Statistical Characterization of Rough Terrain

摘要

A technique is presented to statistically characterize rough terrain surfaces. The approach is described in terms of a specific example but in principle is quite general and can be adapted to fit any number of situations. The starting point is the use of observed surface height data to generate statistical parameters for probability density functions (PDF) that potentially characterize the data. This involves the use of parameter estimation techniques. The estimated parameters are then used in an hypothesis test to determine the best PDF for the given data. The accuracy of the analysis depends on the available data, the estimators employed, and the determination of appropriate PDF's. The example consists of a large number of terrain regions in an area of eastern Massachusetts. Experimental data in the form of electromagnetic scattering from the surface are available for this site. Because of the number of cases involved, the complexity of the multivariate height distributions, and the type of measurements available, only a single observation of the multivariate data is used in the present analysis. Techniques to improve the parameter estimation are being pursued. The results of the statistical analysis in terms of mean height, variance, correlation, PDF, and a geologic feature characterizing each subregion are presented. These will be used in an electromagnetic scattering formulation to allow comparison with the experimental data. Once agreement is reached, other areas of interest can then be analyzed in a similar fashion. (Author)