ANSI-C program: ptarg_cal_MLI.c
NAME
ptarg_cal_MLI: Point target analysis and
radiometric calibration of slant-range and ground-range (GRD)
intensity images
SYNOPSIS
ptarg_cal_MLI
<MLI_par> <MLI> <r_samp> <az_samp>
<psigma> <c_r_samp> <c_az_samp>
<ptr_image> <r_plot> <az_plot> <pcal>
[osf] [win] [pltflg] [psz] [csz]
| <MLI_par> | (input) slant-range or ground-range intensity image
parameter file |
| <MLI> | (input) ground-range or slant range detected image in
FLOAT format |
| <r_samp> | point target range sample number |
| <az_samp> | point target azimuth line number |
| <psigma> |
radar cross-section of the
calibration target in m**2 |
| <c_r_samp> |
clutter region center
range sample number |
| <c_az_samp> |
clutter region center
azimuth line number |
| <ptr_image> | (output) oversampled point
target image fcomplex format, width: osf*psz |
| <r_plot> | (output) range point target response plot data (text format) |
| <az_plot> | (output) azimuth point target response plot data (text format) |
| [pcal] |
(output) measured point
target parameters and radiometric calibration factor (text
format) |
| [osf] | image over-sampling factor, 2, 4, 8, 16, 32, 64 (enter - for default: 16) |
| [win] |
(input) maximum search
window offset (samples) (enter - for default: 1) |
| [pltflg] |
plotting mode flag: 0: none 1: output plots in PNG format (default) 2: screen output and PNG format plots 3: output plots in PDF format |
| [psz] | point target region size (samples) (enter - for default: 16) |
| [csz] | clutter region size (samples) (enter - for default: 16) |
| theta_inc | incidence angle (RISAT-1, required for calibration of terrain corrrected images) |
EXAMPLE
ptarg_cal_MLI 19990421.mli
19990421.mli.par 203 48 10.0 203 70 cr2.slc
cr2.r_plot cr2.az_plot cr2.pcal
Integrates the target backscatter over an 16x16 window about
the coordinates specified on the command line. Determines the
exact point target location of a calibration target, oversamples
the MLI about the target, and determines the range and azimuth
response. The average clutter level is determine from a region
close to the target. This information is used to determine
the absolute radiometric calibration factor. A text file
pcal is generated that
contains the results of the radiometric analysis.
DESCRIPTION
ptarg_cal_MLI is a point target analysis tool used to determine the radiometric calibration factor for slant-range or ground-range intensity images. These images must contain a calibration target with known radar cross section (RCS) in order to radiometrically calibrate the scene. The input MLI (FLOAT format) is read and the response of the point target at or near the indicated location is analyzed. To estimate the approximate point target location the SLC can be displayed using the display program dispwr. The user must also provide the center coordinates of a clutter region chosen close to the point target and that is representative of the area where the point target is located. The radiometric calibration factor is determined using the integral algorithm described by Gray et al. 1990 [1].
When using this method with MLI data it is best that the data
have as few looks as possible since multilooking strongly reduces
the image resolution.
We start from an MLI or ground-range image that is corrected
for antenna pattern, slant range spreading loss, ellipsoidal
corrected reference area, except for the calibration
constant. This means that the SLC image is radiometrically
corrected except for the calibration constant and terrain
topography effects. If this is not the case then the SLC has to
be corrected using radcal_MLI. Sigma0 values have units
of m**2/m**2 (dimensionless) and Radar Cross Section is in units
of m**2.
Then RCS (radar cross section) of a calibrated (sigma0) MLI pixel
is K*A*Pi, where
K is the radiometric
calibration factor and Pi is the intensity of
the pixel. The defined reference area A is given by
ground_range_pixel_spacing * azimuth_pixel_spacing. The
RCS of an area is given by the sum of RCS of the pixels
within the area. This includes the calibration target RCS
and clutter RCS.
RCS_CR_with_clutter = K * SUM(A
* P_i)
over the 16x16 region where the sum is of the individual pixel
intensities Pi.
The clutter cross-section is determined using an area without the
target but has the same characteristics (area with the same
characteristics close to the calibration target).
RCS_clutter_per_pixel = K*
SUM(A*P_i)/N_c
over the clutter estimation region(32*32), where N_c is the number of pixels in the
clutter region(32*32).
RCS_CR = RCS_CR_with_clutter
- (N * RCS_clutter_per_pixel)
where N is number of samples in the 8x8 point target
region.
We can now solve for the calibration constant K:
K = RCS_CR/(SUM(A * P_i)-
(N * SUM(A * P_i)/N_c))
Note: the two sums are for the region with the target, and the
clutter region without the target respectively. This
expression can be simplied by extracting A from within the
sums:
K = RCS_CR/(A *
(SUM(P_i)- (N * SUM(P_i)/N_c))
The output pcal file has the following format:
#r_pix, az_pix,
r_pos, az_pos,
pk_mag, pk_phase_deg, r_3dB_width,
az_3dB_width, r_10dB_width, az_10dB_width, r_pslr_dB,
az_pslr_dB
202.871 47.350 681125.223
81.996 1.48090e+03
-123.904
1.771
2.966
3.118
5.277
-21.669 -23.996
#cl_r_pix, cl_az_pix, cl_ave_pwr,
ptr_range, inc_angle, ptr/clutter_dB, beta0_cl_dB,
sigma0_cl_dB, K_beta0_dB, K_sigma0_dB
203
70
1.2616e+04 681125.399
0.74661 26.7047
-20.4380
-22.1183 -61.4471
-63.1274
The first line describes the point target parameters including
the following parameters:
rpix:
range pixel position of the target peak (samples)
az_pix:
azimuth pixel position of the target peak (lines)
rpos:
slant range of the pixel peak
azpos:
meters along-track position of the pixel peak
pk_mag:
magnitude of the peak
pk_phase_deg: phase in degrees of the pixel
peak
r_3dB_width: 3 dB range width of the peak
in meters
az_3dB_width: 3 dB azimuth width of the peak in
meters
r_10dB_width: 10 dB range width of the peak in
meters
az_10dB_width 10 dB azimuth width of the peak
in meters
r_pslr_dB range peak
sidelobe ratio in dB
az_pslr_dB azimuth peak
sidelobe ratio in dB
The second line describes the clutter statistics and
radiometric calibration parameters:
cl_r_pix
clutter region center range pixel
cl_az_pix clutter
region center azimuth pixel (line)
cl_ave_pwr clutter average
power level in
ptr_range point target
range
inc_angle incidence
angle at the point target
ptr/clutter_dB ratio of the point target energy to
clutter in dB
beta0_cl_dB beta0 of the clutter
region dB
sigma0_cl_dB sigma0 of the clutter region
dB
K_beta0_dB calibration
constant for beta0 dB
k_sigma0_dB calibration constant
for sigma0 dB
For the point target response analysis, the MLI is oversampled
with an oversampling factor that can be set on the command
line using the osf
parameter. The location of the intensity maximum is determined.
Centered at the intensity maximum the range and azimuth response
is determined and stored as text files and optionally plotted
using Gnuplot. Parameters estimated from the data include the
sub-pixel peak location, expressed in input SLC samples, the
signal intensity in dB, the signal intensity in linear scale, and
the signal phase. The over-sampled MLI around the point target is
saved as a 2-D image in FLOAT format.
ptarg_cal_MLI estimates the the 3 dB and 10 dB peak widths in range and azimuth, and the Peak to Side-Lobe Ratio (PSLR). However these values are not nearly as accurate as with an SLC image because the phase information is not available in an intensity image and the image is somewhat undersampled. The search region for the point target peak is located in the center of the data segment extracted from the entire image as specified by the r_samp and az_samp parameters command-line parameters. The size of the search region defaults to +/- 1 pixel of the center coordinates but can be increase to +/- 4 samples by entering a value for the win command-line parameter. The interpolated image has dimensions determined by the over-sampling factor. The oversampled point target image is 256x256 for an 16x16 image oversampled with osf=16 and is fcomplex format with 0.0 phase.
The user can specify on the command line the size of the point-target window and the size of the clutter window using the command line parameters psz and csz The clutter window should be at least as large as the window used for extracting the point target. Increasing the size of the window for the point target beyond where there is any return will only increase errors since the contribution of the clutter has increased relative to the point target energy. It should usually not be necessary to change the size of the windows to differ from the default.
In the case when using this program with RISAT-1 intensity images that have been terrain geocoded, the incidence angle for a point must be provided by the user on the command line. The incidence angle can be determined using the (*_grd.txt) geometry grid file.
[1] Gray, L.A., P.
Vachon, C. Livingstone, T. Lukowski, "Synthetic Aperture Radar
Calibration using Reference Reflectors," IEEE Transactions on
Geoscience and Remote Sensing, Vol. 28, No. 3, pp. 374-383, May
1990.
SEE ALSO
Users Guide, ptarg_cal_SLC, radcal_MLI, ptarg_SLC