INSCOREG

Description

INSCOREG automatically coregisters and resamples a dependent file to a reference file. Control points are acquired automatically, outliers removed, and pixels resampled to match the reference file. The metadata of the resampled output file is updated to reflect the modified geolocation information. Non-overlapping areas are flagged as NoData.

Parameters

inscoreg(filref, dbic_ref, fili, dbic_dep, dbic, modeltyp, numpts, minscore, searchr, filo)

Name	Type	Caption	Length	Value range
FILREF *	str	Database reference file name	1 -
DBIC_REF	List[int]	Database reference input channel	0 - 1
FILI *	str	Database dependent file name	1 -
DBIC_DEP	List[int]	Database dependent input channel	0 - 1
DBIC	List[int]	Database dependent input channel list	0 -
MODELTYP	str	The type of model to fit the GCPs	0 -
NUMPTS	List[int]	Number of candidate GCPs	0 - 1	1 - Default: 500
MINSCORE	List[float]	Minimum acceptance score	0 - 1	0.72 - 1 Default: 0.75
SEARCHR	List[int]	Search radius	0 - 1
FILO *	str	Database output file name	1 -

* Required parameter

Parameter descriptions

FILREF

The name of the file to use as the reference data set. The file can be in a supported vendor format or a PCIDSK (.pix) file created by SARINGEST. With interferometric processing, the file must contain single-look complex (SLC) valued data.

DBIC_REF

The channel number of the reference file to use for image-to-image matching. You can specify only one channel.

The default value is 1.

FILI

The name of the file to use as the dependent data set. The file can be in a supported vendor format or a PCIDSK (.pix) file created by SARINGEST. With interferometric processing, the input file must be SLC data and the acquisition parameters (wavelength, polarization, and so on) and viewing geometry (location, look direction, pass direction) of the dependent file must match the reference file. The interferometric parameters can be identified by the INSINFO algorithm.

DBIC_DEP

The single channel of the dependent file to match with the corresponding channel of the reference file.

DBIC

The channels to resample and coregister. The result is written sequentially to the output file in the order specified.

The default value is 1.

MODELTYP

The type of model to fit to the collected GCPs, the model is then used to interpolate pixel shifts between GCPs. The valid options are:

TPS: Apply a thin plate spline
POLY#: Apply polynomial model, you can add a number after to specify the order

The default is a 2nd order polynomial model.

NUMPTS

The number of grid points to distribute over the reference image. The corresponding location, based on geocoding information, is used as the center of the search radius in the dependent file.

The value must be greater than or equal to 1. The default value is 500.

MINSCORE

The minimum acceptable correlation score. Match points with a lower correlation score will be rejected. The value must be in the range of 0 (no correlation) to 1 (perfect correlation). The default value is 0.75, and is the recommended minimum value.

SEARCHR

The search radius, in pixels, to examine to find a match between the reference and dependent images. The value must be greater than or equal to 1. The default value is 50.

FILO

The file name to use for the resampled dependent data set. The resampled dependent pixels will align one to one with the reference pixels and contain the number of channels specified by the dependent input channel list. Non-overlapping areas are flagged as NoData. The file name you specify must not exist already.

Return Value

Returns:

Type: PCI_INT

Details

The reference channel and the dependent channel you specify are used to generate image match points. The number of candidate points you specify are distributed evenly (in a grid) over the reference image. With each candidate reference point, a corresponding location in the dependent file is determined. The area surrounding the dependent location is scanned to find the image match with the highest score.

Dependent points with a correlation below the threshold you specify are removed. The remaining match points are used to derive the mapping of the low-order-polynomial, which is applied to all channels specified in the input channel list. The resampled complex-valued output contains the number of channels, and the order of occurrence, corresponding to the input channel list.

Examples

Coregister the first channel (default) of the dependent file to the first channel (default) of the reference file. The goal is to use 500 points with a search radius of 10 pixels and a minimum correlation of 0.80. The resampled image will be the same size (number of lines and pixels) as the reference file and contain one channel. Non-overlapping areas will be identified as NoData. The GCP and image-matching parameters are written to the report.

from pci.inscoreg import inscoreg

filref   =  "Reference_File.pix"
dbic_ref =  []
fili     =  "Dependent_File.pix"
dbic_dep =  []
dbic     =  []
modeltyp =  ""
numpts   =  [500]
minscore =  [0.8]
searchr  =  [10]
filo     =  "Resampled_File.pix"

inscoreg(filref, dbic_ref, fili, dbic_dep, dbic, modeltyp, numpts, minscore, searchr, filo)

Coregister the first channel of the dependent file (default) to the first channel of the reference file by using the default settings. The resampled image will be the same size (number of lines and pixels) as the reference file and contain four channels in the specified order. Nonoverlapping areas will be identified as NoData. The GCP and image-matching parameters are written to the report.

from pci.inscoreg import inscoreg

filref   =  "Multi_Pol_Reference_File.pix"
dbic_ref =  []
fili     =  "Multi_Pol_Dependent_File.pix"
dbic_dep =  []
dbic     =  [1,2,3,4]
modeltyp =  "TPS"
numpts   =  []
minscore =  []
searchr  =  []
filo     =  "Multi_Pol_Resampled_File.pix"

inscoreg(filref, dbic_ref, fili, dbic_dep, dbic, modeltyp, numpts, minscore, searchr, filo)

Coregister the fourth channel of the dependent file to the fourth channel of the reference file by using the default settings. The resampled image will be the same size (lines and pixels) as the reference file and contain two channels (the first and fourth) of resampled data in the specified order.

from pci.inscoreg import inscoreg

filref   =  "Multi_Pol_Reference_File.pix"
dbic_ref =  [4]
fili     =  "Multi_Pol_Dependent_File.pix"
dbic_dep =  [4]
dbic     =  [1,4]
modeltyp =  "POLY2"
numpts   =  []
minscore =  []
searchr  =  []
filo     =  "Co_Pol_Resampled_File.pix"

inscoreg(filref, dbic_ref, fili, dbic_dep, dbic, modeltyp, numpts, minscore, searchr, filo)

Algorithm

Points with an absolute error of less than 0.1 of a pixel in both the x and y directions are stored as ground control points (GCP). The transformation determines the exact image location of the output pixel. INSCOREG applies a transformation of the second-order polynomial to the dependent data. If less than 10 GCPs are found, the order of the polynomial transformation is reduced to one; that is, an affine transformation is applied.

The value of the output pixel is determined by using a two-dimensional sinc function; that is, sin (x)/x), as the resampling kernel over a 16-by-16 neighborhood.

Acknowledgements

The PCI Interferometric SAR (InSAR) project was funded in part by the Canadian Space Agency under the Earth Observation Application Development Program (EOADP) contract (9F043-130644/006/MTB), Application Development for Environmental Monitoring and Remediation.

Environments	PYTHON :: EASI :: MODELER
Quick links	Description :: Parameters :: Parameter descriptions :: Return Value :: Details :: Examples :: Algorithm :: Acknowledgements :: Related