Stefano Salvini and Stefan J. Wijnholds, "High-compression Baseline Dependent Averaging," URSI General Assembly and Scientific Symposium (URSI GASS), Montreal (Canada), 19-26 August 2017.

abstract:
Baseline dependent averaging (BDA) can be used to reduce the volume of visibility data significantly. Most current BDA schemes perform (weighted) averaging over a certain time interval. This quickly causes decorrelation due to time averaging. We propose to reduce this decorrelation by representing the visibilities by polynomial coefficients. The high compression made feasible by this approach may cause fast-changing calibration parameters to become undersampled. We propose the Compress-Expand-Compress (CEC) method to mitigate this. All compression and expansion methods proposed herein are very simple and cause negligible computation overhead. We demonstrate the effectiveness of our scheme in a simulation emulating a highdynamic range imaging problem.

back to publication list