Péter Steinbach, Balázs Székely

ELTE, Geophysical Department

One of the most effective way of exploring and mapping under-surface object has been proven the geophysical magnetic prospecting in the last decades. The fact that magnetic instruments have became sensible enough to detect very weak magnetic effects of archaeological objects made it possible routinely to apply this approach in the archaeological research prior to the excavations.

Magnetic surveys yield values of the magnetic vectorian field, sampled in the vicinity of the studied object. Anomaly in the undisturbed background field, appearing as variation in the spatial data distribution is coupled to under-surface object. The rough picture of the magnetic anomaly (or anomalies), like a map gives only a first order assumption about the location and extent of the archaeological site or single object to be explored without the knowledge of physical properties. Much more detailed, quantitative information can be extracted from the recorded dataset by proper modelling the object caused the measured magnetic anomaly.

The parameter estimation means determining the parameters of an arbitrary chosen model based on the fitting the calculated field values to the measured ones. The accuracy and the stability of the result strongly depends on the complexity of the used model (number of parameters to estimate), the mathematical tool of the optimising phase and, of course, on the quality and volume of the measured data. The information of the physical characteristics of the unexcavated objects, served by a proper modelling is undoubtedly worth to complete this step.

A unique, statistical, 3D optimising procedure is described in details. This algorithm not only determines the geometric and magnetic properties of a 3D vertical prism model object, but also the error of the fitted parameters and their covariance matrix is given. Numerous examples of completed model estimation prove the applicability and accuracy of this method, as well as draw its limitations.