| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.1 MB | Adobe PDF |
Advisor(s)
Abstract(s)
O presente estudo tem como objetivo testar e validar um conjunto de técnicas espectrais aplicadas a
imagens de satélite SAR (Synthetic Aperture Radar) em que sejam bem visíveis as condições de
ondulação da superfície do oceano, permitindo estimar a batimetria com base na deteção da variação
dos comprimentos de onda.
Este trabalho visa melhorar a compreensão do comportamento das ondas de gravidade, que é
extremamente importante para a previsão da hidrodinâmica costeira e para a correta caracterização da
morfologia do fundo do mar.
A análise espectral surge como uma ferramenta para apoiar a nossa compreensão e observar o
comportamento das ondas geradas pelo vento. Aplicou-se um conjunto de algoritmos elaborados em
Matlab para manipular imagens de satélite SAR e obter informações sobre comprimentos de onda em
vários pontos ao longo da área de estudo. Reconhece-se que a Transformada de Fourier (FFT - Fast
Fourier Transform) apresenta algumas limitações quando se trata de ondas não lineares e não
estacionárias. Assim, este estudo analisa outras técnicas espectrais com o objetivo de superar as
limitações da FFT, fazendo uso da Transformada por Ôndulas e da Transformada de Hilbert-Huang
(HHT - Hilbert-Huang Transform).
A título de exemplo, analisou-se uma imagem de satélite SAR da região de Aveiro, localizada na costa
noroeste portuguesa. A data da aquisição da imagem de satélite é 17 de fevereiro de 2015,
correspondendo a um dia com pouco vento, permitindo este tipo de análise. Embora esta metodologia
esteja numa fase inicial, as técnicas espectrais aplicadas parecem fornecer uma boa resposta nas zonas
de águas pouco profundas, corroborando e ampliando trabalhos anteriores desenvolvidos dentro deste
âmbito.
The present study aims to test and validate a set of spectral techniques applied to SAR (Synthetic Aperture Radar) satellite images in which ocean surface ripple conditions are well visible, allowing estimation of bathymetry based on the detection of variation of wavelengths. This work aims to improve the understanding of gravity wave behavior, which is extremely important for the prediction of coastal hydrodynamics and for the correct characterization of seafloor morphology. Spectral analysis emerges as a tool to support our understanding and observe the behavior of wind generated waves. A set of algorithms elaborated in Matlab was used to manipulate SAR satellite images and to obtain information of the wavelengths at various points along the study area. It is recognized that the Fourier Transform (FFT - Fast Fourier Transform) has some limitations when it comes to non-linear and non-stationary waves. Thus, this study analyzes other spectral techniques with the objective of overcoming the limitation of the FFT, making use of the Wavelet Transform and the Hilbert-Huang Transform (HHT). As an example, a SAR satellite image of the region of Aveiro, located on the northwest coast of Portugal, was analyzed. The date of acquisition of the satellite image is February 17, 2015, corresponding to a day with little wind, allowing this type of analysis. Although this methodology is at an early stage, the applied spectral techniques seem to provide a good response in the shallow water zones, corroborating and amplifying previous works developed within this scope.
The present study aims to test and validate a set of spectral techniques applied to SAR (Synthetic Aperture Radar) satellite images in which ocean surface ripple conditions are well visible, allowing estimation of bathymetry based on the detection of variation of wavelengths. This work aims to improve the understanding of gravity wave behavior, which is extremely important for the prediction of coastal hydrodynamics and for the correct characterization of seafloor morphology. Spectral analysis emerges as a tool to support our understanding and observe the behavior of wind generated waves. A set of algorithms elaborated in Matlab was used to manipulate SAR satellite images and to obtain information of the wavelengths at various points along the study area. It is recognized that the Fourier Transform (FFT - Fast Fourier Transform) has some limitations when it comes to non-linear and non-stationary waves. Thus, this study analyzes other spectral techniques with the objective of overcoming the limitation of the FFT, making use of the Wavelet Transform and the Hilbert-Huang Transform (HHT). As an example, a SAR satellite image of the region of Aveiro, located on the northwest coast of Portugal, was analyzed. The date of acquisition of the satellite image is February 17, 2015, corresponding to a day with little wind, allowing this type of analysis. Although this methodology is at an early stage, the applied spectral techniques seem to provide a good response in the shallow water zones, corroborating and amplifying previous works developed within this scope.
Description
Keywords
Imagens Análise espetral Comprimento de onda Transformada por Ôndulas Transformada de Hilbert-Huang Batimetria Images Spectral analysis Wavelength Wavelet Transform Hilbert-Huang Transform Bathymetry