A Smoothed Filtering and Correction Factor Proposal for ERT2D and TDIPT2D Data Collected in Difficult Terrain to Improve Inversion Models

When collecting data from ERT2D (electrical resistivity tomography) and TDIPT2D (time domain IP tomography), different phenomena occur, whether natural or anthropogenic noise, which contaminate the data, and it can make their processing, analysis and interpretation difficult. Different techniques have been developed to eliminate or reduce these effects on the data, such as noise filtering or the development of new techniques for better data collection in the field. In the present work an iterative weighted least squares filter is employed after voltage normalization by current and a geometrical factor correction applied on strong topographic terrains. The selection of the filter basis function is recommended to be able to represent the natural behavior of the function to be filtered. Stationary or variable voltages in electrical prospecting decay with the inverse of the distance, which can be represented by an expansion in Legendre polynomials. On the other hand, uneven spacing of the electrodes leads to not using the correct geometric factor which results in an error in the calculation of the electrical anomaly. The efficiency of the proposed technique is analyzed and tested with field examples using different filters and applying the proposed correction factor or without using it. The results indicate a low RMS and L2-Norm error, whereby a better definition of the inverted resistivity image is obtained. For the TDIP case a better correspondence between the inverted images of resistivity and chargeability can be reached.