decorr stretch
Description
Perform a decorrelation stretch on an RGB image by applying a whitening transform with stretch.
This works by
- calculating the covariance matrix
- calculating a transform to a space in which the diagonals of the covariance matrix are 1 and the other elements are minimised (i.e. the bands are decorrelated) - so this is a form of PCA
- applying that transform
- applying a stretch factor to increase those diagonal variances - the principal components
- apply the inverse transform
- restore original image scaling
- fix up the image mean
- normalise each band after clipping a given percentile of outliers
Ignores DQ and uncertainty
Connections
Inputs
| Index | Name | Type | Desc |
|---|---|---|---|
| 0 | rgb | img | (none) |
Outputs
| Index | Name | Type | Desc |
|---|---|---|---|
| 0 | rgb | img | (none) |
| 1 | eigs | number | (none) |
| 2 | sds | number | (none) |
Parameters
| decorr stretch | stretch: float (default 1.0) | stretch factor |
| clip: float (default 5.0) | percentile outliers to clip in postprocessing |
Automatically generated by generate_autodocs.py
Date: 2026-07-03T14:16:39.921215