COMPUTER PROCESSING OF MULTI-DIMENSIONAL DATA

摘要

The present invention relates to a solution for processing source data into target data. The source data is here divided into a number data blocks (B1<SUB>r0</SUB>, B2<SUB>r1</SUB>, B3<SUB>r2</SUB>) B4<SUB>r3</SUB>) each block repre- senting an information space having at least two dimensions. The blocks (B1<SUB>r0</SUB>, B2<SUB>r1</SUB> , B3<SUB>r2</SUB>, B4<SUB>r3</SUB>) are delimited by a respective set of block edges (b<SUB>1</SUB> , b<SUB>2</SUB> b<SUB>3</SUB>, b<SUB>4</SUB>). The blocks (B1<SUB>r0</SUB>, B2<SUB>r1</SUB> , B3<SUB>r2</SUB>, B4<SUB>r3</SUB>) have at least two selectable resolution levels (r<SUB>0</SUB>, r<SUB>1</SUB> , r<SUB>2</SUB>, r<SUB>3</SUB>), where the resolution levels may be selected individually based on a level-of-detail scheme, such that different blocks represent information at different resolution levels. Moreover, a block sample boundary (ß<SUB>SB1</SUB> , ß<SUB>sB2</SUB>, ß<SUB>sB3</SUB>. ß<SUB>sB4</SUB>) around each block (B1<SUB>r0</SUB>, B2<SUB>r1</SUB> , B3<SUB>r2</SUB>, B4<SUB>r3</SUB>) is defined by a set of surfaces spanned by at least one sample which in each respective dimension of the information space is positioned a longest distance from a geometric center point of this block. A rim distance (d<SUB>4</SUB>) between the block sample boundary (ß<SUB>SB</SUB>) and the block edge (b<SUB>4</SUB>) is relatively short for a comparatively high resolution level (r3), and the rim distance (d<SUB>1</SUB>) is relatively long for a comparatively low re- solution level (r0). In connection with the production of the target data (D<SUB>?</SUB>) at least one interpolation parameter (?<SUB>SI</SUB>, e(P<SUB>S?</SUB>)) is determined for at least one interpolated sample between a first block (B2) neighboring a second block (B3) at least based on a first rim distance (d<SUB>2</SUB>) of the first block (B2) and a second rim distance (d<SUB>3</SUB>) of the second block (B3).