CONCLUSION

This trace interpolation scheme works well on evenly sampled data sets which have a substantial low-frequency component. The nonstationarity of seismic data makes it necessary to perform the interlacing on small windows of data which are then patched back together. The algorithm has difficulty when there are many crossing events because this tends to alias the shaping filter. This problem is partially overcome by performing the interpolation in overlapping patches of data. For the cross-well data set, the presence of numerous events causes the shaping filter to be aliased. As a result the final interpolation has artifacts in it.

It may be possible to smooth the shaping filter to eliminate the artifacts but the method will eventually fail when there are too many crossing events. In general the method is better suited to relatively sparse and slowly varying data sets.