2-D Fourier Phrase Analysis On Target Cloud And A Study On Its Characteristics

Fourier phase analysis is proposed as a new technique to address the problem of "sub-pixel motion" which is met in deriving cloud motion wind (CMW) from high temporal resolution images by the maximum correlation technique. The technique is more efficient than the maximum correlation because only two templates in same size are involved in computing procedure. Moreover, it obtains not only CMW vectors but also velocity spectrum and variance.However, the noises in phase-difference image affect the precision of calculation of CMW. The filters for eliminating the noise are presented. And FFT is used in Fourier phase analysis to accelerate the calculating velocity.This article is composed of six parts. In the first part, a summary on the research progress and significance of cloud motion wind based on literature is introduced and the purpose of this study is stated. In the second and third parts, the 1-D and 2-D Fourier analysis techniques based on fluctuation theory and Fourier phase analysis are described, respectively, for application in calculating cloud motion wind vectors. In the fourth part, the causes of the noises in phase-difference image are analyzed and the filters for eliminating the noise are presented. A consecutive satellite images with one-minute interval are used for experiment to track clouds with the two-Dim Fourier phase analysis and noise filters. The result of wind vectors is obtained after quality control. In the sixth part, we concluded our research and put up with shortages of this article.