|
|
|
Abstract
Selection of the Best WaveletA in Smoothing Spectral Reflectance Data from Soil Spectra Using Statistical Method
Basharudin Abdul Hadi
UiTM Kampus Perlis,
Malaysia
Email: Basha_uitma@yahoo.co.uk
The main purpose of field spectroradiometer is to calibrate airborne and satellite remote sensing signals from various sources for the purpose of target identification and classification. Field spectoradiometer measures the reflectance value of soil on earth with sun as the source of energy. However, when we are using field spectroradiometer in an uncontrolled real-world environment, the spectra measured could be subjected to various ¡®spectral noise¡¯ sources. In this paper, we are looking at the use of a new mathematical algorithm based on wavelet transform in order to analyse the presence of noise in field spectroradiometer data. The results indicate that there are significant noise levels in the data especially after derivative analysis. The use of different wavelets are also investigated in order to evaluate their performance in denoising. In general, wavelet-based denoising is effective in reducing noise in the data, thus allowing more advanced technique to be conducted such as the spectral derivative analysis. The performance of the different wavelets however will differ depending on the characteristics of the wavelet functions and the spectral curve patterns. Then, to validate the result in order to search out the best suitable wavelet in de noising we use statistical method to varify the data. From the results it can be concluded, statistical method have a good potential in selection of the best wavelet in de noising process.
|
|
|