Creating full 3-D angle gathers through cross-correlation gathers
is currently computationally impractical given the volume size that
needs to be read/written for every shot.
Compressive sensing offers a potential solution to this problem by collecting
a subset of the correlation gathers and then forming the entire volume after
all shot contributions have been stacked.
The StOMP algorithm appears to be an effective method to obtain a sparse
basis function necessary for a successful compressive sensing effort.
Phase encoding to encode multiple correlations into every
data point appears to offer an improved result.
Further work on full 3-D shift gathers is needed to prove the
feasibility of the method.