Next: Implementation Up: FILTPHASE Previous: Input Cards Contents

Output Description

The process control flag for this step is switched on in the products result file:

 filtphase:              1

And in the same result file a section will be added like (except if PF_IN_FILE is specified):

 *******************************************************************
 *_Start_filtphase:
 *******************************************************************
 Method: goldstein: size, alpha, overlap:        32 0.5 4
 Input_file:                             Outdata/cint.srp.raw
 Data_output_file:                       cint.0.5gf
 Data_output_format:                     complex_real4
 First_line (w.r.t. original_master):    1073
 Last_line (w.r.t. original_master):     4302
 First_pixel (w.r.t. original_master):   148
 Last_pixel (w.r.t. original_master):    985
 Multilookfactor_azimuth_direction:      10
 Multilookfactor_range_direction:        2
 Number of lines (multilooked):          323
 Number of pixels (multilooked):         419
 *******************************************************************
 * End_filtphase:_NORMAL
 *******************************************************************

**Figure 29.1:** Magnitude of unfiltered complex interferogram.
$\epsfig{file=Figures/cint_beforefilter_mag_lq.eps,width=\linewidth}$

**Figure 29.2:** Phase of unfiltered complex interferogram.
$\epsfig{file=Figures/cint_beforefilter_pha_lq.eps, width=\linewidth}$

**Figure 29.3:** Magnitude of filtered complex interferogram. (Method: spatialconv.) A spatial convolution with a kernel [1 4 9 4 1] was used. Clearly a lot of detail is lost.
$\epsfig{file=Figures/cint_spat14941_mag_lq.eps, width=\linewidth}$

**Figure 29.4:** Phase of filtered complex interferogram. (Method: spatialconv.) A spatial convolution with a kernel [1 4 9 4 1] was used. Clearly a lot of detail is lost.
$\epsfig{file=Figures/cint_spat14941_pha_lq.eps,width=\linewidth}$

**Figure 29.5:** Magnitude of filtered complex interferogram. (Method: spectral.) A pointwise multiplication in the spectral domain by a 32 point hamming filter was used, a blocksize of 32, and an overlap of 4.
$\epsfig{file=Figures/cint_hamm32_overl4_mag_lq.eps, width=\linewidth}$

**Figure 29.6:** Phase of filtered complex interferogram. (Method: spectral.) A pointwise multiplication in the spectral domain by a 32 point hamming filter was used, a blocksize of 32, and an overlap of 4.
$\epsfig{file=Figures/cint_hamm32_overl4_pha_lq.eps,width=\linewidth}$

**Figure 29.7:** Magnitude of filtered complex interferogram. (Method: goldstein.) parameters used are alpha=0.5, smooth=3, overlap=4. This filter seems to preserve most detail.
$\epsfig{file=Figures/cint_gold_a4_s3_o4_mag_lq.eps,width=\linewidth}$

**Figure 29.8:** Phase of filtered complex interferogram. (Method: goldstein.) parameters used are alpha=0.5, smooth=3, overlap=4. This filter seems to preserve most detail.
$\epsfig{file=Figures/cint_gold_a4_s3_o4_pha_lq.eps, width=\linewidth}$

Next: Implementation Up: FILTPHASE Previous: Input Cards Contents

Leijen 2009-04-14