Stochastic Signal Processing > Spatial Data Quality Assessment
  • Step 1: spatial data analysis.
    S Q A is performed on the amplitude pre stack gather data set shown in Figure a. The experimental variogram is computed in the offset direction on the entire gather and displayed in figure b in order to highlight the global spatial noise and signal features of the gather.
    S Q A probability model only relies on a local stationarity assumption meaning that local computations are performed inside a local neighborhood defined around each sample location. Figure c displays the local energy of the gather.

  • Step 2: variogram interpretation.
    The seismic gather contains both signal and noise as highlighted in figure a. Noise signatures appear on the experimental variogram as a nugget effect and a short 7 offset range structure as shown in Figure b.
    • Nugget effect : spatial variability between 0 and 1 offset indicates non-correlated noise from trace to trace;
    • Spatial variability for short 2 to 6 offset range indicates correlated noise from trace to trace;
    • Spatial variability for longer offset range indicates signal
    The interpretation of the noise on the variogram allows for computing the energy of noise and signal inside local neighborhoods centered on each gather data, as shown in Figure c

  • Step 3: computation of the Spatial Quality Index.
    Spatial Quality Index or S Q I is computed as the ratio between local variance of the signal and the local variance of the signal plus noise as shown in figure b.