A carregar...
Publicação
Metodologias de avaliação e caracterização das condições fluviais com recurso a veículos aéreos não tripulados
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 7.71 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A caracterização das condições fluviais providencia elementos que integram as
avaliações de eficácia dos Regimes de Caudais Ecológicos em aproveitamentos
hidroelétricos, concebidas para minimizar os impactes da alteração do regime hidrológico
a jusante de barragens. Esta caracterização pode ser melhorada através da utilização de
Veículos Aéreos não Tripulados (VAnT): reduzindo alguma subjetividade na recolha de
elementos no terreno, imposta por limitações de segurança e de acessibilidade em alguns
troços fluviais; efetuando análises temporais das alterações ocorridas; cartografando os
diversos elementos fluviais através da criação de mapas. Os objetivos centram-se no
desenvolvimento de um conjunto de procedimentos metodológicos automáticos e
semiautomáticos, de modo a caracterizar as alterações nos troços fluviais a jusante de
barragens, sendo, progressivamente, aplicados a diferentes troços em contextos
diversificados, e introduzidos nas caracterizações fluviais tradicionais. Assim, foram
elaboradas diversas tarefas: compilação de metodologias para caracterizar as condições
fluviais; elaboração de mapas, utilizando ferramentas e softwares de Sistemas de
Informação Geográfica (SIG), provenientes de ortofotos, vídeos e imagens LiDAR,
obtidos na área de estudo; utilização dos dados obtidos por VAnT para analisar as
características hidrogeomorfológicas e fitogeográficas com o pós-processamento das
imagens nos SIG. Os métodos e técnicas utilizados foram a obtenção de imagens de alta
definição através de VAnT e o seu processamento no software Pix4D. Foram também
utilizadas ferramentas em softwares de SIG para classificar os ortomosaicos, através de
classificações semiautomáticas com algoritmos no ArcGIS Pro. Verificam-se alguns
problemas nos resultados obtidos nas classificações, devido às diversas orientações das
vertentes do vale e, consequentemente, aos contrastes na incidência da radiação solar.
Apesar disso, identificou-se a classificação supervisionada do ArcGIS Pro com o
algoritmo Support Vector Machine e os atributos de segmentação Active chromaticity
color e Mean digital number, como o método mais eficaz para caracterizar e classificar
semiautomaticamente as condições fluviais neste tipo de relevo.
The characterization of river conditions provides elements that are part of the effectiveness assessments of Environmental Flow Regimes in hydroelectric dams, that aims at minimizing the impacts of altering the hydrological regime of dams downstream. This characterization can be improved using Unmanned Aerial Vehicles (UAVs): reducing some subjectivity in field data collection, imposed by safety and accessibility limitations on some river stretches; allowing temporal analysis of the changes occurred in river sections, as well as a more rigorous characterization by mapping various river elements. The objectives are focused on developing a set of automatic and semi-automatic methodological procedures to characterize changes in river sections downstream of dams, which can be progressively applied to different sections in diverse contexts and introduced into traditional river characterizations. Therefore, several tasks were developed: compilation of methodologies to characterize river conditions; creation of maps, using Geographic Information Systems (GIS) tools, derived from orthophotos, videos and LiDAR images, obtained in the study area; use of the data to analyze the hydrogeomorphological and phytogeographical characteristics with the post-processing of the images in GIS. The methods and techniques applied in this project consisted of the utilization of high-resolution images using UAVs, that were processed in Pix4D software to create orthomosaics and Digital Surface Models. Additionally, GIS software tools were used to classify the orthomosaics through semi-automatic classifications with algorithms in ArcGIS Pro. The analysis revealed some problems with the results obtained from this classification, due to the different valley slopes orientation which leads to a high variability of solar radiation incidence. Despite this, it was determined that the supervised classification of ArcGIS Pro, using the Support Vector Machine algorithm, the Active Chromaticity Color and Mean Digital Number segmentation attributes are the most effective methods for semi-automatically characterizing and classifying river conditions in this type of terrain.
The characterization of river conditions provides elements that are part of the effectiveness assessments of Environmental Flow Regimes in hydroelectric dams, that aims at minimizing the impacts of altering the hydrological regime of dams downstream. This characterization can be improved using Unmanned Aerial Vehicles (UAVs): reducing some subjectivity in field data collection, imposed by safety and accessibility limitations on some river stretches; allowing temporal analysis of the changes occurred in river sections, as well as a more rigorous characterization by mapping various river elements. The objectives are focused on developing a set of automatic and semi-automatic methodological procedures to characterize changes in river sections downstream of dams, which can be progressively applied to different sections in diverse contexts and introduced into traditional river characterizations. Therefore, several tasks were developed: compilation of methodologies to characterize river conditions; creation of maps, using Geographic Information Systems (GIS) tools, derived from orthophotos, videos and LiDAR images, obtained in the study area; use of the data to analyze the hydrogeomorphological and phytogeographical characteristics with the post-processing of the images in GIS. The methods and techniques applied in this project consisted of the utilization of high-resolution images using UAVs, that were processed in Pix4D software to create orthomosaics and Digital Surface Models. Additionally, GIS software tools were used to classify the orthomosaics through semi-automatic classifications with algorithms in ArcGIS Pro. The analysis revealed some problems with the results obtained from this classification, due to the different valley slopes orientation which leads to a high variability of solar radiation incidence. Despite this, it was determined that the supervised classification of ArcGIS Pro, using the Support Vector Machine algorithm, the Active Chromaticity Color and Mean Digital Number segmentation attributes are the most effective methods for semi-automatically characterizing and classifying river conditions in this type of terrain.
Descrição
Palavras-chave
Veículo Aéreo Não Tripulado Classificação semiautomática Condições Fluviais Caudal Ecológico Troço fluvial