Reconstruction of spatial data using isometric mapping and multiple-point statistics

摘要: Only partial spatial information in studied fields is a ubiquitous problem the reconstruction of data and major cause uncertainty for reconstructed results. This not likely to change since there will always be some unsampled volumes simulated regions where no direct available. Multiple-point statistics (MPS) can powerful tool address this issue because it extract features training images copy them using sparse conditional or even without any data. Because from are linear, previous MPS methods linear dimensionality reduction suitable deal with nonlinear situation. A new method isometric mapping (ISOMAP) that achieve proposed reconstruct The patterns image classified clustering after reduced. simulation performed by comparing current event average all class finding out one most similar event. experiments show structural characteristics reconstructions those images.