Dina Research SchoolDina Research School
We consider problems where in the real world, it is impossible to get exhaustive values of data at every desired point because of practical constraints. Thus, interpolation is important and fundamental to graphing, analysing and understanding of multidimensional data. To create a digital terrain model, for example, it is for instance required to interpolate elevation measurements to a 2D grid. In soil science, spatial interpolation is often used to construct maps of soil properties like nutrient contents, soil humidity, pH etc.
The workshop will consider both "deterministic" approaches to interpolation like multidimensional splines, thin plate splines and mesh smoothing, as well as kriging which is based on statistical criteria of optimal prediction for random surfaces. In applications of kriging, the so-called variogram is used to quantify spatial variation and the kriging interpolator minimizes the expected squared error of the prediction.
The theoretical sessions formulates the problem of spatial interpolation from scattered data as a method for prediction and representation of multi-variate fields. The role and specific issues of interpolation for various applications are discussed and properties of interpolation methods are illustrated by examples of Texture Based Volume Visualization. Future directions focus on robust data analysis with automatic choice of spatially variable interpolation parameters, and model or process-based interpolation.
Case studies presentations discusses how the methodology is used in real applications in soil science and agriculture.
Author: phd@dina.kvl.dk. Updated:
11 april 2006