Centro-Sul do estado do Paraná, associando os componentes da paisagem à parâmetros fitossociológicos da comunidade arbórea; (2) Quantificar a taxa de reconhecimento de Hovenia dulcis Thunb. (Uva-do-Japão), espécie exótica invasora e de Araucaria angustifolia (Bertol.) Kuntze (Araucária), conífera nativa ameaçada de extinção, por algoritmos de aprendizado de máquina; e (3) Aprimorar a extração de atributos para a classificação de H. dulcis, em um modelo de classificação pixel a pixel de apenas uma classe. Para o cumprimento dos objetivos, os métodos empregados foram: (1) Para duas imagens WolrdVew2 dos anos de 2012 e 2016, utilizando apenas as bandas multiespectrais do sensor (2m de resolução espacial) foram aplicadas a segmentação multirresolução e Máquina de Vetores de Suporte (Support Vector Machine - SVM). A partir das classificações de maior acurácia foram extraídas nove métricas da paisagem submetidas a uma Análise de Componentes Principais (Principal Component Analysis - PCA). Esses componentes foram associados a modelos de regressão gerados por Generalized Estimation Equation (GEE) para nove parâmetros fitossociológicos calculados a partir de nove parcelas permanentes de inventário florestal, com três medições 2011, 2014 e 2017 distribuídas na área da sub-bacia em questão; (2) Para o reconhecimento de A. angustifolia e H. dulcis foi realizada a interpretação visual e recorte das copas das árvores validadas por imagens de RPA (Remotely Piloted Aircraft) na imagem WorldView-2 de 2016 fusionada (0,5m de resolução espacial). Os recortes tiveram atributos texturais extraídos por filtros de bordas (Edge Filter) e por Histogramas de Gradiente Orientados em Pirâmide (Pyramid Histogram of Oriented Gradients – PHOG). Os atributos texturais e os valores espectrais dos pixels das imagens foram submetidos a mineração de dados utilizando algoritmo genético. Finalmente foram submetidos à dois classificadores, Redes Neurais Artificiais (Artificial Neural Network - ANN) e Floresta de Decisão Randômica (Ramdom Forest - RF) e; (3) Em duas imagens de RPA de 2018 foram selecionados pixels aleatórios de H. dulcis e da classe negativa (não-H. dulcis), de maneira balanceada. Em seguida, foram criados modelos de classificação pixel a pixel de apenas uma classe (One-class classification), pelo algoritmo Random Forest, utilizando apenas os atributos espectrais dos pixels para a classificação. De maneira geral encontrou-se elevada acurácia, aproximadamente 90%, para todas as classificações empregadas, com as diferentes técnicas de classificação aplicadas às diferentes imagens e alvos. Com relação ao Capítulo 1, nos quatro anos de análise encontrou-se uma redução na área de cobertura de FOM de 6,4% e uma expansão agrícola de 8,07%, devido principalmente ao aumento do processo de fragmentação da paisagem. Ainda foi possível encontrar associações da fragmentação na comunidade arbórea, a qual apresentou mudanças principalmente relacionadas ao incremento de biomassa, indicando processos de regeneração. No Capítulo 2 a classificação das espécies foi realizada com alta acurácia, chegando a 95% de acertos para o Cross-Validation. RF superou as taxas de classificação da ANN e ainda se mostrou mais estável e veloz para o treinamento e posterior classificação das espécies. Para o Capítulo 3, a classificação das imagens de RPA também apresentaram elevada acurácia, com cerca de 95% de acertos. Com apenas três bandas espectrais: Red (R), Green (G) e Blue (B) os atributos espectrais da faixa do vermelho e do verde se mostraram mais relevantes para a classificação, e o modelo de classificação pixel a pixel de apenas uma classe se mostrou eficiente na detecção de H. dulcis. Concluiu-se que o sensor multiespectral WorldView-2 e a imagens de RPAs, associadas ou não, têm elevado potencial de fornecer informações para o reconhecimento de padrões florestais da FOM, variando de elementos da paisagem, e.g. remanescentes florestais, até copas de espécies arbóreas, como H. dulcis e A. angustifolia. Dessa forma, esses instrumentos podem ser efetivos na contribuição da conservação e manejo sustentável da FOM.Submitted by Fabiano Jucá (fjuca@unicentro.br) on 2021-11-05T12:46:25Z No. of bitstreams: 1 tese - Enzo Luigi Crisigiovanni.pdf: 5631570 bytes, checksum: 70a9c9e065cd3c5a5e59ebfc17860036 (MD5)Made available in DSpace on 2021-11-05T12:46:25Z (GMT). No. of bitstreams: 1 tese - Enzo Luigi Crisigiovanni.pdf: 5631570 bytes, checksum: 70a9c9e065cd3c5a5e59ebfc17860036 (MD5) Previous issue date: 2021-06-10Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPESapplication/pdfhttp://tede.unicentro.br:8080/jspui/retrieve/7324/tese%20-%20Enzo%20Luigi%20Crisigiovanni.pdf.jpghttp://tede.unicentro.br:8080/jspui/retrieve/7325/tese%20-%20Enzo%20Luigi%20Crisigiovanni.pdf.jpgporUniversidade Estadual do Centro-OestePrograma de Pós-Graduação em Ciências Florestais (Doutorado)UNICENTROBrasilUnicentro::Departamento de Ciências FlorestaisSensoriamento remotoAprendizado de máquinasEcologia da paisagemBioinvasãoRemote sensingMachine learningLandscape ecologyBioinvasionCIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTALRECURSOS FLORESTAIS E ENGENHARIA FLORESTAL::MANEJO FLORESTALRECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTAinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesis2828826774026714864600600600600600-5938256993918186975-604049389552879283-76664982548871839512075167498588264571info:eu-repo/semantics/openAccessreponame:Biblioteca Digital de Teses e Dissertações do UNICENTROinstname:Universidade Estadual do Centro-Oeste (UNICENTRO)instacron:UNICENTROTHUMBNAILtese - Enzo Luigi Crisigiovanni.pdf.jpgtese - Enzo Luigi Crisigiovanni.pdf.jpgimage/jpeg2142http://localhost:8080/tede/bitstream/jspui/1725/4/tese+-+Enzo+Luigi+Crisigiovanni.pdf.jpgdfc695e5377b2154eb02bf7e5896ab1eMD54THUMBNAILtese - Enzo Luigi Crisigiovanni.pdf.jpgtese - Enzo Luigi Crisigiovanni.pdf.jpgimage/jpeg2142http://localhost:8080/tede/bitstream/jspui/1725/4/tese+-+Enzo+Luigi+Crisigiovanni.pdf.jpgdfc695e5377b2154eb02bf7e5896ab1eMD54TEXTtese - Enzo Luigi Crisigiovanni.pdf.txttese - Enzo Luigi Crisigiovanni.pdf.txttext/plain248544http://localhost:8080/tede/bitstream/jspui/1725/3/tese+-+Enzo+Luigi+Crisigiovanni.pdf.txt0c8254c1123d526e9215b46e876fc441MD53TEXTtese - Enzo Luigi Crisigiovanni.pdf.txttese - Enzo Luigi Crisigiovanni.pdf.txttext/plain248544http://localhost:8080/tede/bitstream/jspui/1725/3/tese+-+Enzo+Luigi+Crisigiovanni.pdf.txt0c8254c1123d526e9215b46e876fc441MD53ORIGINALtese - Enzo Luigi Crisigiovanni.pdftese - Enzo Luigi Crisigiovanni.pdfapplication/pdf5631570http://localhost:8080/tede/bitstream/jspui/1725/2/tese+-+Enzo+Luigi+Crisigiovanni.pdf70a9c9e065cd3c5a5e59ebfc17860036MD52LICENSElicense.txtlicense.txttext/plain; charset=utf-82003http://localhost:8080/tede/bitstream/jspui/1725/1/license.txt6544a715df32d52b08aa3def94c4dddeMD51jspui/17252021-11-25 13:02:33.015oai:localhost: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Biblioteca Digital de Teses e Dissertaçõeshttp://tede.unicentro.br:8080/jspui/PUBhttp://tede.unicentro.br/tde_oai/oai3.phprepositorio@unicentro.br||fabianoqueiroz@yahoo.com.bropendoar:2021-11-25T15:02:33Biblioteca Digital de Teses e Dissertações do UNICENTRO - Universidade Estadual do Centro-Oeste (UNICENTRO)false |
dc.title.por.fl_str_mv |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
title |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
spellingShingle |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA Crisigiovanni, Enzo Luigi Sensoriamento remoto Aprendizado de máquinas Ecologia da paisagem Bioinvasão Remote sensing Machine learning Landscape ecology Bioinvasion CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL::MANEJO FLORESTAL |
title_short |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
title_full |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
title_fullStr |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
title_full_unstemmed |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
title_sort |
RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA |
author |
Crisigiovanni, Enzo Luigi |
author_facet |
Crisigiovanni, Enzo Luigi |
author_role |
author |
dc.contributor.advisor1.fl_str_mv |
Figueiredo Filho, Afonso |
dc.contributor.advisor1Lattes.fl_str_mv |
http://lattes.cnpq.br/4151544991447365 |
dc.contributor.advisor-co1.fl_str_mv |
Pesck, Vagner Alex |
dc.contributor.advisor-co1Lattes.fl_str_mv |
http://lattes.cnpq.br/4670102252045134 |
dc.contributor.advisor-co2.fl_str_mv |
Lima, Vanderlei Aparecido de |
dc.contributor.advisor-co2Lattes.fl_str_mv |
http://lattes.cnpq.br/9090461949264421 |
dc.contributor.authorID.fl_str_mv |
069.450.439-40 |
dc.contributor.authorLattes.fl_str_mv |
http://lattes.cnpq.br/3734724830416138 |
dc.contributor.author.fl_str_mv |
Crisigiovanni, Enzo Luigi |
contributor_str_mv |
Figueiredo Filho, Afonso Pesck, Vagner Alex Lima, Vanderlei Aparecido de |
dc.subject.por.fl_str_mv |
Sensoriamento remoto Aprendizado de máquinas Ecologia da paisagem Bioinvasão |
topic |
Sensoriamento remoto Aprendizado de máquinas Ecologia da paisagem Bioinvasão Remote sensing Machine learning Landscape ecology Bioinvasion CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL::MANEJO FLORESTAL |
dc.subject.eng.fl_str_mv |
Remote sensing Machine learning Landscape ecology Bioinvasion |
dc.subject.cnpq.fl_str_mv |
CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL::MANEJO FLORESTAL |
description |
Habitat loss and bioinvasion are the main causes of species extinctions associated with anthropic processes. One of the main tools for the Spatio temporal analysis of these processes is remote sensing. With the focus on recognizing patterns in the Mixed Ombrophylous Rainforest (MOF) at different spatial and temporal scales, three chapters were sought: (1) Classify and measure elements of the landscape of a hydrographic sub-basin in the Mixed Rainforest in Center-South Paraná state, associating the landscape components to the phytosociological parameters of the tree community; (2) Quantify the recognition rate of Hovenia dulcis Thunb. (Japanese raisin tree), invasive alien species and Araucaria angustifolia (Bertol.) Kuntze (Araucária), native conifer threatened with extinction, by machine learning algorithms; and (3) Improve the extraction of attributes for the classification of H. dulcis, in a pixel-by-pixel classification applied in a One-class-Classification (OCC) model. To achieve the objectives, the methods employed were: (1) For two WolrdVew2 images from the years 2012 and 2016, using only the multispectral bands of the sensor (2m spatial resolution), multiresolution segmentation, and Support Vector Machine (SVM). Nine landscape metrics were extracted and submitted to a Principal Component Analysis (PCA) from the most accurate classifications. These components were associated with regression models generated by Generalized Estimation Equation (GEE) for nine phytosociological parameters calculated from nine permanent forest inventory plots, with three measurements 2011, 2014, and 2017 distributed in the area of the basin in question; (2) For the recognition of A. angustifolia and H. dulcis, the visual interpretation and clipping of the treetops validated by RPA images (Remoted piloted aircraft) was performed in the 2016 WorldView-2 image fused (0.5 m spatial resolution). The cutouts had textural attributes extracted by Edge Filter and Pyramid Histogram of Oriented Gradients (PHOG), the textural attributes, and the images' pixels' spectral values were submitted to data mining using a genetic algorithm. Finally, they were submitted to two classifiers, Artificial Neural Networks (ANN) and Random Forest (RF); and (3) In two 2018 RPA images, random pixels from H. dulcis and the negative class (non-H. dulcis) were selected in a balanced way. Then, OCC models classified pixel-by-pixel were created by the Random Forest algorithm, using only the pixels' spectral attributes for the classification. In general, approximately 90% accuracy was found for all classifications used, with the different classification techniques applied to different images and targets. Concerning Chapter 1, there was a reduction in the area of MOF coverage of 6.4% and an agricultural expansion of 8.07% in the four years of analysis, mainly due to the increase in the landscape fragmentation process. It was still possible to find associations of fragmentation in the tree community, which showed changes mainly related to the increase in biomass, indicative of regeneration processes. Chapter 2, on the other hand, species classification was performed with high accuracy, reaching 95% of correct answers for Cross-Validation. RF surpassed the ANN classification rates and still proved to be more stable and faster for training and later species classification. For chapter 3, the classification of RPA images was also highly accurate, with around 95% correct answers. With only three spectral bands: Red (R), Green (G) and Blue (B), the spectral attributes of the red and green range were more relevant for the classification, and the pixel-by-pixel classification model of only one class proved efficient in the detection of H. dulcis. It was concluded that the WorldView-2 multispectral sensor and images of RPAs, associated or not, have a high potential to provide information for recognizing MOF forest patterns, ranging from landscape elements, e.g. forest remnants, to tree species canopies, such as H. dulcis and A. angustifolia. Thus, it can be effective in contributing to the conservation and sustainable management of the MOF. |
publishDate |
2021 |
dc.date.accessioned.fl_str_mv |
2021-11-05T12:46:25Z |
dc.date.issued.fl_str_mv |
2021-06-10 |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/doctoralThesis |
format |
doctoralThesis |
status_str |
publishedVersion |
dc.identifier.citation.fl_str_mv |
Crisigiovanni, Enzo Luigi. RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA. 2021. 147 f. Tese (Programa de Pós-Graduação em Ciências Florestais - Doutorado) - Universidade Estadual do Centro-Oeste, Irati-PR. |
dc.identifier.uri.fl_str_mv |
http://tede.unicentro.br:8080/jspui/handle/jspui/1725 |
identifier_str_mv |
Crisigiovanni, Enzo Luigi. RECONHECIMENTO DE PADRÕES ESPAÇO-TEMPORAIS EM IMAGENS WorldView-2 E RPAs POR APRENDIZADO DE MÁQUINA EM FLORESTA OMBRÓFILA MISTA. 2021. 147 f. Tese (Programa de Pós-Graduação em Ciências Florestais - Doutorado) - Universidade Estadual do Centro-Oeste, Irati-PR. |
url |
http://tede.unicentro.br:8080/jspui/handle/jspui/1725 |
dc.language.iso.fl_str_mv |
por |
language |
por |
dc.relation.program.fl_str_mv |
2828826774026714864 |
dc.relation.confidence.fl_str_mv |
600 600 600 600 600 |
dc.relation.department.fl_str_mv |
-5938256993918186975 |
dc.relation.cnpq.fl_str_mv |
-604049389552879283 -7666498254887183951 |
dc.relation.sponsorship.fl_str_mv |
2075167498588264571 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Universidade Estadual do Centro-Oeste |
dc.publisher.program.fl_str_mv |
Programa de Pós-Graduação em Ciências Florestais (Doutorado) |
dc.publisher.initials.fl_str_mv |
UNICENTRO |
dc.publisher.country.fl_str_mv |
Brasil |
dc.publisher.department.fl_str_mv |
Unicentro::Departamento de Ciências Florestais |
publisher.none.fl_str_mv |
Universidade Estadual do Centro-Oeste |
dc.source.none.fl_str_mv |
reponame:Biblioteca Digital de Teses e Dissertações do UNICENTRO instname:Universidade Estadual do Centro-Oeste (UNICENTRO) instacron:UNICENTRO |
instname_str |
Universidade Estadual do Centro-Oeste (UNICENTRO) |
instacron_str |
UNICENTRO |
institution |
UNICENTRO |
reponame_str |
Biblioteca Digital de Teses e Dissertações do UNICENTRO |
collection |
Biblioteca Digital de Teses e Dissertações do UNICENTRO |
bitstream.url.fl_str_mv |
http://localhost:8080/tede/bitstream/jspui/1725/4/tese+-+Enzo+Luigi+Crisigiovanni.pdf.jpg http://localhost:8080/tede/bitstream/jspui/1725/4/tese+-+Enzo+Luigi+Crisigiovanni.pdf.jpg http://localhost:8080/tede/bitstream/jspui/1725/3/tese+-+Enzo+Luigi+Crisigiovanni.pdf.txt http://localhost:8080/tede/bitstream/jspui/1725/3/tese+-+Enzo+Luigi+Crisigiovanni.pdf.txt http://localhost:8080/tede/bitstream/jspui/1725/2/tese+-+Enzo+Luigi+Crisigiovanni.pdf http://localhost:8080/tede/bitstream/jspui/1725/1/license.txt |
bitstream.checksum.fl_str_mv |
dfc695e5377b2154eb02bf7e5896ab1e dfc695e5377b2154eb02bf7e5896ab1e 0c8254c1123d526e9215b46e876fc441 0c8254c1123d526e9215b46e876fc441 70a9c9e065cd3c5a5e59ebfc17860036 6544a715df32d52b08aa3def94c4ddde |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 MD5 MD5 MD5 MD5 |
repository.name.fl_str_mv |
Biblioteca Digital de Teses e Dissertações do UNICENTRO - Universidade Estadual do Centro-Oeste (UNICENTRO) |
repository.mail.fl_str_mv |
repositorio@unicentro.br||fabianoqueiroz@yahoo.com.br |
_version_ |
1801859386414465024 |