Medical Care |

Medical Care



Chaque forme pharmaceutique présente ses propres avantages et inconvénients acheter du amoxil.

mais n'ont pas d'effets néfastes pour l'organisme dans son ensemble.

StaMPS (Stanford Method for PS) Manual Nordic Volcanological Centre Institute of Earth Sciences University of Iceland Askja, 101 Reykjavik Version 2.1, June 1st, 2007 This manual provides a guide to running StaMPS, but does not attempt to explain all the process-ing. For some details on the inner workings, see Hooper et al. (2007), Hooper and Zebker (2007)and Hooper, Ph.D Thesis.
There are two pre-processing steps before getting to the PS processing proper. The first is to focusthe raw data, and the second is to form interferograms. ROI PAC is used for focusing and DORISfor interferogram formation. If starting with SLC images, rather than raw data, the focusing stepcan be skipped and the images imported directly into DORIS.
Both the ROI PAC and DORIS processing are non-standard and various shell scripts, matlabscripts and programs are included in this package to produce interferograms that are PS friendly.
The PS processing itself includes C++ programs and matlab scripts to identify PS pixels, and toextract the deformation signal for these pixels. Typing help followed by the name of the matlabscript should, in most cases, provide a brief description of the processing.
Throughout this manual, commands to be entered on the command line are in bold blue andentries that are specific to the data set being processed and may require modification are in boldred. The presence of >> before a command indicates that the command is a matlab script.
Install StaMPS:tar -xvf StaMPS v2.1.tarcd StaMPS/srcmakemake install dismph is a complex phase display program complied in the install that uses X11 Open Motif. Thisis installed as standard under most unix/linux operating systems (I think), and can be installed on OS X systems from - copy include/Xm to /usr/X11R6/includeand libXm.a to /usr/X11R6/lib. Alternatively another display program may be used instead ofdismph.
Details on installing ROI PAC (if needed) can be found here: http://www.openchannelfoundation.
org/projects/ROI PAC Details on installing DORIS can be found here: The Triangle program is used in 3-D phase unwrapping of PS pixels and can be found here:∼quake/triangle.html Edit the first 6 lines of StaMPS CONFIG to point to the correct directories in your case.
source StaMPS CONFIG Create SLCs (using ROI PAC) If you already have CEOS level 1 type SLCs skip this section.
Scripts have been updated to run with ROI PAC version 2.3. If you have version 2.2 installed, youshould run make slcs ers V2.2, make slcs envi V2.2, and remake slcs V2.2, instead of the currentversions.
In your processing directory:mkdir SLCcd SLCcp $MY SCR/roi.proc .
mkdir yyyymmdd for each scene and put raw data (IMAGERYyyymmdd) and leader file (SAR-LEADERyyyymmdd) here, or the ASA file for Envisat data.
Choose a master based on minimizing perpendicular, Doppler and temporal baselines (see Hooperet al., 2007). Substitute your master date in the format yyyymmdd wherever master date appearsbelow.
While in the SLC directory: echo master date > make slcs.listmake slcs ers or make slcs envicd master datemultilook master date.slc 6312 (or whatever is the filewidth)dismph master date.slc.4looks 1578 (former filewidth/4) Choose area of interest and note first and last azimuth line numbers (multiplying by 20) e.g. 15000and 20000 and first and last range pixel numbers (multiplying by 4) e.g. 2400 and 3400 cd . (back to SLC directory)vi roi.proc (or use your favorite editor) and change the following:before z ext = -12500 -(first az line minus 2500) N.B., include the minus signnumber of patches = 3 ((last line - first line + 2500)/3000 rounded up)mean pixel rng = 2900 (mean range pixel of area of interest) cd master datemultilook master date.slc 6312 (or whatever is the filewidth)dismph master date.slc.4looks 1578 (former filewidth/4) Find your area of interest and note the new first and last azimuth line numbers (multiplying by20).
cp $MY SCR/master .
vi master and update the crop area. first l and last l are the first and last azimuthline numbers, first p and last p are the first and last range pixels. To ensure round numbers,first l & first p should end in a 1 and last l & last p should end in a 0.
step master setupcd . (back to SLC directory) ls -d [1,2]* > make slcs.listvi make slcs.list and remove the entry for master datemake slcs ers or make slcs envi If needed, remake slcs will recreate SLCs for all entries in make slcs.list without rerunning themake step of ROI PAC.
It may become apparent later that one or more scenes are offset from the master by so much thatthe focussed image does not include the cropped master image. In which case copy roi.proc fromthe SLC directory to the relevant yyyymmdd directory, rename it yyyymmdd.proc and edit it sothat the the part processed includes the cropped master image. Edit make slcs.list to leave onlyscenes that need recreating and run remake slcs.
CEOS SLC (Level 1 product) If you have created SLCs with ROI PAC, skip this section.
mkdir SLCcd SLCmkdir yyyymmdd for each scene and put the corresponding ASA IMP 1P file in each directory,renamed to image.slc Choose a master based on minimizing perpendicular, Doppler and temporal baselines (see Hooperet al., 2007). Substitute your master date in the format yyyymmdd wherever master date appearsbelow.
cd master datestep master read wholedismph image.slc.4looks 1294 (filewidth/4) Choose area of interest and note first and last azimuth line numbers (multiplying by 20) e.g. 15000and 20000 and first and last range pixel numbers (multiplying by 4) e.g. 2400 and 3400 cp $MY SCR/master .
vi master and update the crop area. first l and last l are the first and last azimuthline numbers, first p and last p are the first and last range pixels. To ensure round numbers,first l & first p should end in a 1 and last l & last p should end in a 0.
cd . (back to SLC directory) ls -d [1,2]* > make slcs.listvi make slcs.list and remove the entry for master datemake read Create IFGs (using DORIS) In the same directory where SLC and INSAR master date reside:mkdir DEM Place your DEM in this directory. Resolution should be of order 20m.
cd INSAR master date If the SLCs weren't created by ROI PAC create master date.slc.rsc with the following line, sub-stituting the correct heading:HEADING vi dem.dorisin and update the following fields: CRD IN FORMAT real4CRD IN DEM /data/T156/DEM/NED THIRD ARCSEC.flt // posting in degrees 46.42181 -122.46252 // lat and lon of upper left CRD IN NODATA -9999 step master orbit ls -d ./SLC/[1,2]* > slcs.list (N.B. If the SLC directories are not in the same directory asINSAR master date, the full path must be given)vi slcs.list and remove master datemake orbits Check output from the Start coarse correl step in coreg.out in the yyyymmdd subdirectoriesthat are created: Coarse correlation translation lines: Coarse correlation translation pixels: These values should be approximately the median values from the data below them. Since DORISv 3.17, this is sometimes not the case and if wrong by more than a couple of pixels, you shouldmanually change them.
It may also be the case that there is a timing error in the orbit info and the approximate values in Start coarse orbits are too far from the real values for coarse correlation to work. In this case, estimate the coarse offsets yourself (look at the SLCs), update them in Start coarse orbits incoreg.out and rerun just the DORIS COARSECORR step.
Manual DEM offset correction Choose a slave close in time and space. In the yyyymmdd subdirectory for the chosen slave:step coregstep demstep resamplestep ifg matlab -nojvm -nosplash>>make amp dem(az down,rg right,red contrast,green brightness) • this will combine amplitude (red) and DEM (green).
• the first time, enter make amp dem(0,0). Estimate by eye the offset of the DEM in az and range, and rerun until ‘good' fit.
• you can adjust red contrast and green brightness (default 0.5 and 1) to vary contrast between amplitude image and DEM.
• Two offsets are output. Once happy with the fit, update these values in dem.dorisin and geocode.dorisin in the INSAR master date directory.
In the yyyymmdd directory:step dem>>make amp dem(0,0) and check that the fit is correct.
In the INSAR master date directory: make coreg (long runtime)By default all images with baseline < 100m are coregistered directly to the master and those withlarger baselines are coregistered to the 3 closest slave images with a smaller baseline. These defaultvalues can be changed by copying $DORIS SCR/make coreg to INSAR master date, editing thevalues at the top and running ./make coreg.
If rerunning, make coreg does not re-coregister scenes that have already been processed. If thisis required, delete the corresponding CPM Data.n1.n2 files in the coreg subdirectory, where n1and n2 refer to the order of the two coregistered scenes in make coreg.list (or 0 for the master), ordelete the entire coreg subdirectory to re-coregister all scenes.
Also by default, all cross-correlations with coherence greater than 0.3 are selected initially byDORIS. This value can be increased (by editing coreg.dorisin) if there is generally good coherenceto make run times faster or decreased if coherence is generally particularly bad, though I generallyfind that any cross-correlation with coherence below 0.12 is never correct.
make dems (long runtime)make resample or make filtazi resample The former does not filter in azimuth, and is therecommended approach for PS processing. The latter filters in azimuth before resampling.
make ifgsxv */*.ras and check that each interferogram looks OK make orbits processes all images listed in slcs.list make dems, make resample, make filtazi resample and make ifgs process all images listedin make ifgs.list, which is output by make orbits, but can be edited to add or drop images.
make coreg processes all images listed in make coreg.list in the coreg subdirectory. Extra imagescan be added to the bottom of this file, but no lines should ever be deleted, as n1 and n2 inCPM Data.n1.n2 files refer to the order of make coreg.list The following individual steps can be rerun in the individual yyyymmdd subdirectories of IN-SAR master date:step coreg coregisters the slave image directly to the master (may be different to results frommake coreg which includes slave-slave coregistration).
step resample resamples the slave image.
step filtazi resample filters both master and slave in azimuth and resamples the slave.
step dem creates the simulated dem interferogram.
step ifg creates the interferogram.
cd to your favorite interferogram directory and run:step geo (calculates the latitude and longitude of each pixel, only needs to be run once) Possible reasons for DORIS SIGERV error • master.res or slave.res (as specified in dorisin file) is missing • orbits are missing from master.res or slave.res • higher order coefficients in coregpm are too large - makes resampling impossible • slave SLC doesn't completely overlap the master cropped SLC. See discussion above on recre- ating the SLC using ROI PAC.
In the INSAR master date directory runps prep 0.4 3 2 50 200 where amplitude dispersion number of patches in range (default 1) number of patches in azimuth, (default 1) overlapping pixels between patches in range (default 50) overlapping pixels between patches in azimuth (default 200) The number of patches you choose will depend on the size of your area and the memory on yourcomputer. Generally, patches should be < 5 million SLC pixels.
matlab -nojvm -nosplash>>ps parms default You can modify the default parameters using >>setparm (refer to my thesis for meaning of mostparameters).
The default is to run all steps. A subset of steps can also be selected - see >>help stamps fordetails.
Steps 1 to 7 run by default on individual patches. Step 8 merges the patches into one patch andStep 9 runs on the merged patch. After merging, it is also possible to run Step 6 and/or 7 on themerged patch by setting the patch flag to ‘n', e.g.,>>stamps(6,7,‘n')This should in general produce more reliable results, but will take longer to run.
Step 6: Phase Unwrapping Processing is controlled by the following parameters: Unwrapping method.
Index to interferograms to be unwrapped.
unwrap prefilter flag Prefilter phase before unwrapping to reduce noise. Otheroption (not generally recommended) ‘n'.
unwrap patch phase Use the patch phase from Step 3 as prefiltered phase. If setto ‘n' (recommended), PS phase is filtered using a Goldsteinadaptive phase filter.
Resampling grid spacing. If unwrap prefilter flag is setto ‘y', phase is resampled to a grid with this spacing.
unwrap gold n win Window size for Goldstein filter Smoothing window (in days) for estimating phase noise dis-tribution for each pair of neighboring pixels. The time seriesphase for each pair is smoothed using a Gaussian windowwith standard deviation of this size. Original phase minussmoothed phase is assumed to be noise, which is used fordetermining probability of a phase jump between the pairin each interferogram.
Note that if re-running Step 6 and Step 7 has been run, estimates of SCLA and master atmosphereand orbit error will be subtracted before unwrapping. If you do not wish this to occur, reset these estimates before running Step 6 with>>scla reset Note also, however, that subtraction of SCLA and master atmosphere and orbit error has not beenimplemented with the unwrap prefilter flag = ‘n' option.
After running Step 6, display the output with>>ps plot(‘u') Check for unwrapping errors i.e., phase jumps in space which are uncorrelated in time. Unwrappingerrors are more likely to occur in longer perpendicular baseline interferograms. This is for tworeasons, firstly there is more noise associated with each PS, and secondly, the phase due to anyspatially-correlated look angle (SCLA) error is larger, as it is proportional to perpendicular baseline.
Noise is reduced by spatial filtering before unwrapping, but it is also possible to reduce the SCLAerror phase by estimating the SCLA error from the interferograms that unwrap OK (Step 7). IfStep 6 is re-run after Step 7 has been run, SCLA error phase is temporarily subtracted from thewrapped phase before unwrapping. Unwrapping accuracy is further improved by also temporarilysubtracting the atmosphere and orbit error phase due to the master image, present in all theinterferograms, which is also estimated in Step 7.
Step 7: Spatially-Correlated Look Angle Error Spatially-uncorrelated look angle error is calculated in Step 3 and removed in Step 5. After unwrap-ping, spatially-correlated look angle (SCLA) error is calculated which is due to spatially-correlatedDEM error (this includes error in the DEM itself, and incorrect mapping of the DEM into radarcoordinates). Master atmosphere and orbit error phase is estimated simultaneously.
Processing is controlled by the following parameters: Index to interferograms to be used in the SCLA estimation.
Display the estimate of SCLA error with>>ps plot(‘d') Units are phase per m of perpendicular baseline, with 0.01 rad/m correspondingto about 12 m of DEM error for ENVISAT I2 swath.
Display the estimate of master atmosphere and orbit error phase with>>ps plot(‘m') Unwrapped phase minus either or both of the above can be plotted with ‘u-d', ‘u-m' or ‘u-dm'.
After running, check that the estimates seem reasonable, i.e., ps plot(‘u-dm') looks generallysmoother than ps plot(‘u') (note that the default colour scales will be different). If not generallysmoother, one or more interferograms has probably unwrapped incorrectly. Drop it/them fromrecalc index and rerun Step 7, e.g., to drop the 13th and 14th interferograms,>>setparm(‘recalc in',[1:12,15:17])>>stamps(7,7) Once happy that all included interferograms are generally smoother, rerun Step 6. Step 6 willsubtract the estimates of SCLA and master atmosphere/orbit errors, before unwrapping and addthem back in afterwards (as long as unwrap prefilter flag = ‘y'). If more interferograms becomereliably unwrapped on re-running, add them into recalc index before running Step 7. This can be repeated until all interferograms are reliably unwrapped, or until no further improvement is seen.
If there is non-steady deformation present in some interferograms and, by chance, it correlateswith perpendicular baseline, it can get mapped into the SCLA error. This may be evidenced aspropagation of any deformation in ps plot(‘u-dm')to all interferograms (though the sign for eachwill depend on the perpendicular baseline sign), or correlation of ps plot(‘d') with ps plot(‘m').
If you suspect this is occurring, you can attempt to remove the deformation/baseline correlationby adding or subtracting interferograms from recalc index. Note that time and baseline info canbe displayed with>>ps info If some interferograms are still not reliably unwrapped, try setting unwrap patch phase to ‘n' andrerunning Step 6. This will use the filtered phase of the PS pixels only, rather than that derived inStep 3 for unwrapping. Try also increasing unwrap grid size to 200 m or more. This will reducethe effects of noise by smoothing more, but do not set it higher than the distance over which youexpect deformation phase to vary by about π/2. Another thing to try is dropping noisier pixels bysetting weed standard dev to a lower value, and re-running from Step 4.
The following matlab scripts plot the data in various ways (use >>help in matlab for options) Plots all multilooked interferograms Plots a series of PS phase values, on various backgrounds Plots a value for each PS, on various backgrounds You can select a reference area by setting parameters ref lon and ref lat. All plots will then bereferenced to the mean value for this area.
N.B. This list is not comprehensive.
• Initial beta release.
• Addition of make resample and make filtazi resample to give the option of filtering in azimuth. As this involves updating master.res differently for every image pair, a separatemaster.res is now maintained in each individual slave directory.
• Update to make coreg to be more efficient (uses a different strategy for picking which images to coregister).
• Addition of step coreg to allow coregistration for an individual slave image with the master • Update to make amp dem.m to display the image in matlab instead of using disrg • Error in ps load initial.m fixed so that individual PS bperp and look angle values are now • Addition of ps load dem.m to allow plotting of PS on shaded relief topography.
• Other tidying of code.
• Processing added to enable input of CEOS Level 1 SLC data.
• step master setup added.
• Extra step added to ps weed.m to drop pixels that are not correlated in time with surround- • Ability to process data in smaller patches added.
• Changes to way data saved, for efficiency.
• Changes to ps est gamma quick.m to make it restartable and to make convergence criteria more reliable.
• New statistical cost function 3-D unwrapping algorithm.
• Look angle bug fixed.
• Some bug fixes.
• Flattening/DEM processing changed back to that in Version 1.1 (to remove a bug that was • Changes for compatibility with 64-bit machines.
• New scripts for working with ESA level 1 SLCs.
• Change to ps weed.m to handle duplicate lat/lon assignment by DORIS.
• Estimation of spatially-correlated look angle (DEM) error and master atmosphere and orbit error added (Step 7).
• Unwrapping (Step 6) now uses estimates from Step 7 if present.
• Merging of patches is now an explicit step (Step 8).
• Estimation of spatially-correlated noise is now Step 9.
• ps info added Links to PDF files for all the following references can be find at:∼ahooper/pubs.html Hooper, A., P. Segall and H. Zebker, Persistent Scatterer InSAR for Crustal Deformation Analysis,with Application to Volcan Alcedo, Galapagos, Journal of Geophysical Research, in press, 2007.
Hooper, A., and H. Zebker, Phase unwrapping in three dimensions with application to InSAR timeseries, Journal of the Optical Society of America A (Optics, Image Science and Vision), in press,2007.
Hooper, A., H. Zebker, P. Segall, and B. Kampes, A New Method for Measuring Deformation onVolcanoes and Other Natural Terrains Using InSAR Persistent Scatterers, Geophys. Res. Letters,31, L23611, doi:10.1029/2004GL021737, 2004 Hooper, A., Persistent Scatterer Radar Interferometry for Crustal Deformation Studies and Mod-eling of Volcanic Deformation, Ph.D. thesis, Stanford University, 2006.



[ RESIDENT'S CASE PROBLEM ] MICHAEL S. CROWELL, PT, DPT¹šDEHC7DM$=?BB" PT, DSC, OCS, FAAOMPT² Medical Screening and Evacuation: Cauda Equina Syndrome in a Combat Zone Low back pain (LBP) is a prevalent condition, particularly in cific LBP, nerve root syndrome (radicu- primary care clinics, with billions of dollars spent each year

Services Available from Ryan's Dispensing of all Types of Prescriptions Regardless of what type of valid prescription you may have, we Tel: (045) 524747 are happy to fill it for you. This includes Medical Card (GMS) Pre-scriptions, Hospital Prescriptions, Psychiatric Prescriptions, Health Fax: (045) 528706 Amendment Prescription, Private (including Drug Payment