Adaptive interpolation of multidimensional signals for differential compression

Algorithms of interpolation of multidimensional signals for differential compression are investigated. We propose an approach for constructing adaptive interpolators based on the automatic selection of an interpolating function at each point of the signal. The selection is made with the use of attributes calculated from the point’s local neighborhood. An adaptive multidimensional interpolator is developed with this approach. It automatically selects an interpolating function at each point of the signal, providing improved contour interpolation accuracy. The choice is made by a decision rule based on a local characteristic of distinctness and direction of the contour. The proposed interpolator is implemented for a three-dimensional case. The interpolator switches between six interpolating functions: an averaging function and functions that take into account contours of five directions. An experimental study of the proposed algorithm is carried out on three-dimensional hyper spectral remote sensing data. The proposed interpolator allows increasing the efficiency of differential compression.