The problem of 2D wavenumber spectrum estimation from randomly sampled data is examined. Two new estimators are proposed, one based on a spatial sample and hold procedure and one based on the projection slice theorem, commonly used in tomography. Each estimator is tested using simulated PTV data. Initial results indicate that these estimators have the potential to extend spectral estimates beyond the convential Nyquist lim-its imposed when using regularly sampled data.
展开▼
