Abstract
Remote sensing instrumentations signal electric-magnetic waves and detect their reflections to obtain the characteristics of objects. There is no requirement of physical contacts between the probes and the objects. This feature contributes to the cost control and enhancing the capability of wave monitoring since the direct impacts from the relevant environmental loads to the instrumentations can be avoided.Reflections of electric-magnetic signals are represented as grey-scale images, in which information of waves can be extracted. Due to the topography effects, however, the wave images are often pronouns high non-homogeneity. To retrieve the wave characteristics from those wave field images, Wavelet Transform Analysis is adopted in present study to establish the 2-D Wavelet Transform directional wave-number spectrum estimation method. Unlike traditional Fourier Transform Method, Wavelet Transform Analysis is capable of constructing the spatial and spectral structures from non-homogeneous images. Moreover, the optimal resolution of both the spatial and spectral domain could be yielded by adjusting the scale and sifting the location of mother wavelet function. The relationship between the Wavelet coefficients and wave spectral parameters as well as the size of modulated window and the wave image are derived accordingly afterward. From the numerical simulation tests, the validity and feasibility of proposed method are examined.
To justify the accuracy of this method, images of two central wave sources conditions and random wave fields are simulated and used as the inputs of propose method. From the results, it is suggested that the distance between the modulated window centers to the image border should not less than 1.08 times the peak frequency wave length.
Furthermore, the parameter n0 of the Morlet mother wavelet function should be selected in the range 5~10 to obtain optimal resolution in both spatial and spectral domain.
| Date of Award | 2002 |
|---|---|
| Original language | English |
| Supervisor | Zsu-Hsin Chuang (Supervisor) |