PSEABA

Description

PSEABA creates an image of four parameters for a fully polarimetric SAR (POLSAR) data set: entropy, alpha angle, beta angle, and anisotropy. These parameters characterize the properties of partially coherent scattering by computing the proportion and type of the scattering mechanism for all features in your image. This information can be used further for image classification. The eigenvalues and eigenvectors used to compute these parameters can also be output, if required.

Parameters

pseaba(fili, filo, geteigen)

Name	Type	Caption	Length	Value range
FILI*	str	Input polarimetric SAR image	1 -
FILO*	str	Output Cloude-Pottier decomposed raster	1 -
GETEIGEN	List[int]	Retrieve eigenvalues and eigenvectors	0 - 1	Default: 0

* Required parameter

Parameter descriptions

FILI

The name of the input polarimetric SAR data set, which must be either non-symmetrized or symmetrized fully polarimetric (quad-polarization) complex data. The input data set must be in covariance, coherence, or Kennaugh matrix format.

The input data set must be a data set that is already imported into the PCIDSK (.pix) format by SARINGEST or it can be the key file name of any GDB-supported POLSAR data set in its distribution format. For more information and a complete list of supported polarimetric SAR sensors and data products, see the SARINGEST Help.

The input data set should have an equivalent number of looks (ENL) of at least 25, which could be achieved by applying a boxcar filter when the input data set is single-look complex. The decomposition for single-look complex data generates only one eigenvalue; the entropy is zero and the anisotropy is undefined.

FILO

The name of the output file to which to write the discriminators.The output file has the same dimensions as the input file, but the number of channels depends on the output matrix type and whether or not the eigenvalues and eigenvectors are required. All output channels contain floating-point pixel values.

The specified file must not already exist.

GETEIGEN

Specifies whether to extract the eigenvalues and eigenvectors. When the value of GETEIGEN is 0, only four channels containing entropy, anisotropy, alpha, and beta angles (in degrees) are required. When the value of GETEIGEN is 1, 12 additional channels are created in the output file, which results in a larger file size. These additional channels contain the three eigenvalues, stored in order of magnitude, followed by the three polarization elements of their corresponding eigenvectors, each of which represents an orthogonal scattering mechanism.

Details

PSEABA computes four parameters that describe the properties of partially coherent scattering, through a Cloude-Pottier decomposition, for every pixel of a fully polarimetric SAR (POLSAR) data set.

The input data set must contain either non-symmetrized or symmetrized, fully polarimetric (quad-polarization) data in one of the following matrix formats: covariance (c4r6c or C3r3c), coherency (t4r6c or T3r3c), or Kennaugh (k16r or K9r). If required, the input data set is converted by this algorithm to the symmetrized coherency matrix, T3r3c, to perform the calculations. When the conversion is not supported, the function stops processing and produces and an error message.

The core algorithm used in PSEABA is based on the article in References. It decomposes the symmetrized coherency matrix into eigenvalues (used to determine the entropy and anisotropy values), and eigenvectors (used to compute the alpha and beta angles). Each eigenvector represents an orthogonal scattering mechanism and is specified as a Pauli component. The diagonal matrix has three real, non-negative elements that represent the eigenvalues of the three scattering mechanisms. The eigenvalues are sorted in decreasing order along the diagonal.

The entropy measures the amount of mixing between the three scattering mechanisms, where values close to 0 indicate a single scattering mechanism, and values close to 1 indicate an equal mixture of three scattering mechanisms (equal eigenvalues).

The anisotropy characterizes the amount of mixing between the second and third scattering mechanism. The anisotropy A=0 indicates that the two mechanisms are mixed in equal proportions and that their eigenvalues are equal. An anisotropy value close to 1 indicates that the second mechanism dominates over the third mechanism and that the second eigenvalue is much larger than the third eigenvalue. The anisotropy is set to 0 when both eigenvalues are 0.

Each eigenvector can be characterized by the alpha and beta angles. The alpha angle (between 0 and 90 degrees) characterizes the scattering mechanism. For example, alpha=0 degrees indicates a trihedral scatterer or a smooth surface; alpha=45 degrees indicates a dipole scatterer (often assigned to volume structures); alpha=90 degrees indicates a dihedral scatterer (often related to double-bounce).

The beta angle (between 0 and 90 degrees) is twice the preferred orientation angle of the scatterer, and therefore characterizes the dominant polarization.

The overall values of the alpha and beta angles at every pixel are derived as weighted averages of the values for the three eigenvectors. The weight of each value is computed from its eigenvalue divided by the sum of all three eigenvalues.

The processing in PSEABA is similar to that in PSCLOPOT. The two functions differ in that PSCLOPOT performs an unsupervised classification using the entropy, alpha angle, and anisotropy space and creates a classification map, while the PSEABA algorithm provides access to the individual parameter values at every pixel.

Example

In the following example, compute the four parameters for an SLC RADARSAT-2 image. The SLC image is first averaged by the PSBOXCAR function using a 5 x 5 window. PSBOXCAR also converts the image from the original non-symmetrized scattering matrix (s4c) format to the non-symmetrized covariance matrix format (c4r6c). The averaged image is then used as an input to PSEABA. The eigenvalue and eigenvector channels are also created.

from pci.pseaba import *
from pci.psboxcar import *

fili	="rast2.pix"
filo	="r2_enl25.pix"
flsz	=[5]
psboxcar( fili, filo, flsz )

fili	="r2_enl25.pix"
filo	="r2_pseaba.pix"
geteigen=1
pseaba( fili, filo, geteigen )

Acknowledgements

PCI Geomatics gratefully acknowledges the financial support provided by the Canadian Space Agency through the Earth Observation Application Development Program (EOADP), contract number 9F028-034946.

References

Cloude, S.R., and E. Pottier. "An entropy based classification scheme for land applications of polarimetric SAR", IEEE Trans. Geosci. Remote Sensing, 35, no. 1 (1997): 68-78.