Multiple frequency compositing of spatially coincident GPR data sets

The compositing of spatially coincident GPR data sets acquired over a range of different antenna frequencies is a possible method for expanding the spectral bandwidth. A modelling study using the Berlage wavelet to represent the GPR source signature shows that the best results are obtained by applying an appropriate shift and weighting to the individual frequency components before compositing. It was also determined that application of only the weighting procedure significantly enhanced the composite signal. Compositing was performed on two spatially coincident multiple frequency data sets to examine its implementation with actual field data. A pre-compositing linear mute effectively suppressed the lower frequency direct wave interference with the very shallow reflections in higher frequency profiles. The best compositing results were obtained by using an optimal spectral whitening procedure to estimate the component weightings.