Fissile material is relatively dense, and consequently will have an increased gravitational effect compared to most materials. We want to explore the possibility that a Tikhonov regularization approach can be used to identify such high density material. Since it is a largely parameter based method there are many different aspects of the problem to optimize, some of which may be dependent on others.
In this talk I will be going over the method I followed, which involves implementing a logarithmic barrier term in conjunction with the Newton method and conjugate gradient method. The Huber norm is also used to decrease the spreading in the recovered model, and other weighting functions are explored to improve on the penetration problems usually associated with gravity gradiometry. Accuracy is the main factor when analyzing the reconstructions. However the hope is that this method might be used to scan cargo containers, and so efficiency and ease of use must also be considered.