The principal components decomposition creates a set of decorrelated data bands whose energies are ranged in amplitude. This process requires knowledge of the covariance matrix (automatically estimated, in this case). The resultant bands can also be recomposed into the original values space, using the generated matrix.1
Inverse PCA: If checked, an inverse PCA transformation will be performed. An input PCA matrix is required.
Output raster data type: Select the output raster pixel data type (Double precision is required for the direct transformation).
PCA Matrix file: File name for the created/required PCA Matrix (required for inverse PCA transformation, optional for the direct transformation).
Bands: The raster bands to be processed.
Press NEXT to jump to the next wizard page, where an output raster file name must be selected.
1 Digital Image Processing: PIKS Inside, Third Edition. William K. Prat. ISBN: 0-471-37407-5 (Hardback)