GPR Data Visualization Enhancement by Deconvolution and Some Image Attributes

Abstract

Deconvolution is a filtering process which removes a wavelet effect from the recorded trace by reversing the process of convolution. Predictive deconvolution is an attempt to attenuate multiples which involve the surface or near-surface reflectors. The GPR data visualization is based on selected amplitude-colour scale, opaque, view angle and selected data volume range which are called 2D/3D GPR visualization attributes to indicate discontinuities, infrastructures and buried remains. The aim of this study is to indicate a predictive deconvolution processing for the ground penetrating radar (GPR) data to remove vertical noise including source wavelet multiples and therefore reflected wave multiples. The second aim is to show how GPR data imaging attributes visualize buried remains in very complex data. We used a special data gathered on a stone paved archaeological site with large stair steps in Sapinuwa, Agilonu- Tasdosem (stonepavement), Ortakoy-Corum city, Turkey. We obtained depth slices and 3D sub-volumes of the GPR data using visualization attributes. The results of the 2D profile sections and interactive 3D time slices and their sub-volumes indicated that the deconvolved data was cleaned from the source wavelet multiples and revealed the buried important archaeological structures hidden in the the staircase stone structure which had approximately 3.0 m high.