CCDWM

Description

CCDWM combines the results from multiple change detection techniques in a weighted manner. Areas indicating change for multiple techniques (intensity, coherence, Wishart) are much more likely to indicate real changes. For each change file, the user weights are applied to the change ranking channel and the results summed. Areas ranking high for all change metrics are highlighted. For convenience, the weighted values are ranked from 0 (lowest value) to 100 (highest value).

Parameters

ccdwm(mfile, weights, dbic, filo)

Name	Type	Caption	Length	Value range
MFILE *	str	Input change metric rasters	1 -
WEIGHTS	List[float]	Weights	0 -
DBIC	List[int]	Input raster channel	0 - 1	1 - Default: 4
FILO *	str	Output weighted change raster	1 -

* Required parameter

Parameter descriptions

MFILE

Specifies a list of co-registered change metric files. All input data must be in the same projection and pixel size.

It is expected that the input change metric rasters have been produced by CCDINTEN, CCDPHASE, or CCDWISH.

MFILE can be set using any of the following options:

A specific file name pattern or all of the files in a folder; for example, "c:\data\*.pix" or "c:\data"
An existing text file that contains references to multiple files; for example, "c:\list.txt"

If the text file option is used, the following general rules apply:

Specify a .txt extension for the file containing multiple files.
Specify string parameters within quotation marks.
A quotation mark preceded by a backslash is used as part of the string. For example, 'ab\"c' is used as: ab"c.
Use an exclamation mark (!) before writing a comment in the text file.

WEIGHTS

Optionally specifies a list of zero- or positive-valued weights to be applied to ranked change metrics, as defined by the input image. The first weight will be applied to the first image, the second weight to the second image, and so on. You must specify the same number of weights as there are image files to compare; otherwise, CCDWM will error and abort. By default the images are weighted equally, with a value of 1 for each image.

DBIC

Specifies the input channels to be compared. The specified channel must exist for all input images. All input channels must be the same bit depth, and must be all real-valued (that is, non-complex). The default is 4 because that will correspond to the percentile change created by PCI's other Coherent Change Detection algorithms.

FILO

Specifies the name of the output file to be created. This file will contain the weighted metric in channel 1 and the weighted metric cumulative percentile values in channel 2.

The specified file must not already exist.

Details

CCDWM rates the results of different change detection techniques to produce a unique map of overall change. Depending on the change and features of interest, different change detection techniques might enhance different changes. For instance, when the backscatter in two images is very similar but the phase between the co-polarization has changed, the Wishart test statistic is much more sensitive to the differences than test statistics based only on the intensity. However, depending on the gradient of change, you might want to use results from both CCDINTEN and CCDWISH but weight them differently to determine the overall changes.

Example

In this example, the change detection results from CCDINTEN, CCDPHASE, and CCDWISH are combined. The user has already determined the weights, based on visual inspection of the results for each technique on their features of interest.

from pci.ccdwm import ccdwm

mfile	=	"ccdfiles.txt"	# text file listing file names
#  to process
weights	=	[2, 1, 1]	# results from first file in MFILE will be weighted twice
dbic	=	[3]	# use channel 3 to compute weighted change
filo	=	"weightedmetrics.pix"	# output file name

ccdwm(mfile, weights, dbic, filo)

Environments	PYTHON :: EASI :: MODELER
Quick links	Description :: Parameters :: Parameter descriptions :: Details :: Example :: Related