A carregar...
Publicação
Automatização de Software Open Source (WEBODM) para Modelação e Mapeamento de Barragens com Drones
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 22.2 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Os sistemas de monitorização da saúde estrutural das barragens são cruciais para identificar
comportamentos anómalos e desta forma minimizar ou eliminar os seus efeitos.
A caraterização das barragens, no que diz respeito aos princípios de engenharia, é viabilizada
com o uso de equipamentos avançados de monitorização como por exemplo
piezómetros e medidores de caudal, que descrevem o desempenho mecânico das mesmas,
evitando desastres naturais e mitigando riscos socioeconómicos.
A colocação de um número limitado de pontos de monitorização é, geralmente, insuficiente
para barragens de grande porte, uma vez que muitas das áreas das barragens são
inacessíveis. A solução encontrada para esta dificuldade passa pelo uso de métodos em
larga escala e sem contacto. Consequentemente, as tecnologias baseadas na visão têm sido
consideradas como uma abordagem eficiente para a monitorização da saúde estrutural
das barragens. Sensores baseados em visão computacional, combinados com o constante
aperfeiçoamento na resolução digital e capacidade de computação, surgiram como uma
ferramenta promissora para a aferição remota de estruturas.
Outra aplicação das imagens é a criação de modelos tridimensionais (3D) das estruturas.
Esta metodologia consiste numa renderização baseada em imagens, capaz de identificar
danos e realizar medições óticas sem contacto, que podem ser aplicadas na prevenção
de desastres. Para obter o número necessário de imagens do objeto, em tempo útil e a
baixo custo, é necessária uma plataforma flexível que possa ser controlada remotamente
e suportar diversos tipos de configurações.
Devido aos recentes avanços no uso e disponibilidade de plataformas de Veículos
Aéreos Não Tripulados (VANT) e o desenvolvimento de software de processamento de
imagem de fácil manuseamento, a fotogrametria baseada em VANT está a ser adotada
cada vez mais para produzir topografia de alta resolução com o intuito de estudar alterações
nas superfícies. Os VANT equipados com câmeras são adequados para o levantamento
da superfície do terreno, devido à geração de topografia de alta resolução para
documentar as caraterísticas topográficas em tempo real, pelo que permitem a avaliação
da integridade da infraestrutura e a deteção de danos de grandes estruturas.
Tendo em conta que um dos fatores com maior influência na precisão de um modelo tridimensional é o uso de Ground Control Points (GCPs), o seu número e distribuição
pelo terreno são importantíssimos. Torna-se necessário colocar na superfície da barragem
um número específico de GCPs, de forma a obter um modelo georreferenciado de elevada
precisão e à escala.
Nesta dissertação, propõe-se desenvolver um processo automático de identificação
de GCPs usando fotogrametria de VANT, para deteção de danos em barragens e monitorização
de emergências. Este processo acabará por aumentar a eficiência, bem como a
precisão, na criação de ortofotos e no desenvolvimento de modelos tridimensionais (3D)
dessas estruturas.
Dam health monitoring systems are crucial for identifying anomalous behaviours and eliminating or minimising their effects. Advanced monitoring equipment like piezometers and flow meters allows the characterization of the dams, regarding the principles of engineering, and describes their mechanical performance, thus avoiding natural disasters and mitigating socio-economic risk. The placement of a limited number of monitoring points is usually insufficient for large-scale dams, since many areas of the dams are inaccessible. Large-scale and noncontact methods applied for dam health monitoring presented themselves as a solution. Vision-based technologies have been considered as an efficient approach for structural health monitoring. Camera and computer vision-based sensors combined with the improvements in camera resolution and computation ability, have emerged as a promising tool for the non-contact remote measurement of structural responses. Another application of images is the creation of three-dimensional (3D) models of the structures. This methodology consists of an image-based rendering capable of identifying damage and performing non-contact, optical-based measurements that can be applied for disaster prevention. In order to obtain the necessary number of photographs of the object, in a timely and inexpensive manner, a flexible platform is required that should be controlled remotely and support several types of payloads. Due to the recent advances in the use and availability of Unmanned Aerial Vehicle (UAV) platforms and the development of easy-to-operate image processing software, UAV-based photogrammetry is being adopted increasingly to produce high-resolution topography for studying surface processes. UAVs equipped with cameras are suitable for surveying the terrain surface due to the generation of high-resolution topography to document the real-time topographic features. The infrastructure integrity assessment and damage detection of large structures are often performed by Unmanned Aerial Vehicles (UAVs). A specific number of GCPs, which are usually measured by traditional surveying instruments, need to be deployed on the object to generate the UAV-based model with the geolocation and scale. One of the factors with the greatest influence on the accuracy of a UAV is the number and distribution of the GCPs. In this thesis, a semi-automated process of Ground Control Point (GCP) identification using UAV photogrammetry is proposed for dam damage detection and emergency monitoring. This process will ultimately increase efficiency, as well as accuracy, in the creation of orthophotos, and in the development of three-dimensional (3D) models of these structures.
Dam health monitoring systems are crucial for identifying anomalous behaviours and eliminating or minimising their effects. Advanced monitoring equipment like piezometers and flow meters allows the characterization of the dams, regarding the principles of engineering, and describes their mechanical performance, thus avoiding natural disasters and mitigating socio-economic risk. The placement of a limited number of monitoring points is usually insufficient for large-scale dams, since many areas of the dams are inaccessible. Large-scale and noncontact methods applied for dam health monitoring presented themselves as a solution. Vision-based technologies have been considered as an efficient approach for structural health monitoring. Camera and computer vision-based sensors combined with the improvements in camera resolution and computation ability, have emerged as a promising tool for the non-contact remote measurement of structural responses. Another application of images is the creation of three-dimensional (3D) models of the structures. This methodology consists of an image-based rendering capable of identifying damage and performing non-contact, optical-based measurements that can be applied for disaster prevention. In order to obtain the necessary number of photographs of the object, in a timely and inexpensive manner, a flexible platform is required that should be controlled remotely and support several types of payloads. Due to the recent advances in the use and availability of Unmanned Aerial Vehicle (UAV) platforms and the development of easy-to-operate image processing software, UAV-based photogrammetry is being adopted increasingly to produce high-resolution topography for studying surface processes. UAVs equipped with cameras are suitable for surveying the terrain surface due to the generation of high-resolution topography to document the real-time topographic features. The infrastructure integrity assessment and damage detection of large structures are often performed by Unmanned Aerial Vehicles (UAVs). A specific number of GCPs, which are usually measured by traditional surveying instruments, need to be deployed on the object to generate the UAV-based model with the geolocation and scale. One of the factors with the greatest influence on the accuracy of a UAV is the number and distribution of the GCPs. In this thesis, a semi-automated process of Ground Control Point (GCP) identification using UAV photogrammetry is proposed for dam damage detection and emergency monitoring. This process will ultimately increase efficiency, as well as accuracy, in the creation of orthophotos, and in the development of three-dimensional (3D) models of these structures.
Descrição
Palavras-chave
Open source Modelos 3D Mapeamento Aéreo Barragens Veículos Aéreos Não Tripulados (Drones) Pontos de Controlo