COMPRESSION OF GEOPHYSICAL DATA

The idea of automated processing of imagery has been the desire of technology since devices to collect volumes of images have been in operation. In this proposal we discuss the relative merits of automated image processing in the spatial and transform domain. We suggest that with the development of wavelets and vector quantization (VQ) used in compression, the possibility exists to use the values in the transform space to perform pattern recognition. In addition, we have combined the wavelet transform with the affine transform of fractal compression to provide a means to represent features in a rotation and magnification independent format. Once can use these transform coefficients or identifiers (in the case of VQ and fractals) to represent the components of the image. Using these values, it is possible to apply finite inductive sequences (FI) to strings of these identifiers for the purpose of pattern recognition. The FI technique is briefly explained and its application to this process is described. The proposed data structure for storage and communication is the coefficients from the transform space, whatever they turn out to be.