Effective automatic picking of traveltime data with high precision

Traveltime data has to be carefully collected to ensure sufficient quality, because absolute traveltimes are needed to compare tomographic reconstructions of velocity and Q distributions of different objects. With respect to the huge amount of data an effective automatic algorithm is required to pick these times. To test different methods, a simulation of the propagation of radar waves close to reality was applied. Noise conditions and wavelet deformation due to transmission in absorbent media were taken into special considerations. It is shown that common algorithms, like simple threshold methods or correlation, are unsuitable for detecting traveltime differences. For this reason, a combination of different strategies and algorithms was developed. This procedure yields more precise results than all single algorithms separately.