The investigation of principles of the correlation processing in the first arrival times picking of seismic refraction data

In this thesis a new technique consisting of automatic/semi-automatic algorithms based on correlation process, which are widely used in many phases in seismic reflection data processing was developed to, calculate the first arrival times from seismic refraction data which have difficuities in picking the first arrival times. Since the success of considered automatic/semi-automatic algorithm is directly correlated with the susceptibility of defined source wave operators, correlation operators were designed in three ways; power spectrum, directly wave forms on clean channel or take into account by used source, minimum phase wavelet models. The type of operator was determined with respect to general complexity characteristic of data.The developed technique was calculated for different ground models and it was applied to the real field data recorded from different geologic environments. The obtained results were discussed with respect to cause and affect relationship in detail. It was seen that first arrival picking with the correlation technique is very useful as the noise content of data increase and picking difficulty occurs due to upset first arrival waveform along the offset. Also, effects of time differences between manual and semi-automatic picking on seismic refraction results were evaluated with seismic tomography and delay time methods solutions. In the comparisons it was seen that the tomography solutions done with the first arrival times obtained from correlation method has a better lateral resolution, while this time differences may change seismic velocity and layer thicknesses unexpectedly in delay time solutions