Full text loading...
The enhanced horizontal derivative (EHD) is a well known filter whose maxima provide estimated locations of the sources’ boundaries. This filter is defined as the total horizontal gradient (THG) of a weighted sum of vertical derivatives (VDR) of increasing order. A modified implementation of EHD, which we refer to as mEHD, is obtained when we switch the order of THG and VDR; that is, when we consider the weighted sum of total horizontal gradients of vertical derivatives. It turns out that mEHD is more stable than the original one, as demonstrated with examples with noise-contaminated synthetic anomalies. We also address the choice of the weighting coefficients of the vertical derivatives.