The Research On The Separating Methods For Potential Field In Space Domain

Potential field separation is an important technique in the processing of potential field data, including the same and different dimensions methods, the same and different positions methods. The potential separation methods of the same dimensions and the same positions can be divided into two categories:space domain method and frequency domain method. For the potential separation methods of the different dimensions and the different positions, all derivative method and continuation method are two important methods. In this paper we focus on the trend analysis method and the interpolation-cut method in space domain.Although a tremendous amount of research works have been done on the potential field separation technique, there are still some problems that need to be solved. In this paper, the formula of the three kinds of fixed order times trend analysis methods (general, sliding, sliding iteration) and change order times trend analysis method are introduced, and the filtering properties of the fixed order times sliding trend analysis method and the multiple sliding iteration trend analysis method are discussed. To shed some new light on the pathological over-determined problem in the fixed trend analysis method, all the coordinates quantities are parallel transported and normalized such that the stability of the solution is improved. Then we work out the formula of the interpolation-cut method. In order to improve the accuracy of potential field separation, we adopt the minimum curvature extending edge method. Finally, under the Fortran 90 standard, the codes of the corresponding programs have been finished and the result of the synthetic data and field data are analyzed using the trend analysis method and the interpolation-cut method respectively.We obtain several understanding in this work. First, the minimum curvature extending edge method is superior to cosine function extension edge method. Second, the trend analysis method is low pass filtering. Third, the effect of change order multiple iterative sliding trend analysis is better among all the methods that considered. Finally, our result of field data is consistent with the distribution of exposed rock, which provide favorable basis of studying the distribution of rocks.