# create Nansat object
n = Nansat(iPath + fileName)
# list bands and georeference of the object
print n
# Reprojected image into Lat/Lon WGS84 (Simple Cylindrical) projection
# 1. Cancel previous reprojection
# 2. Get corners of the image and the pixel resolution
# 3. Create Domain with stereographic projection, corner coordinates and resolution 1000m
# 4. Reproject
# 5. Write image
n.reproject() # 1.
lons, lats = n.get_corners() # 2.
pxlRes = distancelib.getPixelResolution(array(lats), array(lons), n[1])
pxlRes = array(pxlRes)*360/40000 # great circle distance
srsString = "+proj=latlong +datum=WGS84 +ellps=WGS84 +no_defs"
#~ extentString = '-lle %f %f %f %f -ts 3000 3000' % (min(lons), min(lats), max(lons), max(lats))
extentString = '-lle %f %f %f %f -tr %f %f' % (min(lons), min(lats), \
max(lons), max(lats), pxlRes[1], pxlRes[0])
d = Domain(srs=srsString, ext=extentString) # 3.
print d
n.reproject(d) # 4.
# get array with watermask (landmask) b
# it must be done after reprojection!
# 1. Get Nansat object with watermask
# 2. Get array from Nansat object. 0 - land, 1 - water
#wm = n.watermask(mod44path='/media/magDesk/media/SOLabNFS/store/auxdata/coastline/mod44w/')
wm = n.watermask(mod44path='/media/data/data/auxdata/coastline/mod44w/')
lns = lns + 180
try:
ncepGFSmodelWindSwath['wind_speed'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['wind_speed']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['wind_dir'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['wind_dir']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['u'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['u']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['v'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['v']).flatten(), (lats_2, lons_2), method='cubic')
except:
# ncepGFSmodelWindSwath['wind_speed'] = RectBivariateSpline(lts, lns, flipud(ncepGFSmodelWindSwath['wind_speed']), kx=2, ky=2)(lts_2, lns_2)
pass
# FOR SOME REASON RectSphereBivariateSpline doesn't work very good, using griddata instead
pxlResWindSwath = asarray(distancelib.getPixelResolution(lts_2, \
lns_2, \
lns_2.shape, 'km'))
print "Interpolated Wind cell resolution, %s km" % pxlResWindSwath
del lts, lns, lts_2, lns_2
# calculate bearing from initial lats/lons for further wind calculation
# Taking initial values as bearing is more accurate after
# interpolation than vice versa
bearing = zeros((GEOgrid['lons'].shape[0]-1, GEOgrid['lons'].shape[1]))
for n in range(0, GEOgrid['lons'].shape[1]):
col = ([GEOgrid['lats'][:-1, n], GEOgrid['lons'][:-1, n]],
[GEOgrid['lats'][1:, n], GEOgrid['lons'][1:, n]])
for m in range(0, GEOgrid['lons'].shape[0]-1):
def processingS1(inpath, fn, scale=1, resolution=80):
#~ Setting default resolution to 80m, which in most cases sorresponds to scale=8
pxlResSARm = resolution
s1 = readS1(inpath=inpath, fn=fn, resolution=pxlResSARm, min_lat=35)
if s1.lat_skip:
return None
# s1.__dict__['raw_counts']
# get vars from s1 class
# polarization
# raw_counts
# incidenceAngle_2
# sigmaNought_2
# noiseLut_2
# sigma0
# lons_2
# lats_2
# GEOgrid
# cLUTs
# nLUTs
# manifest
polarization = s1.polarization
incidenceAngle_2 = s1.incidenceAngle_2
sigma0 = s1.sigma0
lons_2 = s1.lons_2
lats_2 = s1.lats_2
GEOgrid = s1.GEOgrid
manifest = s1.manifest
del s1
gc.collect()
sigma0w = {}
roughness = {}
print "Scale set to: \'%s\' " % scale
for p in polarization:
print "Filtering Image: \'%s\' polarization" % p
# filter the image
sigma0w[p] = wiener(
sigma0[p][::scale, ::scale], mysize=(7, 7), noise=None
)
# sigma0w[p] = sigma0[p]
del sigma0
gc.collect()
# S1 Pixel resolution
# we use pxlResSAR for further GSHHS rasterizing and
# reprojecting data with pyresample
lonlim = (lons_2[::scale, ::scale].min(), lons_2[::scale, ::scale].max())
latlim = (lats_2[::scale, ::scale].min(), lats_2[::scale, ::scale].max())
# enlarge lonlims for cropping a bit larger area for masking
lonlimGSHHS = (lonlim[0]-1.0, lonlim[1]+1.0)
latlimGSHHS = (latlim[0]-1.0, latlim[1]+1.0)
# Get first guess pixel resolution
pxlResSARm = asarray(
distancelib.getPixelResolution(
lats_2[::scale, ::scale], lons_2[::scale, ::scale],
lons_2[::scale, ::scale].shape, 'km'
)
)*1e3
pxlResSARdeg = asarray(
distancelib.getPixelResolution(
lats_2[::scale, ::scale], lons_2[::scale, ::scale],
lons_2[::scale, ::scale].shape, 'deg'
)
)
# Define areas with pyresample
swath_def = pr.geometry.SwathDefinition(
lons=lons_2[::scale, ::scale], lats=lats_2[::scale, ::scale]
)
area_def_4326 = swath_area_def(name='Temporal SWATH EPSG Projection 4326',
proj='eqc', lonlim=lonlimGSHHS,
latlim=latlimGSHHS, ellps="WGS84",
res=pxlResSARm)
# Get the SAR pixel resolution from the area_def
# for further identical shapes
up = min(latlimGSHHS)
down = max(latlimGSHHS)
left = min(lonlimGSHHS)
right = max(lonlimGSHHS)
area_extent_deg = (left, down, right, up)
area_extent_deg_shape = area_def_4326.shape
pxlResSARdeg = asarray(
(abs(area_extent_deg[2] - area_extent_deg[0]) /
float(area_extent_deg_shape[1]),
abs(area_extent_deg[3] - area_extent_deg[1]) /
float(area_extent_deg_shape[0]))
)
pxlResSARm = asarray(
(area_def_4326.pixel_size_x, area_def_4326.pixel_size_y)
)
print "S1 cell resolution, %s deg" % str(pxlResSARdeg)
print "S1 cell resolution, %s m" % str(pxlResSARm)
# Apply Mask from GSHHS
reload(gshhs_rasterize)
# ESRI shapefile containing land polygons
shapefile = '/media/SOLabNFS/store/auxdata/coastline/GSHHS_shp/f/GSHHS_f_L1.shp'
# reproject GSHHS onto S1 grid before calculations
print "Rasterizing Land Mask"
mask_arr_4326 = gshhs_rasterize.gshhs_rasterize_4326(
lonlimGSHHS, latlimGSHHS, pxlResSARdeg, area_def_4326.shape,
True, shapefile
)
del pxlResSARdeg
mask_arr_swath = pr.kd_tree.resample_nearest(
area_def_4326, mask_arr_4326, swath_def,
radius_of_influence=4*pxlResSARm.max(), epsilon=0.5, fill_value=None
)
# Commented until roughness is needed
# # Nice Image (Roughness)
# sigma0wAvg = {}
# roughnessNrmlzd = {}
# if len(polarization[0]) >= 2: # if 2 polarizations
# for p in polarization:
# print "Nice Image: \'%s\' polarization" % p
# NB!!! DO NOT MASK LAND for Sigma0w and Roughness!!!!
# roughness[p] = ma.masked_where(mask_arr_swath, sigma0w[p])
# sigma0wAvg[p] = ma.median(roughness[p], axis=0)
# roughnessNrmlzd[p] = (roughness[p]-sigma0wAvg[p])/sigma0wAvg[p]
# elif len(polarization[0]) == 1: # if only 1 polarization
# p = polarization
# print "Nice Image: \'%s\' polarization" % p
# roughness[p] = ma.masked_where(mask_arr_swath, sigma0w[p])
# sigma0wAvg[p] = ma.median(roughness[p], axis=0)
# roughnessNrmlzd[p] = (roughness[p]-sigma0wAvg[p])/sigma0wAvg[p]
# del roughness, sigma0wAvg
# Adding Model wind
# import xmltodict
# zf = zipfile.ZipFile(inpath+fn, 'r')
# manifest = zf.read(fn[:-4] + '.SAFE/manifest.safe')
# manifest = xmltodict.parse(manifest) # Parse the read document string
# zf.close()
Objects = manifest['xfdu:XFDU']['metadataSection']['metadataObject']
for Object in Objects:
try:
startTime = datetime.datetime.strptime(
Object['metadataWrap']['xmlData']['safe:acquisitionPeriod']['safe:startTime'],
"%Y-%m-%dT%H:%M:%S.%f"
)
break
except:
pass
ncepGFSmodelWind = ncepGFSmodel(startTime, lats_2, lons_2)
# Reprojecting data
# Pixel resolution
# we use pxlResWind/pxlResSAR for further pyresample
# radius_of_influence and sigmas
pxlResWind = asarray(
distancelib.getPixelResolution(
ncepGFSmodelWind['lats_wind'],
ncepGFSmodelWind['lons_wind'],
ncepGFSmodelWind['lons_wind'].shape, 'km'
)
)
# Note pxlResWind is in KM, multiply by 1e3 for meters
print "S1 cell resolution, %s m" % pxlResSARm
print "Wind cell resolution, %s km" % pxlResWind
# reproject NCEP onto S1 grid before calculations
# Using RectSphereBivariateSpline - Bivariate spline approximation over a rectangular mesh on a sphere
# as it is much more efficiant for full resolution
# as well as smoothes nicely the image
# We don't want to work with full res wind so scaling the image for about 100m resolution
# Adjust scale to get appropriate value
# scale = 5
lts = ncepGFSmodelWind['lats_wind']
lns = ncepGFSmodelWind['lons_wind']
ncepGFSmodelWindSwath = {}
# check that lns increasing, if not
if ~all(diff(lns[0,:]) > 0):
lns_ = lns[0,:]
# we must start lns from -180 increasing to +180
# so we reconcatenate lns array and data array, by cropping part of array and putting in front
lns_ = concatenate((lns_[(lns_>=-180) & (lns_<0)], lns_[(lns_>=0)]),0)
ncepGFSmodelWindSwath['wind_speed'] = (ncepGFSmodelWind['wind_speed'])
ncepGFSmodelWindSwath['wind_speed'] = concatenate((ncepGFSmodelWind['wind_speed'][:,(lns_>=-180) & (lns_<0)],\
ncepGFSmodelWind['wind_speed'][:,(lns_>=0)]),1)
ncepGFSmodelWindSwath['wind_dir'] = (ncepGFSmodelWind['wind_dir'])
ncepGFSmodelWindSwath['wind_dir'] = concatenate((ncepGFSmodelWind['wind_dir'][:,(lns_>=-180) & (lns_<0)],\
ncepGFSmodelWind['wind_dir'][:,(lns_>=0)]),1)
# only then we do interpolate
lts_2 = lats_2[::scale,::scale]
lns_2 = lons_2[::scale,::scale]
# RectSphereBivariateSpline uses lats and lons within the intervals (0, pi), (0, 2pi).
# Make sure the latitude is between 0 .. 180
if lts_2.min()<0:
lts_2 = lts_2 + 90
lts = lts + 90
if lns_2.min()<0:
lns_2 = lns_2 + 180
lns = lns + 180
try:
ncepGFSmodelWindSwath['wind_speed'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['wind_speed']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['wind_dir'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['wind_dir']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['u'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['u']).flatten(), (lats_2, lons_2), method='cubic')
ncepGFSmodelWindSwath['v'] = griddata((lts.flatten(), lns.flatten()), (ncepGFSmodelWind['v']).flatten(), (lats_2, lons_2), method='cubic')
except:
# ncepGFSmodelWindSwath['wind_speed'] = RectBivariateSpline(lts, lns, flipud(ncepGFSmodelWindSwath['wind_speed']), kx=2, ky=2)(lts_2, lns_2)
pass
def main( argv=None ):
year = '2012'
useMask = False
if argv is None:
argv = sys.argv
if argv is None:
print ( "Please specify the path/year to the asar folder! \n")
return
# Parse arguments
try:
opts, args = getopt.getopt(argv,"hi:o:",["year=","oPath=","iPath=","useMask="])
except getopt.GetoptError:
print 'readASAR.py -year <year> ...'
sys.exit(2)
for opt, arg in opts:
if opt == '-h':
print 'readASAR.py -year <year> ...'
sys.exit()
elif opt in ("-year", "--year"):
year = arg
elif opt in ("-oPath", "--oPath"):
oPath = arg
elif opt in ("-iPath", "--iPath"):
iPath = arg
elif opt in ("-useMask", "--useMask"):
useMask = arg
oPath = '/media/SOLabNFS2/tmp/roughness/' + year + '/'
iPath = '/media/SOLabNFS2/store/satellite/asar/' + year + '/'
if not os.path.exists(oPath):
os.makedirs(oPath)
dirNames=os.listdir(iPath)
for dirName in dirNames:
fileNames=os.listdir(iPath+dirName)
for fileName in fileNames:
figureName = oPath + fileName[0:27] + '/' + fileName + '_proj.png'
kmlName = oPath + fileName[0:27] + '/' + fileName + '.kml'
if not os.path.exists(oPath + fileName[0:27] + '/'):
os.makedirs(oPath + fileName[0:27] + '/')
if os.path.isfile(kmlName):
print "%s already processed" % (fileName)
continue
else:
print "%s" % (fileName)
# try to create Nansat object
try:
n = Nansat(iPath + dirName + '/' + fileName, mapperName='asar', logLevel=27)
except Exception as e:
print "Failed to create Nansat object:"
print str(e)
os.rmdir(oPath + fileName[0:27] + '/' )
continue
#~ Get the bands
raw_counts = n[1]
inc_angle = n[2]
#~ NICE image (roughness)
pol = n.bands()[3]['polarization']
if pol == 'HH':
ph = (2.20495, -14.3561e-2, 11.28e-4)
sigma0_hh_ref = exp( ( ph[0]+inc_angle*ph[1]+inc_angle**2*ph[2])*log(10) )
roughness = n[3]/sigma0_hh_ref
elif pol == 'VV':
pv = (2.29373, -15.393e-2, 15.1762e-4)
sigma0_vv_ref = exp( ( pv[0]+inc_angle*pv[1]+inc_angle**2*pv[2])*log(10) )
roughness = n[3]/sigma0_vv_ref
#~ Create new band
n.add_band(bandID=4, array=roughness, \
parameters={'name':'roughness', \
'wkv': 'surface_backwards_scattering_coefficient_of_radar_wave', \
'dataType': 6})
# Reproject image into Lat/Lon WGS84 (Simple Cylindrical) projection
# 1. Cancel previous reprojection
# 2. Get corners of the image and the pixel resolution
# 3. Create Domain with stereographic projection, corner coordinates 1000m
# 4. Reproject
# 5. Write image
n.reproject() # 1.
lons, lats = n.get_corners() # 2.
# Pixel resolution
#~ pxlRes = distancelib.getPixelResolution(array(lats), array(lons), n.shape())
#~ pxlRes = array(pxlRes)*360/40000 # great circle distance
pxlRes = array(distancelib.getPixelResolution(array(lats), array(lons), n.shape(), 'deg'))
ipdb.set_trace()
if min(lats) >= 65 and max(lats) >= 75 and max(lats)-min(lats) >= 13:
pxlRes = array([0.00065, 0.00065])*2 # make the resolution 150x150m
#~ pxlRes = pxlRes*7 # make the resolution worser
srsString = "+proj=latlong +datum=WGS84 +ellps=WGS84 +no_defs"
#~ extentString = '-lle %f %f %f %f -ts 3000 3000' % (min(lons), min(lats), max(lons), max(lats))
extentString = '-lle %f %f %f %f -tr %f %f' % (min(lons), min(lats), \
max(lons), max(lats), pxlRes[1], pxlRes[0])
d = Domain(srs=srsString, ext=extentString) # 3.
n.reproject(d) # 4.
if useMask:
# get array with watermask (landmask) b
# it must be done after reprojection!
# 1. Get Nansat object with watermask
# 2. Get array from Nansat object. 0 - land, 1 - water
#wm = n.watermask(mod44path='/media/magDesk/media/SOLabNFS/store/auxdata/coastline/mod44w/')
wm = n.watermask(mod44path='/media/data/data/auxdata/coastline/mod44w/')
wmArray = wm[1]
#~ ОШИБКА numOfColor=255 не маскирует, потому что в figure.apply_mask: availIndeces = range(self.d['numOfColor'], 255 - 1)
#~ n.write_figure(fileName=figureName, bands=[3], \
#~ numOfColor=255, mask_array=wmArray, mask_lut={0: 0},
#~ clim=[0,0.15], cmapName='gray', transparency=0) # 5.
n.write_figure(fileName=figureName, bands=[4], \
mask_array=wmArray, mask_lut={0: [0,0,0]},
clim=[0,2], cmapName='gray', transparency=[0,0,0]) # 5.
else:
n.write_figure(fileName=figureName, bands=[1], \
clim=[0,2], cmapName='gray', transparency=[0,0,0]) # 5.
# open the input image and convert to RGBA for further tiling with slbtiles
input_img = Image.open(figureName)
output_img = input_img.convert("RGBA")
output_img.save(figureName)
# make KML image
n.write_kml_image(kmlFileName=kmlName, kmlFigureName=figureName)
#~ Change the file permissions
os.chmod(oPath, 0777)
os.chmod(oPath + fileName[0:27] + '/', 0777)
os.chmod(kmlName, 0777)
os.chmod(figureName, 0777)
#~ Change the owner and group
#~ os.chown(oPath, 1111, 1111)
#~ os.chown(oPath + fileName[0:27] + '/', 1111, 1111)
#~ os.chown(kmlName, 1111, 1111)
#~ os.chown(figureName, 1111, 1111)
#~ garbage collection
gc.collect()
def main( argv=None ):
# default values
pn = '/media/data/data/OTHER/RS2 Agulhas and Lion/RS2_FQA_1xQGSS20101218_173930_00000005/'
xOff=0
yOff=0
xS=None
yS=None
xBufScale=None
yBufScale=None
s = 'abs'
if argv is None:
argv = sys.argv
if argv is None:
print ( "Please specify the path to the RS2 folder! \n" + \
"See USAGE for more details \n")
return Usage()
# Parse arguments
try:
opts, args = getopt.getopt(argv,"hi:o:",["pn=","xoff="])
except getopt.GetoptError:
print 'readRS2.py -pn <inputfile> ...'
sys.exit(2)
for opt, arg in opts:
if opt == '-h':
print 'readRS2.py -pn <inputfile> ...'
sys.exit()
elif opt in ("-pn", "--pn"):
pn = arg
elif opt in ("-xoff", "--xoff"):
xOff = arg
elif opt in ("-yoff", "--yoff"):
yOff = arg
elif opt in ("-xS", "--xS"):
xS = arg
elif opt in ("-yS", "--yS"):
yS = arg
elif opt in ("-xScale", "--xScale"):
xBufScale = arg
elif opt in ("-yScale", "--yScale"):
yBufScale = arg
print "Path to the RS2 file:\n", pn
if core_ct < 4:
# calibrate the scene
print "Calibrating"
currtime = time()
calib = calibRS2(pn, xOff, yOff, xS, yS, xBufScale, yBufScale, 'sigma0')
# get the incidence angle
print "Incidence Angle and Lats/Lons"
IncidenceAngle = calib.incidence_angle()
IncidenceAngle = rad2deg(IncidenceAngle)
# get the lats/lons
(lat, lon, line, pixel) = calib.xml2geo()
print "Filtering"
# 1. Remove speckle noise (using Lee-Wiener filter)
calib.speckle_filter('wiener', 7)
print 'Serial: time elapsed:', time() - currtime
elif core_ct >= 4:
# calibrate the scene
print "Calibrating"
currtime = time()
calibPar = calibRS2par(pn, xOff, yOff, xS, yS, xBufScale, yBufScale, s)
# get the incidence angle
print "Incidence Angle and Lats/Lons"
IncidenceAngle = calibPar.incidence_angle()
IncidenceAngle = rad2deg(IncidenceAngle)
# get the lats/lons
(lat, lon, line, pixel) = calibPar.xml2geo()
print "Filtering"
# 1. Remove speckle noise (using Lee-Wiener filter)
calibPar.speckle_filter('wiener', 7)
print 'Parallel: time elapsed: %f' % ( time() - currtime )
# export to geotiff
# calibPar.save_tiff()
print "Some Sigma calculations..."
SigmaHHwnr = calibPar.SigmaHHwnr
SigmaVVwnr = calibPar.SigmaVVwnr
# SigmaHVwnr = calibPar.SigmaHVwnr
# Free up some memory
del calibPar.SigmaHHwnr, calibPar.SigmaVVwnr, calibPar.SigmaHVwnr
# # Calculate some wind
# # testing that with sar=-0.387 wind speed is 10m/s
# w = rcs2wind(sar=-0.387*ones((1,1)), cmdv=4, windir=0, theta=20*ones((1,1)))
# print "Testing CMOD4 passed, Wind =", w
#
# sar=10*log10(SigmaVVwnr[::10,::10])
## sar=10*log10(SigmaVVwnr)
# w = rcs2wind(sar, cmdv=4, windir=130, theta=33*ones(sar.shape))
# print "Mean CMOD4 Wind =", w.mean()
# calculate the conjugate
SigmaVVConjHH = calibPar.S_VV*conj(calibPar.S_HH)
SigmaVVConjHHWnr = wiener(SigmaVVConjHH, mysize=(7,7), noise=None)
# all sigmas in linear units
SigmaVVConjHHdelta = SigmaVVConjHHWnr/(SigmaVVwnr - SigmaHHwnr)
# # Check that imagenary part is much less then the real one
# SigmaVVConjHHr = real(SigmaVVConjHH)
# SigmaVVConjHHi = imag(SigmaVVConjHH)
# SigmaVVConjHHr.mean()
# SigmaVVConjHHi.mean()
# Free up some memory
del SigmaVVConjHH, calibPar.S_HH, calibPar.S_VV
# 2. Find Poolarization ratio PR in dB and delta - linear
# delta - is a Bragg component, which describes wind field impact
# PR = 10*log10(SigmaHHwnr) - 10*log10(SigmaVVwnr)
# # PR must not be > 0, be careful with masking
#P[P>0] = 0
PR = SigmaHHwnr/SigmaVVwnr
delta = SigmaVVwnr - SigmaHHwnr
# # Masking Land from HV polarization
# maskedLand = SigmaHVwnr>0.003
# 3. Linear fit of Sigma and delta: Sigma = A*delta
# A = sum(SigmaVVwnr[~maskedLand]*delta[~maskedLand])/sum(delta[~maskedLand]^2);
# B = sum(SigmaHHwnr[~maskedLand]*delta[~maskedLand])/sum(delta[~maskedLand]^2);
# imposing from theory
A = linspace(2.3,1.8,IncidenceAngle.size) # dependence on inc angle
A = tile(A, (line[-1,0]+1, 1))
# 4. Calculate SigmaVVwb = SigmaVVwb === SigmaHHwb
# substracting wind field variability contribution, the rest is impact of
# current and film on the image
SigmaWb = SigmaVVwnr - ( A*delta )
# Checking that SigmaWb from VV and HH are the same
# SigmaVVwb = SigmaVVwnr - ( A*delta )
# SigmaHHwb = SigmaHHwnr - ( B*delta )
# SigmaWb = SigmaVVwb
# mean(SigmaHHwb(:)./SigmaVVwb(:))
# clear SigmaVVwb SigmaHHwb
# Cropping the oil slicks
# проверить почему переворот изображения!!!
# print ( "Cropping...")
# plt.figure()
# plt.imshow(delta)
# plt.clim(-0.01,0.03)
# plt.gray()
# plt.show()
# pts = imzoom()
ptsCropSouth = array([[ 3200, 4550],
[ 3800, 5100]])
oilCropSouth = imcrop(ptsCropSouth, delta)
ptsCropNorth = array([[ 3400, 0],
[ 4900, 1650]])
oilCropNorth = imcrop(ptsCropNorth, delta)
# reduce the speckle noise
oilN = wiener(oilCropNorth, mysize=(3,3), noise=None)
oilS = wiener(oilCropSouth, mysize=(3,3), noise=None)
# use median filter to smooth
oilN = median_filter(oilN, radius=37)
oilS = median_filter(oilS, radius=37)
# find the global threshold
global_threshN = threshold_otsu(oilN)
global_threshS = threshold_otsu(oilS)
# create a binary mask
markersN = ones(oilN.shape, dtype=uint)
markersS = ones(oilS.shape, dtype=uint)
markersN[oilN < global_threshN*0.84] = 0
markersS[oilS < global_threshS*0.84] = 0
# Find the relation of PR in/out-side the Slicks
oilCropNorthPR = imcrop(ptsCropNorth, PR)
oilCropPN = oilCropNorthPR[markersN==0].mean()/oilCropNorthPR[markersS==1].mean()
oilCropSouthP = imcrop(ptsCropSouth, PR)
oilCropPS = oilCropSouthP[markersS==0].mean()/oilCropSouthP[markersS==1].mean()
# Checking that calculations are OK
oilCropNorthKVV = imcrop(ptsCropNorth, SigmaVVwnr)
KVV = oilCropNorthKVV[markersN==0].mean()/oilCropNorthKVV[markersN==1].mean()
oilCropNorthKHH = imcrop(ptsCropNorth, SigmaHHwnr)
KHH = oilCropNorthKHH[markersN==0].mean()/oilCropNorthKHH[markersN==1].mean()
PSlick = KHH/KVV*oilCropNorthKHH[markersN==1].mean()/oilCropNorthKVV[markersN==1].mean()
print "Polarization ratio Out/In-side Slick = \
\n %0.2f for Northern \n %0.2f for Southern" \
%(1/oilCropPN, 1/oilCropPS)
oilCropNorthVV = imcrop(ptsCropNorth, SigmaVVwnr)
oilCropNorthDelta = imcrop(ptsCropNorth, delta)
oilCropNorthSigmaWb = imcrop(ptsCropNorth, SigmaWb)
# Interpolating lat/lon to image size for future pcolormesh
from numpy import linspace, arange, round, floor, ceil
from scipy.interpolate import interp2d
def interpLL(l, pixel, line, gx, gy):
"""
Interpolating lat/lon to image size for future pcolormesh
"""
fl = interp2d(pixel[0,:], line[:,0], l, kind='linear')
l_new = fl(gx, gy)
return l_new
ptsCrop = ptsCropNorth
RasterXSize = round((-ptsCrop[0,0]+ptsCrop[-1,0]))
RasterYSize = round((-ptsCrop[0,-1]+ptsCrop[-1,-1]))
gx = linspace(ptsCrop[0,0], ptsCrop[-1,0], RasterXSize+1)
gy = linspace(ptsCrop[0,-1], ptsCrop[-1,-1], RasterYSize+1)
RasterXSize = round(pixel[0,-1])
RasterYSize = round(line[-1,0])
gx = linspace(0, pixel[0,-1], RasterXSize+1)
gy = linspace(0, line[-1,0], RasterYSize+1)
lat_new = interpLL(lat, pixel, line, gx, gy)
lon_new = interpLL(lon, pixel, line, gx, gy)
# save to MAT file
from scipy.io import savemat
print "Exporting"
pn = '/home/mag/Documents/MATLAB/Leon_Agulhas_RS2/'
expFn = 'OilSlickTrns/OilSlickTrns.mat'
if not path.isfile(pn + expFn):
print "Exporting to:\n" , pn + expFn
savemat(pn + expFn, mdict={ \
'VV':oilCropNorthVV, \
'PD':oilCropNorthDelta, \
'SigmaWb':oilCropNorthSigmaWb, \
'PR':oilCropNorthPR, \
'lat':lat_new, 'lon':lon_new, \
}, do_compression=False)
expFn = 'CurrentFrontTrns/CurrentFrontTrns.mat'
if not path.isfile(pn + expFn):
print "Exporting to:\n" , pn + expFn
savemat(pn + expFn, mdict={ \
'PD':delta, \
'PR':PR, \
'VV':SigmaVVwnr, \
'SigmaWb':SigmaWb, \
'lat':lat_new, 'lon':lon_new, \
}, do_compression=False)
import h5py
pn = '/home/mag/Documents/MATLAB/Leon_Agulhas_RS2/'
expFn = 'CurrentFrontTrns/CurrentFrontTrns.mat'
expFn = 'OilSlickTrns/OilSlickTrns.mat'
f = h5py.File(pn + expFn)
SigmaWb = f["SigmaWb"]
SigmaWb = flipud(rot90(array(SigmaWb)))
lat = f["lat"]
lat = flipud(rot90(array(lat)))
lon = f["lon"]
lon = flipud(rot90(array(lon)))
y = f["tWb"]['transx']
y = array(y)
x = f["tWb"]['transy']
x = array(x)
pts = zeros((2,2))
pts[:,0] = x
pts[:,1] = y
# transec_init = f["tWb"]['transec_init']
# transec_init = array(transec_init)
import distancelib
reload(distancelib)
# Get pixel resolution in Col/Row direction
PxResCol, PxResRow = distancelib.getPixelResolution(lat, lon, SigmaWb)
# Get the length of the transect. Input coords in (row,col)
trnsLength = distancelib.getDistancePx([pts[0,1], pts[0,0]], [pts[1,1], pts[1,0]], lat, lon, SigmaWb)
import transect
reload(transect)
transect.plotTrns(SigmaWbtrans, pts, lat, lon, SigmaWb, pn)
# Making transect
plt.figure()
plt.imshow(oilCropNorthDelta)
plt.gray()
plt.clim(0, 0.02)
plt.colorbar()
pts = transect.imagePts()
plt.savefig('/home/mag/PDtrns.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
# Transect North Oil Slick
pts = array([[610, 871], [822, 1007]])
# [1] returns only Mean Transect
PDtrans = transect.transect(oilCropNorthDelta, pts, 30)[1]
PRtrans = transect.transect(oilCropNorthPR, pts, 30)[1]
VVtrans = transect.transect(oilCropNorthVV, pts, 30)[1]
SigmaWbtrans = transect.transect(oilCropNorthSigmaWb, pts, 30)[1]
# PDtrans = transect.transect(oilCropNorthDelta, pts, 100)[0]
plt.figure()
plt.imshow(CurrentWb)
plt.gray()
plt.clim(0, 0.03)
plt.colorbar()
pts = transect.imagePts()
plt.savefig('/home/mag/PDtrns.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
# Transect Current Front
pts = array([[1207., 4318.], [2473., 3763.]])
PDtrans = transect.transect(delta, pts, 30)[1]
PRtrans = transect.transect(PR, pts, 30)[1]
VVtrans = transect.transect(SigmaVVwnr, pts, 30)[1]
SigmaWbtrans = transect.transect(SigmaWb, pts, 30, method='cubic')[1]
# use median filter to smooth
from scipy.ndimage import median_filter as mflt
PRtransSmthd = mflt(PRtrans, size=21, mode='nearest')
PDtransSmthd = mflt(PDtrans, size=21, mode='nearest')
VVtransSmthd = mflt(VVtrans, size=21, mode='nearest')
SigmaWbtransSmthd = mflt(SigmaWbtrans, size=21, mode='nearest')
plt.figure()
# plt.imshow(SigmaWb)
plt.imshow(SigmaWb)
plt.gray()
plt.clim(0, 0.02)
plt.colorbar()
ax = plt.gca().get_xbound()
ay = plt.gca().get_ybound()
x = pts[:, 0]
y = pts[:, 1]
plt.plot(x, y, '-o')
plt.text(x[0],y[0],"A",bbox=dict(facecolor='w', alpha=0.7),stretch='expanded',fontsize=21)
plt.text(x[1],y[1],"B",bbox=dict(facecolor='w', alpha=0.7),stretch='expanded',fontsize=21)
plt.gca().set_xbound(ax)
plt.gca().set_ybound(ay)
plt.draw()
# window = gtk.Window()
# screen = window.get_screen()
# width = screen.get_width()
# height = screen.get_height()
# mng = plt.get_current_fig_manager()
# mng.resize(width, height)
plt.savefig('/home/mag/PDtrns.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
plt.figure()
plt.plot(PDtransSmthd)
plt.grid()
# window = gtk.Window()
# screen = window.get_screen()
# width = screen.get_width()
# height = screen.get_height()
# mng = plt.get_current_fig_manager()
# mng.resize(width, height)
plt.savefig('/home/mag/PDtrnsLine.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
plt.figure()
plt.plot(PRtransSmthd)
plt.grid()
# window = gtk.Window()
# screen = window.get_screen()
# width = screen.get_width()
# height = screen.get_height()
# mng = plt.get_current_fig_manager()
# mng.resize(width, height)
plt.savefig('/home/mag/PRtransLine.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
plt.figure()
plt.plot(VVtransSmthd)
plt.grid()
# window = gtk.Window()
# screen = window.get_screen()
# width = screen.get_width()
# height = screen.get_height()
# mng = plt.get_current_fig_manager()
# mng.resize(width, height)
plt.savefig('/home/mag/VVtransLine.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
plt.figure()
plt.plot(SigmaWbtransSmthd)
plt.grid()
# window = gtk.Window()
# screen = window.get_screen()
# width = screen.get_width()
# height = screen.get_height()
# mng = plt.get_current_fig_manager()
# mng.resize(width, height)
plt.savefig('/home/mag/SigmaWbtransLine.png', facecolor='w', edgecolor='w', \
dpi=100, bbox_inches="tight", pad_inches=1.75)
plt.close('all')
print ( "Plotting on a map... \nUsing default NSPER Basemap \n")
# setup nsper basemap
# Lat/Lon coords of image corners
ll_lat = lat.min()
ur_lat = lat.max()
ll_lon = lon.min()
ur_lon = lon.max()
cent_lat = lat.mean()
cent_lon = lon.mean()
m = Basemap(llcrnrlat=ll_lat, urcrnrlat=ur_lat,\
llcrnrlon=ll_lon, urcrnrlon=ur_lon, \
resolution='f', projection='nsper', \
satellite_height=798000, \
lat_0=cent_lat,lon_0=cent_lon)
# m = Basemap(llcrnrlat=30, urcrnrlat=50,\
# llcrnrlon=1, urcrnrlon=20, \
# resolution='l', projection='nsper', \
# satellite_height=798000, \
# lat_0=cent_lat,lon_0=cent_lon)
print ( "Plotting figures... \n")
scale = 8 # setting scale factor to resize in 1/scale times
label1 = 'Sigma VV [dB]'
label2 = 'Sigma HH [dB]'
label11 = 'Sigma VV [linear units]'
label22 = 'Sigma HH [linear units]'
label3 = 'PR [dB]'
label33 = 'PR [linear units]'
label4 = 'PD [linear units]'
label5 = 'Wb contribution [linear units]'
label6 = 'SigmaVVConjHH [linear units]'
label7 = 'SigmaVVConjHHdelta [linear units]'
# label8 = 'Wind CMOD4 [m/s]'
import plotRS2
reload(plotRS2)
from plotRS2 import plotRS2
# import mpl_util
# reload(mpl_util)
#
# import gmtColormap
# reload(gmtColormap)
#
# import createMapsEtopo1
# reload(createMapsEtopo1)
# from createMapsEtopo1 import makeMap
#
# import g1sst
# reload(g1sst)
# from g1sst import g1sst
# plotRS2(10*log10(SigmaVVwnr), lat, lon, pixel, line, \
# scale=scale, m=m, clm=(-25,-5), label=label1)
# makeMap(ll_lon, ur_lon, \
# ll_lat, ur_lat, m, name=label1, contour='land')
# plt.close('all')
#
# plotRS2(10*log10(SigmaHHwnr), lat, lon, pixel, line, \
# scale=scale, m=m, clm=(-25,-5), label=label2)
# makeMap(ll_lon, ur_lon, \
# ll_lat, ur_lat, m, name=label2, contour='land')
# plt.close('all')
# plotRS2(PR, lat, lon, pixel, line, \
# scale=scale, m=m, clm=(-2.7,0), label=label3)
# makeMap(ll_lon, ur_lon, \
# ll_lat, ur_lat, m, name=label3, contour='land')
# plt.close('all')
# Coords to plot Figure number
x, y = m(3.15, 42.15)
# Figure 1 a - VV
plotRS2(SigmaVVwnr, lat, lon, pixel, line, \
scale=scale, m=m, clm=(0,0.05), label=label11)
plt.text(x,y,"a",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
makeMap(ll_lon, ur_lon, \
ll_lat, ur_lat, m, name=label11, contour='land')
plt.close('all')
a = label11.split(' ')[0] # split the name if it has spaces
system("rm " + "/home/mag/" + a + ".tiff")
system("mv " + "/home/mag/" + a + "_ETOPO1.tiff" + " " + "/home/mag/Figure1a.tiff")
system("convert -compress lzw /home/mag/Figure1a.tiff /home/mag/Figure1a.tiff")
# Figure 1 b - HH
plotRS2(SigmaHHwnr, lat, lon, pixel, line, \
scale=scale, m=m, clm=(0,0.05), label=label22)
plt.text(x,y,"b",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
makeMap(ll_lon, ur_lon, \
ll_lat, ur_lat, m, name=label22, contour='land')
plt.close('all')
a = label22.split(' ')[0] # split the name if it has spaces
system("rm " + "/home/mag/" + a + ".tiff")
system("mv " + "/home/mag/" + a + "_ETOPO1.tiff" + " " + "/home/mag/Figure1b.tiff")
system("convert -compress lzw /home/mag/Figure1b.tiff /home/mag/Figure1b.tiff")
# Figure 1 c - polarization ratio (PR) (always<1)
plotRS2(PR, lat, lon, pixel, line, \
scale=scale, m=m, clm=(0.4,1), label=label33)
plt.text(x,y,"c",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
makeMap(ll_lon, ur_lon, \
ll_lat, ur_lat, m, name=label3, contour='land')
plt.close('all')
a = label33.split(' ')[0] # split the name if it has spaces
system("rm " + "/home/mag/" + a + ".tiff")
system("mv " + "/home/mag/" + a + "_ETOPO1.tiff" + " " + "/home/mag/Figure1c.tiff")
system("convert -compress lzw /home/mag/Figure1c.tiff /home/mag/Figure1c.tiff")
# Figure 1 d - polarization difference (PD) (always>0)
plotRS2(delta, lat, lon, pixel, line, \
scale=scale, m=m, clm=(0,0.02), label=label4)
plt.text(x,y,"d",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
makeMap(ll_lon, ur_lon, \
ll_lat, ur_lat, m, name=label4, contour='land')
plt.close('all')
a = label4.split(' ')[0] # split the name if it has spaces
system("rm " + "/home/mag/" + a + ".tiff")
system("mv " + "/home/mag/" + a + "_ETOPO1.tiff" + " " + "/home/mag/Figure1d.tiff")
system("convert -compress lzw /home/mag/Figure1d.tiff /home/mag/Figure1d.tiff")
# Figure 4 - wave breaking contribution
plotRS2(SigmaWb, lat, lon, pixel, line, \
scale=scale, m=m, clm=(0,0.03), label=label5)
plt.text(x,y,"d",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
makeMap(ll_lon, ur_lon, \
ll_lat, ur_lat, m, name=label5, contour='land')
plt.close('all')
a = label5.split(' ')[0] # split the name if it has spaces
system("rm " + "/home/mag/" + a + ".tiff")
system("mv " + "/home/mag/" + a + "_ETOPO1.tiff" + " " + "/home/mag/Figure4.tiff")
system("convert -compress lzw /home/mag/Figure4.tiff /home/mag/Figure4.tiff")
import transect
reload(transect)
transect.plotTrns(SigmaWbtrans, pts, lat, lon, SigmaWb, pn)
# Figure 3
# Plot cropped Oil Spills
# Northern
# setup nsper basemap
# Lat/Lon coords of image corners
lat_new, lon_new = intCrpLL(lat, lon, pixel, line, scale=1, ptsCrop=ptsCropNorth)
ll_lat = lat_new.min()
ur_lat = lat_new.max()
ll_lon = lon_new.min()
ur_lon = lon_new.max()
cent_lat = lat_new.mean()
cent_lon = lon_new.mean()
m = Basemap(llcrnrlat=ll_lat, urcrnrlat=ur_lat,\
llcrnrlon=ll_lon, urcrnrlon=ur_lon, \
resolution='c', projection='nsper', \
satellite_height=798000, \
lat_0=cent_lat,lon_0=cent_lon)
# Coords to plot Figure number
xSl, ySl = m(3.473, 42.173)
# Figure 3 a - VV
plotRS2(SigmaVVwnr, lat, lon, pixel, line, \
scale=1, m=m, clm=(0,0.05), label=label11, ptsCrop=ptsCropNorth)
plt.text(xSl,ySl,"a",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
plt.savefig('/home/mag/Figure3a.tiff', facecolor='w', edgecolor='w', \
dpi=300, bbox_inches="tight", pad_inches=1.75)
plt.close('all')
# Figure 3 b - PR
plotRS2(PR, lat, lon, pixel, line, \
scale=1, m=m, clm=(0.5,1), label=label33, ptsCrop=ptsCropNorth)
plt.text(xSl,ySl,"b",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
plt.savefig('/home/mag/Figure3b.tiff', facecolor='w', edgecolor='w', \
dpi=300, bbox_inches="tight", pad_inches=1.75)
plt.close('all')
# # Figure 3 c - PD (always>0)
# plotRS2(delta, lat, lon, pixel, line, \
# scale=1, m=m, clm=(0,0.03), label=label4, ptsCrop=ptsCropNorth)
# plt.text(xSl,ySl,"c",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
# plt.savefig('/home/mag/Figure3c.tiff', facecolor='w', edgecolor='w', \
# dpi=300, bbox_inches="tight", pad_inches=1.75)
# plt.close('all')
system("convert -compress lzw /home/mag/Figure3a.tiff /home/mag/Figure3a.tiff")
system("convert -compress lzw /home/mag/Figure3b.tiff /home/mag/Figure3b.tiff")
# system("convert -compress lzw /home/mag/Figure3c.tiff /home/mag/Figure3c.tiff")
# # Southern
# # setup nsper basemap
# # Lat/Lon coords of image corners
# lat_new, lon_new = intCrpLL(lat, lon, pixel, line, scale=1, ptsCrop=ptsCropSouth)
# ll_lat = lat_new.min()
# ur_lat = lat_new.max()
# ll_lon = lon_new.min()
# ur_lon = lon_new.max()
# cent_lat = lat_new.mean()
# cent_lon = lon_new.mean()
# m = Basemap(llcrnrlat=ll_lat, urcrnrlat=ur_lat,\
# llcrnrlon=ll_lon, urcrnrlon=ur_lon, \
# resolution='c', projection='nsper', \
# satellite_height=798000, \
# lat_0=cent_lat,lon_0=cent_lon)
# # Coords to plot Figure number
# xSl, ySl = m(3.491, 41.973)
# # Figure 3 d - VV
# plotRS2(SigmaVVwnr, lat, lon, pixel, line, \
# scale=1, m=m, clm=(0,0.05), label=label11, ptsCrop=ptsCropSouth)
# plt.text(xSl,ySl,"d",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
# plt.savefig('/home/mag/Figure3d.tiff', facecolor='w', edgecolor='w', \
# dpi=300, bbox_inches="tight", pad_inches=1.75)
# plt.close('all')
# # Figure 3 e - PR (always<1)
# plotRS2(PR, lat, lon, pixel, line, \
# scale=1, m=m, clm=(0.4,1), label=label33, ptsCrop=ptsCropSouth)
# plt.text(xSl,ySl,"e",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
# plt.savefig('/home/mag/Figure3e.tiff', facecolor='w', edgecolor='w', \
# dpi=300, bbox_inches="tight", pad_inches=1.75)
# plt.close('all')
# # Figure 3 f - PD (always>0)
# plotRS2(delta, lat, lon, pixel, line, \
# scale=1, m=m, clm=(0,0.03), label=label4, ptsCrop=ptsCropSouth)
# plt.text(xSl,ySl,"f",bbox=dict(facecolor='w', alpha=0.5),stretch='expanded',fontsize=27)
# plt.savefig('/home/mag/Figure3f.tiff', facecolor='w', edgecolor='w', \
# dpi=300, bbox_inches="tight", pad_inches=1.75)
# plt.close('all')
# system("convert -compress lzw /home/mag/Figure3d.tiff /home/mag/Figure3d.tiff")
# system("convert -compress lzw /home/mag/Figure3e.tiff /home/mag/Figure3e.tiff")
# system("convert -compress lzw /home/mag/Figure3f.tiff /home/mag/Figure3f.tiff")
import transect
reload(transect)
transect.plotTrns(PRtrans, pts, lat, lon, PR, pixel, line, pn='/home/mag/', \
label='PR [linear units]', clm=(0.5,1), \
ptsCrop=ptsCropNorth, m=m, scale=1)
# create Nansat object
#~ n = Nansat(iPath + fileName, mapperName="modis_l1", logLevel=10)
n = Nansat(iPath + fileName, mapperName="modis_l1")
# list bands and georeference of the object
print n
# Reprojected image into Lat/Lon WGS84 (Simple Cylindrical) projection
# 1. Cancel previous reprojection
# 2. Get corners of the image and the pixel resolution
# 3. Create Domain with stereographic projection, corner coordinates and resolution 1000m
# 4. Reproject
# 5. Write image
n.reproject() # 1.
lons, lats = n.get_corners() # 2.
pxlRes = distancelib.getPixelResolution(array(lats), array(lons), n.shape(), units="deg")
srsString = "+proj=latlong +datum=WGS84 +ellps=WGS84 +no_defs"
#~ extentString = '-lle %f %f %f %f -ts 3000 3000' % (min(lons), min(lats), max(lons), max(lats))
extentString = '-lle %f %f %f %f -tr %f %f' % (min(lons), min(lats), \
max(lons), max(lats), pxlRes[1], pxlRes[0])
d = Domain(srs=srsString, ext=extentString) # 3.
print d
n.reproject(d) # 4.
# get array with watermask (landmask) b
# it must be done after reprojection!
# 1. Get Nansat object with watermask
# 2. Get array from Nansat object. 0 - land, 1 - water
#wm = n.watermask(mod44path='/media/magDesk/media/SOLabNFS/store/auxdata/coastline/mod44w/')
wm = n.watermask(mod44path='/media/data/data/auxdata/coastline/mod44w/')
wmArray = wm[1]
sigma0 = raw_counts**2.0 * sin(deg2rad(inc_angle))
sigma0 = 10*log10(sigma0)
n.add_band(bandID=4, array=sigma0)
# 1. Remove speckle noise (using Lee-Wiener filter)
speckle_filter('wiener', 7)
# Reprojected image into Lat/Lon WGS84 (Simple Cylindrical) projection
# 1. Cancel previous reprojection
# 2. Get corners of the image and the pixel resolution
# 3. Create Domain with stereographic projection, corner coordinates and resolution 1000m
# 4. Reproject
# 5. Write image
n.reproject() # 1.
lons, lats = n.get_corners() # 2.
pxlRes = distancelib.getPixelResolution(array(lats), array(lons), n.shape(), units="deg")
srsString = "+proj=latlong +datum=WGS84 +ellps=WGS84 +no_defs"
#~ extentString = '-lle %f %f %f %f -ts 3000 3000' % (min(lons), min(lats), max(lons), max(lats))
extentString = '-lle %f %f %f %f -tr %f %f' % (min(lons), min(lats), \
max(lons), max(lats), pxlRes[1], pxlRes[0])
d = Domain(srs=srsString, ext=extentString) # 3.
n.reproject(d) # 4.
# get array with watermask (landmask) b
# it must be done after reprojection!
# 1. Get Nansat object with watermask
# 2. Get array from Nansat object. 0 - land, 1 - water
#wm = n.watermask(mod44path='/media/magDesk/media/SOLabNFS/store/auxdata/coastline/mod44w/')
wm = n.watermask(mod44path='/media/data/data/auxdata/coastline/mod44w/')
wmArray = wm[1]
def main(argv=None):
year = '2012'
useMask = False
if argv is None:
argv = sys.argv
if argv is None:
print("Please specify the path/year to the asar folder! \n")
return
# Parse arguments
try:
opts, args = getopt.getopt(argv, "hi:o:",
["year=", "oPath=", "iPath=", "useMask="])
except getopt.GetoptError:
print 'readASAR.py -year <year> ...'
sys.exit(2)
for opt, arg in opts:
if opt == '-h':
print 'readASAR.py -year <year> ...'
sys.exit()
elif opt in ("-year", "--year"):
year = arg
elif opt in ("-oPath", "--oPath"):
oPath = arg
elif opt in ("-iPath", "--iPath"):
iPath = arg
elif opt in ("-useMask", "--useMask"):
useMask = arg
oPath = '/media/SOLabNFS2/tmp/roughness/' + year + '/'
iPath = '/media/SOLabNFS2/store/satellite/asar/' + year + '/'
if not os.path.exists(oPath):
os.makedirs(oPath)
dirNames = os.listdir(iPath)
for dirName in dirNames:
fileNames = os.listdir(iPath + dirName)
for fileName in fileNames:
figureName = oPath + fileName[0:27] + '/' + fileName + '_proj.png'
kmlName = oPath + fileName[0:27] + '/' + fileName + '.kml'
if not os.path.exists(oPath + fileName[0:27] + '/'):
os.makedirs(oPath + fileName[0:27] + '/')
if os.path.isfile(kmlName):
print "%s already processed" % (fileName)
continue
else:
print "%s" % (fileName)
# try to create Nansat object
try:
n = Nansat(iPath + dirName + '/' + fileName,
mapperName='asar',
logLevel=27)
except Exception as e:
print "Failed to create Nansat object:"
print str(e)
os.rmdir(oPath + fileName[0:27] + '/')
continue
#~ Get the bands
raw_counts = n[1]
inc_angle = n[2]
#~ NICE image (roughness)
pol = n.bands()[3]['polarization']
if pol == 'HH':
ph = (2.20495, -14.3561e-2, 11.28e-4)
sigma0_hh_ref = exp(
(ph[0] + inc_angle * ph[1] + inc_angle**2 * ph[2]) *
log(10))
roughness = n[3] / sigma0_hh_ref
elif pol == 'VV':
pv = (2.29373, -15.393e-2, 15.1762e-4)
sigma0_vv_ref = exp(
(pv[0] + inc_angle * pv[1] + inc_angle**2 * pv[2]) *
log(10))
roughness = n[3] / sigma0_vv_ref
#~ Create new band
n.add_band(bandID=4, array=roughness, \
parameters={'name':'roughness', \
'wkv': 'surface_backwards_scattering_coefficient_of_radar_wave', \
'dataType': 6})
# Reproject image into Lat/Lon WGS84 (Simple Cylindrical) projection
# 1. Cancel previous reprojection
# 2. Get corners of the image and the pixel resolution
# 3. Create Domain with stereographic projection, corner coordinates 1000m
# 4. Reproject
# 5. Write image
n.reproject() # 1.
lons, lats = n.get_corners() # 2.
# Pixel resolution
#~ pxlRes = distancelib.getPixelResolution(array(lats), array(lons), n.shape())
#~ pxlRes = array(pxlRes)*360/40000 # great circle distance
pxlRes = array(
distancelib.getPixelResolution(array(lats), array(lons),
n.shape(), 'deg'))
ipdb.set_trace()
if min(lats) >= 65 and max(
lats) >= 75 and max(lats) - min(lats) >= 13:
pxlRes = array([0.00065, 0.00065
]) * 2 # make the resolution 150x150m
#~ pxlRes = pxlRes*7 # make the resolution worser
srsString = "+proj=latlong +datum=WGS84 +ellps=WGS84 +no_defs"
#~ extentString = '-lle %f %f %f %f -ts 3000 3000' % (min(lons), min(lats), max(lons), max(lats))
extentString = '-lle %f %f %f %f -tr %f %f' % (min(lons), min(lats), \
max(lons), max(lats), pxlRes[1], pxlRes[0])
d = Domain(srs=srsString, ext=extentString) # 3.
n.reproject(d) # 4.
if useMask:
# get array with watermask (landmask) b
# it must be done after reprojection!
# 1. Get Nansat object with watermask
# 2. Get array from Nansat object. 0 - land, 1 - water
#wm = n.watermask(mod44path='/media/magDesk/media/SOLabNFS/store/auxdata/coastline/mod44w/')
wm = n.watermask(
mod44path='/media/data/data/auxdata/coastline/mod44w/')
wmArray = wm[1]
#~ ОШИБКА numOfColor=255 не маскирует, потому что в figure.apply_mask: availIndeces = range(self.d['numOfColor'], 255 - 1)
#~ n.write_figure(fileName=figureName, bands=[3], \
#~ numOfColor=255, mask_array=wmArray, mask_lut={0: 0},
#~ clim=[0,0.15], cmapName='gray', transparency=0) # 5.
n.write_figure(fileName=figureName, bands=[4], \
mask_array=wmArray, mask_lut={0: [0,0,0]},
clim=[0,2], cmapName='gray', transparency=[0,0,0]) # 5.
else:
n.write_figure(fileName=figureName, bands=[1], \
clim=[0,2], cmapName='gray', transparency=[0,0,0]) # 5.
# open the input image and convert to RGBA for further tiling with slbtiles
input_img = Image.open(figureName)
output_img = input_img.convert("RGBA")
output_img.save(figureName)
# make KML image
n.write_kml_image(kmlFileName=kmlName, kmlFigureName=figureName)
#~ Change the file permissions
os.chmod(oPath, 0777)
os.chmod(oPath + fileName[0:27] + '/', 0777)
os.chmod(kmlName, 0777)
os.chmod(figureName, 0777)
#~ Change the owner and group
#~ os.chown(oPath, 1111, 1111)
#~ os.chown(oPath + fileName[0:27] + '/', 1111, 1111)
#~ os.chown(kmlName, 1111, 1111)
#~ os.chown(figureName, 1111, 1111)
#~ garbage collection
gc.collect()