In this paper, Fourier analysis is applied to multitemporal global vegetation index (GVI) data forextracting the parameters which represent the GVI monthly change. The monthly GVI data is prepared by selecting maximum value in pixels of weekly GVI data for removing cloud effect. The averaged monthly GVI data is made by averaging monthly GVI data through 3 years (April 1982-March 1985), and are used for Fourier analysis. The phase of the first order, the power spectrum of the main, the first, the second and the third order are calculated. It is found that the phase of the first order is used for finding the peak of the monthly NVI data and can be used for monitoring the variation of the season. The usefulness of the power spectrum map of each order and the phase map of the first order is also shown. The paper indicates that these parameters will be used for monitoring vegetation change yearly.
This study refers to a new method to estimate the snow depth in dam basin using Landsat TM data and digital terrain data. Analysis area (433km2, Ouyubari-dam basin) is located in mountain area in Hokkaido, Japan. The correlation analysis between snow depth and TM data was carried out. As a result of this analysis, it became clear that there was a close correlation between snow depth and band-1, 2, 3, 4 data only in the area of the mixed forest. Furthermore, the multiple regression analysis was carried out. In this analysis, the ground truth data of snow depth and S1, S2, B12 data were used. S1 and S2 values can be obtained by HSI transformation of TM data, and B12 value was calculated by ratio process using TM data. As a result of this analysis, the estimation model for measurement of snow depth using TM data was developed. The distribution map of snow dapth in dam basin (analysis area) was made by applying the estimation model developed by this study to Landsat TM data.
The conventional method for classification of remote sensing image is based on Bayes' theorem. The applied condition is to be “each pixel must belong to either of the classes”. In other words, neither the pixel belonging to more than one class ('mixed' pixel) nor the pixel belonging to none of the classes ('unknown' pixel) can be allowed. However, the 'mixed' pixel is necessarily existent in the case of satellite imagery. The existence of 'unknown' pixel is also inevitable as the number of class settings is restricted. This paper proposes the fuzzy classification of the remote sensing image. The classes are defined as fuzzy sets. With this the 'mixed' and 'unknown' pixels can be theoretically considered by the fuzzy set operations. An important problem is how to give a membership function for each class in multi-spectral space. In this paper the membership function is calibrated on the least squares criteria from the training data by the back propagation algorithm of neural network model. The performance of the proposed fuzzy classification method is evaluated in comparison with the conventional supervised classification method. This paper also discusses a method to effectively visualize the fuzzy classification result using RGB color composite.
Three new classification methods for multi-tempotal data are proposed. They are named as a likelihood addition method, a majority method, and a Dempster's rule method. Basic strategies using these methods are to calculate likelihoods for each temporal data and to combine obtained likelihoods for final classification. These three methods use different combining algorithms. From classification experiments, following results were obtained. The method based on Dempster's rule of combination showed about 11% improvment of classification accuracies compared to a conventional method. This method needed about 16% more processing time than that of a conventional method. The other two proposed methods showed 1% to 5% increase of classification accuracies. However, processing times of these two methods were almost the same with that of a conventional method.