Non-uniform sampling issues arising in shallow angle wave profiling LIDAR

Non-uniform sampling has been identified as a fundamental property in the shallow angle remote sensing required to collect the data required to build DSWP prediction models for moving vessels. The dependence of typical DSWP models on traditional discrete spectral techniques introduces the requirement to re-map sets of N non-uniform sea profile data onto uniformly sampled equivalents. The high computational cost of the general change of basis involved in this process and its detrimental effects upon prediction time motivated an examination of approximation methods. A metric IF is discussed which measures the degree of departure from uniformity in terms of the Primary Band Limitation property which is associated with a generalisation of Nyquist´s theorem. Various levels of approximation are examined based upon the extent of the departure of Γ from unity. The major conclusion is that for each set of N remotely sensed data values which constitute a sea profile scan it is possible to assess the level of computational sophistication required to render the data set suitable for use in DSWP. This ability is central to the design of an the intelligent system architecture which is required to operate a viable DSWP unit.