Место издания:State Research Center of the Russian Federation "Concern CSRI Elektropribor", Saint Petersburg
Первая страница:88
Последняя страница:91
Номер статьи:20
Аннотация:The paper addresses the problem of local gravity modeling from terrestrial gravimetry data based on parameterizing the disturbing potential by point masses on a sphere. The task involves determining the magnitudes and depths of the point masses by solving an unconstrained nonlinear minimization problem with L₂-regularization. The solution is obtained via an iterative quasi-Newton algorithm. We also consider an alternative formulation of the problem based on a two-stage minimization of the cost function (first determining the magnitudes of point masses and then estimating their depth). The solution is provided by the least squares method and optimization of a convex single-variable function. The paper presents the results of testing and comparing both algorithms using real airborne and ground gravimetry data.
