==================================================== This is the XML Schema defining the metadata of Oracle GeoRaster object type It consists of two parts: data type definitions and its element content Part 1: Data Types Part 1.1: Data Types for Object Info Part 1.2: Data Types for Raster Info Part 1.3: Data Types for Spatial-Temporal-Band Reference Systems Part 1.3.1: Data Types for Raster Spatial Reference Systems Part 1.3.2: Data Types for Raster Temporal Reference Systems Part 1.3.3: Data Types for Raster Band Reference Systems Part 1.4: Data Types for Layer Metadata Part 2: GeoRaster Metadata Elements or Content Structure ============================================================= =============================================== Part 1: Data Types ================================================ ============================== Part 1.1: Data Types for Object Info ============================== ============================== Part 1.2: Data Types for Raster Info ============================== The "Band" dimension can be treated as any other semantic dimension or any "Layer" if not remote sensing imagery or photographs =========================================================== Part 1.3.1: Data Types for GeoRaster Spatial Reference System Spatial extent (footprint) is recorded as an attribute of georaster object. Its type is SDO_GEOMETRY. So it is not included in the metadata The cell space coordinates are named as (row, column, vertical) The model space coordinates are named as (x, y, z) Spatial unit information is stored in the WKT of the specified SRID ============================================================ The following "S" means "Spectral" for remote sensing imagery Any of X, Y and Z can be horizontal or vertical or any other spatial direction (depending on user interpretation in "modelDimensionDescription") row = pPolynomial(x, y, z) / qPolynomial(x, y, z) column = rPolynomial(x, y, z) / sPolynomial(x, y, z) The following types and definitions are for GCP support. It stores GCP collection for the GeoRaster object. It also optionally specify the Functional Fitting method for generating FFM using the GCP collection. cellDimension can 2 or 3 (only 2 is supported in current release). modelDimension can be 2, 3, -2, -3. cellCoordinate must be (row, column) in current release modelCoordinate must be (X, Y) or (X, Y, Z) in current release ================================================= Part 1.3.2: Data Types for GeoRaster Temporal Reference System The TRS will be modeled by formulas in the future ================================================= ================================================== Part 1.3.3: Data Types for GeoRaster Band Reference System For multispectral remote sensing images, each band is optionally described in the layerDescriptionType The BRS is modeled by formulas for hyperspectral imagery (based on number of spectral segments, min and max wavelength and number of bands for each segment) Detailed radiometric info will be added in the future ===================================================== ==================================================== Part 1.4: Data Types for Layer Metadata For each sub-layer the layerNumber is a positive integer, i.e., layers are logically numbered from 1 to n if the size of the specified layerDimension is n. The layerDimensionOrdinate of each sublayer must be in the range of the dimension and must be in the order of band ordinates. For objectLayer, the layerNumber should be 0 but its layerDimensionOrdinate is not used. ======================================================= value = (a0 + a1 * cellValue) / (b0 + b1 * cellValue) LINEAR bin function: binNumber = numbins * (cellValue - min) / (max - min) + firstBinNumber if (binNumber less than 0) binNumber = firstBinNumber if (binNumber greater than or equal to numbins) binNumber = numbins + firstBinNumber - 1 LOGARITHM bin function: binNumber = numbins * (ln (1.0 + ((cellValue - min)/(max - min)))/ ln (2.0)) + firstBinNumber if (binNumber less than 0) binNumber = firstBinNumber if (binNumber greater than or equal to numbins) binNumber = numbins + firstBinNumber - 1 EXPLICIT bin function means explicit (or direct) value (or value range) for each bin and it will be stored in a table The MAX and MIN in statistic dataset will be used if they are not provided here binTableName is used by EXPLICIT type only ======================================================= Part 2: Metadata Elements / Content Structure of Oracle GeoRaster Object ==========================================================