Gamma ISP: Reference Manual

cc_wave

ANSI-C program: cc_wave.c

NAME
cc_wave - coherence estimation from normalized interferogram and co-registered intensity images.

SYNOPSIS
cc_wave <interf> <pwr1> <pwr2> <corr> <width> [bx] [by] [wgt_flag] [xmax] [ymin] [ymax]

<interf>	complex interferogram
<pwr1>	(input) intensity image of the first scene (or -)
<pwr2>	(input) intensity image of the second scene (or -)
<corr >	output correlation filename
<width>	number of samples/row
[bx]	coherence window size (columns) (default = 5.0)
[by]	coherence window size (rows) (default = 5.0)
[wgt_flag]	weighting function: 0: constant (default) 1: triangular 2: gaussian 3: none (phase only)
[xmin]	starting range pixel offset (default = 0)
[xmax]	last range pixel offset (default = width-1)
[ymin]	starting azimuth row offset, relative to start (default = 0)
<ymax>	last azimuth row offset, relative to start of SLC-1 (default = nlines-1)

NOTE: omitting pwr1 and pwr2 or setting wflg = 3 selects a coherence estimate algorithm that only
uses the complex interferogram values. In the case of wflg = 3, only the interferogram phase is used.

EXAMPLE
cc_wave 1352_1610.flt 1352_1610.pwr1 1352_1610.pwr2 1352_1610.cc 2500 5.0 5.0 1 cc_wave 1352_1610.flt - - 1352_1610.cc 2500 4.5 4.5 2

DESCRIPTION
cc_wave estimates the interferometric coherence from the normalized interferogram and the co-registered intensity images. Alternatively the complex-valued interferogram can be used alone.

The second estimator is selcted when only a normalized complex interferogram is provided. To ommit the intensity images pwr1 and pwr2, place a - on the command line as a place holder. This second type of estimator must be used for interferograms filtered using non-linear filtering as with the adf program.

For both estimators a weighting function can be selected by the user. Supported weighting functions are constant (0), linear (1), gaussian (2) and none (3). If none is selected, the complex interferogram amplitude is ignored and only the interferometric phase is considered.

For constant weigting, the weight is 1.0 for pixels which are located closer than half of the indicated distance of the center pixel, and 0.0 for other pixels.
Triangular weighting decreases linearly with increasing distance from the center pixel. For the gaussian weighting function the weights are calculated using a gauss function of the distance from the center pixel. The advantage of using a weighting function is an improved frequency response of the averaging kernel. The estimated degree of coherence is written as output to the file <corr>.

The correlation can be displayed using the discc program and raster images produced using rascc in the DISP package.

SEE ALSO
discc, rascc.