def main():
usage = '''
Usage:
---------------------------------------------------------
python %s [-p "bandPositions"] [-d "spatialDimensions"]
[-K number of clusters] [-M max scale][-m min scale]
[-t initial annealing temperature] [-s spatial mixing factor]
[-P generate class probabilities image] filename
bandPositions and spatialDimensions are lists,
e.g., -p [1,2,4] -d [0,0,400,400]
If the input file is named
path/filenbasename.ext then
The output classification file is named
path/filebasename_em.ext
and the class probabilities output file is named
path/filebasename_emprobs.ext
--------------------------------------------------------''' %sys.argv[0]
options, args = getopt.getopt(sys.argv[1:],'hp:d:K:M:m:t:s:P')
pos = None
dims = None
K,max_scale,min_scale,T0,beta,probs = (None,None,None,None,None,None)
for option, value in options:
if option == '-h':
print usage
return
elif option == '-p':
pos = eval(value)
elif option == '-d':
dims = eval(value)
elif option == '-K':
K = eval(value)
elif option == '-M':
max_scale = eval(value)
elif option == '-m':
min_scale = eval(value)
elif option == '-t':
T0 = eval(value)
elif option == '-s':
beta = eval(value)
elif option == '-P':
probs = True
if len(args) != 1:
print 'Incorrect number of arguments'
print usage
sys.exit(1)
if K is None:
K = 6
if max_scale is None:
max_scale = 2
else:
max_scale = min((max_scale,3))
if min_scale is None:
min_scale = 0
else:
min_scale = min((max_scale,min_scale))
if T0 is None:
T0 = 0.5
if beta is None:
beta = 0.5
if probs is None:
probs = False
gdal.AllRegister()
infile = args[0]
gdal.AllRegister()
try:
inDataset = gdal.Open(infile,GA_ReadOnly)
cols = inDataset.RasterXSize
rows = inDataset.RasterYSize
bands = inDataset.RasterCount
except Exception as e:
print 'Error: %s --Image could not be read'%e
sys.exit(1)
if pos is not None:
bands = len(pos)
else:
pos = range(1,bands+1)
if dims:
x0,y0,cols,rows = dims
else:
x0 = 0
y0 = 0
class_image = np.zeros((rows,cols),dtype=np.byte)
path = os.path.dirname(infile)
basename = os.path.basename(infile)
root, ext = os.path.splitext(basename)
outfile = path+'/'+root+'_em'+ext
if probs:
probfile = path+'/'+root+'_emprobs'+ext
print '--------------------------'
print ' EM clustering'
print '--------------------------'
print 'infile: %s'%infile
print 'clusters: %i'%K
print 'T0: %f'%T0
print 'beta: %f'%beta
start = time.time()
# read in image and compress
path = os.path.dirname(infile)
basename = os.path.basename(infile)
root, ext = os.path.splitext(basename)
DWTbands = []
for b in pos:
band = inDataset.GetRasterBand(b)
DWTband = auxil.DWTArray(band.ReadAsArray(x0,y0,cols,rows).astype(float),cols,rows)
for i in range(max_scale):
DWTband.filter()
DWTbands.append(DWTband)
rows,cols = DWTbands[0].get_quadrant(0).shape
G = np.transpose(np.array([DWTbands[i].get_quadrant(0,float=True).ravel() for i in range(bands)]))
# initialize membership matrix
n = G.shape[0]
U = np.random.random((K,n))
den = np.sum(U,axis=0)
for j in range(K):
U[j,:] = U[j,:]/den
# cluster at minimum scale
try:
U,Ms,Cs,Ps,pdens = em(G,U,T0,beta,rows,cols)
except:
print 'em failed'
return
# sort clusters wrt partition density
idx = np.argsort(pdens)
idx = idx[::-1]
U = U[idx,:]
# clustering at increasing scales
for i in range(max_scale-min_scale):
# expand U and renormalize
U = np.reshape(U,(K,rows,cols))
rows = rows*2
cols = cols*2
U = ndi.zoom(U,(1,2,2))
U = np.reshape(U,(K,rows*cols))
idx = np.where(U<0.0)
U[idx] = 0.0
den = np.sum(U,axis=0)
for j in range(K):
U[j,:] = U[j,:]/den
# expand the image
for i in range(bands):
DWTbands[i].invert()
G = np.transpose(np.array([DWTbands[i].get_quadrant(0,float=True).ravel() for i in range(bands)]))
# cluster
unfrozen = np.where(np.max(U,axis=0) < 0.90)
try:
U,Ms,Cs,Ps,pdens = em(G,U,0.0,beta,rows,cols,unfrozen=unfrozen)
except:
print 'em failed'
return
print 'Cluster mean vectors'
print Ms
print 'Cluster covariance matrices'
for k in range(K):
print 'cluster: %i'%k
print Cs[k]
# up-sample class memberships if necessary
if min_scale>0:
U = np.reshape(U,(K,rows,cols))
f = 2**min_scale
rows = rows*f
cols = cols*f
U = ndi.zoom(U,(1,f,f))
U = np.reshape(U,(K,rows*cols))
idx = np.where(U<0.0)
U[idx] = 0.0
den = np.sum(U,axis=0)
for j in range(K):
U[j,:] = U[j,:]/den
# classify
labels = np.byte(np.argmax(U,axis=0)+1)
class_image[0:rows,0:cols] = np.reshape(labels,(rows,cols))
rows1,cols1 = class_image.shape
# write to disk
driver = inDataset.GetDriver()
outDataset = driver.Create(outfile,cols1,rows1,1,GDT_Byte)
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
gt = list(geotransform)
gt[0] = gt[0] + x0*gt[1]
gt[3] = gt[3] + y0*gt[5]
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
outBand = outDataset.GetRasterBand(1)
outBand.WriteArray(class_image,0,0)
outBand.FlushCache()
outDataset = None
# write class membership probability file if desired
if probs:
outDataset = driver.Create(probfile,cols,rows,K,GDT_Byte)
if geotransform is not None:
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
for k in range(K):
probs = np.reshape(U[k,:],(rows,cols))
probs = np.byte(probs*255)
outBand = outDataset.GetRasterBand(k+1)
outBand.WriteArray(probs,0,0)
outBand.FlushCache()
outDataset = None
print 'class probabilities written to: %s'%probfile
inDataset = None
if (ext == '') and (K<19):
# try to make an ENVI classification header file
hdr = header.Header()
headerfile = outfile+'.hdr'
f = open(headerfile)
line = f.readline()
envihdr = ''
while line:
envihdr += line
line = f.readline()
f.close()
hdr.read(envihdr)
hdr['file type'] ='ENVI Classification'
hdr['classes'] = str(K+1)
classlookup = '{0'
for i in range(1,3*(K+1)):
classlookup += ', '+str(str(ctable[i]))
classlookup +='}'
hdr['class lookup'] = classlookup
hdr['class names'] = ['class %i'%i for i in range(K+1)]
f = open(headerfile,'w')
f.write(str(hdr))
f.close()
print 'classified image written to: '+outfile
print 'elapsed time: '+str(time.time()-start)
print '--done------------------------'
def main():
gdal.AllRegister()
path = auxil.select_directory('Choose working directory')
if path:
os.chdir(path)
infile = auxil.select_infile(title='Select an image')
if infile:
inDataset = gdal.Open(infile, GA_ReadOnly)
cols = inDataset.RasterXSize
rows = inDataset.RasterYSize
bands = inDataset.RasterCount
else:
return
pos = auxil.select_pos(bands)
if not pos:
return
dims = auxil.select_dims([0, 0, cols, rows])
if dims:
x0, y0, cols, rows = dims
else:
return
m = auxil.select_integer(1000, 'Select training sample size')
K = auxil.select_integer(6, 'Select number of clusters')
outfile, outfmt = auxil.select_outfilefmt()
if not outfile:
return
kernel = auxil.select_integer(1, 'Select kernel: 0=linear, 1=Gaussian')
print '========================='
print ' kkmeans'
print '========================='
print 'infile: ' + infile
print 'samples: ' + str(m)
if kernel == 0:
print 'kernel: ' + 'linear'
else:
print 'kernel: ' + 'Gaussian'
start = time.time()
# input data matrix
XX = np.zeros((cols * rows, bands))
k = 0
for b in pos:
band = inDataset.GetRasterBand(b)
band = band.ReadAsArray(x0, y0, cols, rows).astype(float)
XX[:, k] = np.ravel(band)
k += 1
# training data matrix
idx = np.fix(np.random.random(m) * (cols * rows)).astype(np.integer)
X = XX[idx, :]
print 'kernel matrix...'
# uncentered kernel matrix
KK, gma = auxil.kernelMatrix(X, kernel=kernel)
if gma is not None:
print 'gamma: ' + str(round(gma, 6))
# initial (random) class labels
labels = np.random.randint(K, size=m)
# iteration
change = True
itr = 0
onesm = np.mat(np.ones(m, dtype=float))
while change and (itr < 100):
change = False
U = np.zeros((K, m))
for i in range(m):
U[labels[i], i] = 1
M = np.diag(1.0 / (np.sum(U, axis=1) + 1.0))
MU = np.mat(np.dot(M, U))
Z = (onesm.T) * np.diag(MU * KK * (MU.T)) - 2 * KK * (MU.T)
Z = np.array(Z)
labels1 = (np.argmin(Z, axis=1) % K).ravel()
if np.sum(labels1 != labels):
change = True
labels = labels1
itr += 1
print 'iterations: %i' % itr
# classify image
print 'classifying...'
i = 0
A = np.diag(MU * KK * (MU.T))
A = np.tile(A, (cols, 1))
class_image = np.zeros((rows, cols), dtype=np.byte)
while i < rows:
XXi = XX[i * cols:(i + 1) * cols, :]
KKK, _ = auxil.kernelMatrix(X, XXi, gma=gma, kernel=kernel)
Z = A - 2 * (KKK.T) * (MU.T)
Z = np.array(Z)
labels = np.argmin(Z, axis=1).ravel()
class_image[i, :] = (labels % K) + 1
i += 1
sys.stdout.write("\n")
# write to disk
driver = gdal.GetDriverByName(outfmt)
outDataset = driver.Create(outfile, cols, rows, 1, GDT_Byte)
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
gt = list(geotransform)
gt[0] = gt[0] + x0 * gt[1]
gt[3] = gt[3] + y0 * gt[5]
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
outBand = outDataset.GetRasterBand(1)
outBand.WriteArray(class_image, 0, 0)
outBand.FlushCache()
outDataset = None
inDataset = None
if (outfmt == 'ENVI') and (K < 19):
# try to make an ENVI classification header file
hdr = header.Header()
headerfile = outfile + '.hdr'
f = open(headerfile)
line = f.readline()
envihdr = ''
while line:
envihdr += line
line = f.readline()
f.close()
hdr.read(envihdr)
hdr['file type'] = 'ENVI Classification'
hdr['classes'] = str(K)
classlookup = '{0'
for i in range(1, 3 * K):
classlookup += ', ' + str(str(ctable[i]))
classlookup += '}'
hdr['class lookup'] = classlookup
hdr['class names'] = [str(i + 1) for i in range(K)]
f = open(headerfile, 'w')
f.write(str(hdr))
f.close()
print 'result written to: ' + outfile
print 'elapsed time: ' + str(time.time() - start)
print '--done------------------------'
def main():
gdal.AllRegister()
path = auxil.select_directory('Input directory')
if path:
os.chdir(path)
# input image
infile = auxil.select_infile(title='Image file')
if infile:
inDataset = gdal.Open(infile, GA_ReadOnly)
cols = inDataset.RasterXSize
rows = inDataset.RasterYSize
bands = inDataset.RasterCount
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
gt = list(geotransform)
else:
print 'No geotransform available'
return
imsr = osr.SpatialReference()
imsr.ImportFromWkt(projection)
else:
return
pos = auxil.select_pos(bands)
if not pos:
return
N = len(pos)
rasterBands = []
for b in pos:
rasterBands.append(inDataset.GetRasterBand(b))
# training algorithm
trainalg = auxil.select_integer(1,
msg='1:Maxlike,2:Backprop,3:Congrad,4:SVM')
if not trainalg:
return
# training data (shapefile)
trnfile = auxil.select_infile(filt='.shp', title='Train shapefile')
if trnfile:
trnDriver = ogr.GetDriverByName('ESRI Shapefile')
trnDatasource = trnDriver.Open(trnfile, 0)
trnLayer = trnDatasource.GetLayer()
trnsr = trnLayer.GetSpatialRef()
else:
return
tstfile = auxil.select_outfile(filt='.tst', title='Test results file')
if not tstfile:
print 'No test output'
# outfile
outfile, outfmt = auxil.select_outfilefmt(title='Classification file')
if not outfile:
return
if trainalg in (2, 3, 4):
# class probabilities file, hidden neurons
probfile, probfmt = auxil.select_outfilefmt(title='Probabilities file')
else:
probfile = None
if trainalg in (2, 3):
L = auxil.select_integer(8, 'Number of hidden neurons')
if not L:
return
# coordinate transformation from training to image projection
ct = osr.CoordinateTransformation(trnsr, imsr)
# number of classes
K = 1
feature = trnLayer.GetNextFeature()
while feature:
classid = feature.GetField('CLASS_ID')
if int(classid) > K:
K = int(classid)
feature = trnLayer.GetNextFeature()
trnLayer.ResetReading()
K += 1
print '========================='
print 'supervised classification'
print '========================='
print time.asctime()
print 'image: ' + infile
print 'training: ' + trnfile
if trainalg == 1:
print 'Maximum Likelihood'
elif trainalg == 2:
print 'Neural Net (Backprop)'
elif trainalg == 3:
print 'Neural Net (Congrad)'
else:
print 'Support Vector Machine'
# loop through the polygons
Gs = [] # train observations
ls = [] # class labels
classnames = '{unclassified'
classids = set()
print 'reading training data...'
for i in range(trnLayer.GetFeatureCount()):
feature = trnLayer.GetFeature(i)
classid = str(feature.GetField('CLASS_ID'))
classname = feature.GetField('CLASS_NAME')
if classid not in classids:
classnames += ', ' + classname
classids = classids | set(classid)
l = [0 for i in range(K)]
l[int(classid)] = 1.0
polygon = feature.GetGeometryRef()
# transform to same projection as image
polygon.Transform(ct)
# convert to a Shapely object
poly = shapely.wkt.loads(polygon.ExportToWkt())
# transform the boundary to pixel coords in numpy
bdry = np.array(poly.boundary)
bdry[:, 0] = bdry[:, 0] - gt[0]
bdry[:, 1] = bdry[:, 1] - gt[3]
GT = np.mat([[gt[1], gt[2]], [gt[4], gt[5]]])
bdry = bdry * np.linalg.inv(GT)
# polygon in pixel coords
polygon1 = asPolygon(bdry)
# raster over the bounding rectangle
minx, miny, maxx, maxy = map(int, list(polygon1.bounds))
pts = []
for i in range(minx, maxx + 1):
for j in range(miny, maxy + 1):
pts.append((i, j))
multipt = MultiPoint(pts)
# intersection as list
intersection = np.array(multipt.intersection(polygon1),
dtype=np.int).tolist()
# cut out the bounded image cube
cube = np.zeros((maxy - miny + 1, maxx - minx + 1, len(rasterBands)))
k = 0
for band in rasterBands:
cube[:, :, k] = band.ReadAsArray(minx, miny, maxx - minx + 1,
maxy - miny + 1)
k += 1
# get the training vectors
for (x, y) in intersection:
Gs.append(cube[y - miny, x - minx, :])
ls.append(l)
polygon = None
polygon1 = None
feature.Destroy()
trnDatasource.Destroy()
classnames += '}'
m = len(ls)
print str(m) + ' training pixel vectors were read in'
Gs = np.array(Gs)
ls = np.array(ls)
# stretch the pixel vectors to [-1,1] for ffn
maxx = np.max(Gs, 0)
minx = np.min(Gs, 0)
for j in range(N):
Gs[:, j] = 2 * (Gs[:, j] - minx[j]) / (maxx[j] - minx[j]) - 1.0
# random permutation of training data
idx = np.random.permutation(m)
Gs = Gs[idx, :]
ls = ls[idx, :]
# setup output datasets
driver = gdal.GetDriverByName(outfmt)
outDataset = driver.Create(outfile, cols, rows, 1, GDT_Byte)
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
outBand = outDataset.GetRasterBand(1)
if probfile:
driver = gdal.GetDriverByName(probfmt)
probDataset = driver.Create(probfile, cols, rows, K, GDT_Byte)
if geotransform is not None:
probDataset.SetGeoTransform(tuple(gt))
if projection is not None:
probDataset.SetProjection(projection)
probBands = []
for k in range(K):
probBands.append(probDataset.GetRasterBand(k + 1))
if tstfile:
# train on 2/3 training examples
Gstrn = Gs[0:2 * m // 3, :]
lstrn = ls[0:2 * m // 3, :]
Gstst = Gs[2 * m // 3:, :]
lstst = ls[2 * m // 3:, :]
else:
Gstrn = Gs
lstrn = ls
if trainalg == 1:
classifier = sc.Maxlike(Gstrn, lstrn)
elif trainalg == 2:
classifier = sc.Ffnbp(Gstrn, lstrn, L)
elif trainalg == 3:
classifier = sc.Ffncg(Gstrn, lstrn, L)
elif trainalg == 4:
classifier = sc.Svm(Gstrn, lstrn)
print 'training on %i pixel vectors...' % np.shape(Gstrn)[0]
start = time.time()
result = classifier.train()
print 'elapsed time %s' % str(time.time() - start)
if result:
if trainalg in [2, 3]:
cost = np.log10(result)
ymax = np.max(cost)
ymin = np.min(cost)
xmax = len(cost)
plt.plot(range(xmax), cost, 'k')
plt.axis([0, xmax, ymin - 1, ymax])
plt.title('Log(Cross entropy)')
plt.xlabel('Epoch')
# classify the image
print 'classifying...'
start = time.time()
tile = np.zeros((cols, N))
for row in range(rows):
for j in range(N):
tile[:, j] = rasterBands[j].ReadAsArray(0, row, cols, 1)
tile[:, j] = 2 * (tile[:, j] - minx[j]) / (maxx[j] -
minx[j]) - 1.0
cls, Ms = classifier.classify(tile)
outBand.WriteArray(np.reshape(cls, (1, cols)), 0, row)
if probfile:
Ms = np.byte(Ms * 255)
for k in range(K):
probBands[k].WriteArray(np.reshape(Ms[k, :], (1, cols)), 0,
row)
outBand.FlushCache()
print 'elapsed time %s' % str(time.time() - start)
outDataset = None
inDataset = None
if probfile:
for probBand in probBands:
probBand.FlushCache()
probDataset = None
print 'class probabilities written to: %s' % probfile
K = lstrn.shape[1] + 1
if (outfmt == 'ENVI') and (K < 19):
# try to make an ENVI classification header file
hdr = header.Header()
headerfile = outfile + '.hdr'
f = open(headerfile)
line = f.readline()
envihdr = ''
while line:
envihdr += line
line = f.readline()
f.close()
hdr.read(envihdr)
hdr['file type'] = 'ENVI Classification'
hdr['classes'] = str(K)
classlookup = '{0'
for i in range(1, 3 * K):
classlookup += ', ' + str(str(ctable[i]))
classlookup += '}'
hdr['class lookup'] = classlookup
hdr['class names'] = classnames
f = open(headerfile, 'w')
f.write(str(hdr))
f.close()
print 'thematic map written to: %s' % outfile
if trainalg in [2, 3]:
print 'please close the cross entropy plot to continue'
plt.show()
if tstfile:
with open(tstfile, 'w') as f:
print >> f, 'FFN test results for %s' % infile
print >> f, time.asctime()
print >> f, 'Classification image: %s' % outfile
print >> f, 'Class probabilities image: %s' % probfile
print >> f, lstst.shape[0], lstst.shape[1]
classes, _ = classifier.classify(Gstst)
labels = np.argmax(lstst, axis=1) + 1
for i in range(len(classes)):
print >> f, classes[i], labels[i]
f.close()
print 'test results written to: %s' % tstfile
print 'done'
else:
print 'an error occured'
return
def main():
gdal.AllRegister()
path = auxil.select_directory('Choose working directory')
# path = 'd:\\imagery\\CRC\\Chapters6-7'
if path:
os.chdir(path)
infile = auxil.select_infile(title='Select a class probability image')
if infile:
inDataset = gdal.Open(infile, GA_ReadOnly)
cols = inDataset.RasterXSize
rows = inDataset.RasterYSize
K = inDataset.RasterCount
else:
return
outfile, fmt = auxil.select_outfilefmt()
if not outfile:
return
print '========================='
print ' PLR_reclass'
print '========================='
print 'infile: %s' % infile
start = time.time()
prob_image = np.zeros((K, rows, cols))
for k in range(K):
band = inDataset.GetRasterBand(k + 1)
prob_image[k, :, :] = band.ReadAsArray(0, 0, cols, rows).astype(float)
class_image = np.zeros((rows, cols), dtype=np.byte)
print 'reclassifying...'
for i in range(rows):
if i % 50 == 0:
print '%i rows processed' % i
for j in range(cols):
cls = np.where(prob_image[:, i, j] == np.amax(prob_image[:, i,
j]))[0][0]
if isinstance(cls, int):
class_image[i, j] = cls + 1
# write to disk
driver = gdal.GetDriverByName(fmt)
outDataset = driver.Create(outfile, cols, rows, 1, GDT_Byte)
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
outDataset.SetGeoTransform(geotransform)
if projection is not None:
outDataset.SetProjection(projection)
outBand = outDataset.GetRasterBand(1)
outBand.WriteArray(class_image, 0, 0)
outBand.FlushCache()
outDataset = None
inDataset = None
if (fmt == 'ENVI') and (K < 19):
# try to make an ENVI classification header file
classnames = '{unclassified '
for i in range(K):
classnames += ', ' + str(i + 1)
classnames += '}'
hdr = header.Header()
headerfile = outfile + '.hdr'
f = open(headerfile)
line = f.readline()
envihdr = ''
while line:
envihdr += line
line = f.readline()
f.close()
hdr.read(envihdr)
hdr['file type'] = 'ENVI Classification'
hdr['classes'] = str(K + 1)
classlookup = '{0'
for i in range(1, 3 * (K + 1)):
classlookup += ', ' + str(str(auxil.ctable[i]))
classlookup += '}'
hdr['class lookup'] = classlookup
hdr['class names'] = classnames
f = open(headerfile, 'w')
f.write(str(hdr))
f.close()
print 'result written to: ' + outfile
print 'elapsed time: ' + str(time.time() - start)
print '--done------------------------'
def main():
gdal.AllRegister()
path = auxil.select_directory('Choose working directory')
if path:
os.chdir(path)
infile = auxil.select_infile(title='Select an image')
if infile:
inDataset = gdal.Open(infile, GA_ReadOnly)
cols = inDataset.RasterXSize
rows = inDataset.RasterYSize
bands = inDataset.RasterCount
else:
return
pos = auxil.select_pos(bands)
if not pos:
return
bands = len(pos)
dims = auxil.select_dims([0, 0, cols, rows])
if dims:
x0, y0, cols, rows = dims
else:
return
class_image = np.zeros((rows, cols), dtype=np.byte)
K = auxil.select_integer(6, 'Number of clusters')
max_scale = auxil.select_integer(2, 'Maximum scaling factor')
max_scale = min((max_scale, 3))
min_scale = auxil.select_integer(0, 'Minimum scaling factor')
min_scale = min((max_scale, min_scale))
T0 = auxil.select_float(0.5, 'Initial annealing temperature')
beta = auxil.select_float(0.5, 'Spatial mixing parameter')
outfile, outfmt = auxil.select_outfilefmt(
'Select output classification file')
if not outfile:
return
probfile, probfmt = auxil.select_outfilefmt(
'Select output probability file (optional)')
print '========================='
print ' EM clustering'
print '========================='
print 'infile: %s' % infile
print 'clusters: %i' % K
print 'T0: %f' % T0
print 'beta: %f' % beta
start = time.time()
# read in image and compress
DWTbands = []
for b in pos:
band = inDataset.GetRasterBand(b)
DWTband = auxil.DWTArray(
band.ReadAsArray(x0, y0, cols, rows).astype(float), cols, rows)
for i in range(max_scale):
DWTband.filter()
DWTbands.append(DWTband)
rows, cols = DWTbands[0].get_quadrant(0).shape
G = np.transpose(
np.array([
DWTbands[i].get_quadrant(0, float=True).ravel()
for i in range(bands)
]))
# initialize membership matrix
n = G.shape[0]
U = np.random.random((K, n))
den = np.sum(U, axis=0)
for j in range(K):
U[j, :] = U[j, :] / den
# cluster at minimum scale
try:
U, Ms, Cs, Ps, pdens = em(G, U, T0, beta, rows, cols)
except:
print 'em failed'
return
# sort clusters wrt partition density
idx = np.argsort(pdens)
idx = idx[::-1]
U = U[idx, :]
# clustering at increasing scales
for i in range(max_scale - min_scale):
# expand U and renormalize
U = np.reshape(U, (K, rows, cols))
rows = rows * 2
cols = cols * 2
U = ndi.zoom(U, (1, 2, 2))
U = np.reshape(U, (K, rows * cols))
idx = np.where(U < 0.0)
U[idx] = 0.0
den = np.sum(U, axis=0)
for j in range(K):
U[j, :] = U[j, :] / den
# expand the image
for i in range(bands):
DWTbands[i].invert()
G = np.transpose(
np.array([
DWTbands[i].get_quadrant(0, float=True).ravel()
for i in range(bands)
]))
# cluster
unfrozen = np.where(np.max(U, axis=0) < 0.90)
try:
U, Ms, Cs, Ps, pdens = em(G,
U,
0.0,
beta,
rows,
cols,
unfrozen=unfrozen)
except:
print 'em failed'
return
print 'Cluster mean vectors'
print Ms
print 'Cluster covariance matrices'
for k in range(K):
print 'cluster: %i' % k
print Cs[k]
# up-sample class memberships if necessary
if min_scale > 0:
U = np.reshape(U, (K, rows, cols))
f = 2**min_scale
rows = rows * f
cols = cols * f
U = ndi.zoom(U, (1, f, f))
U = np.reshape(U, (K, rows * cols))
idx = np.where(U < 0.0)
U[idx] = 0.0
den = np.sum(U, axis=0)
for j in range(K):
U[j, :] = U[j, :] / den
# classify
labels = np.byte(np.argmax(U, axis=0) + 1)
class_image[0:rows, 0:cols] = np.reshape(labels, (rows, cols))
rows1, cols1 = class_image.shape
# write to disk
driver = gdal.GetDriverByName(outfmt)
outDataset = driver.Create(outfile, cols1, rows1, 1, GDT_Byte)
projection = inDataset.GetProjection()
geotransform = inDataset.GetGeoTransform()
if geotransform is not None:
gt = list(geotransform)
gt[0] = gt[0] + x0 * gt[1]
gt[3] = gt[3] + y0 * gt[5]
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
outBand = outDataset.GetRasterBand(1)
outBand.WriteArray(class_image, 0, 0)
outBand.FlushCache()
outDataset = None
# write class membership probability file if desired
if probfile:
driver = gdal.GetDriverByName(probfmt)
outDataset = driver.Create(probfile, cols, rows, K, GDT_Byte)
if geotransform is not None:
outDataset.SetGeoTransform(tuple(gt))
if projection is not None:
outDataset.SetProjection(projection)
for k in range(K):
probs = np.reshape(U[k, :], (rows, cols))
probs = np.byte(probs * 255)
outBand = outDataset.GetRasterBand(k + 1)
outBand.WriteArray(probs, 0, 0)
outBand.FlushCache()
outDataset = None
print 'class probabilities written to: %s' % probfile
inDataset = None
if (outfmt == 'ENVI') and (K < 19):
# try to make an ENVI classification header file
hdr = header.Header()
headerfile = outfile + '.hdr'
f = open(headerfile)
line = f.readline()
envihdr = ''
while line:
envihdr += line
line = f.readline()
f.close()
hdr.read(envihdr)
hdr['file type'] = 'ENVI Classification'
hdr['classes'] = str(K + 1)
classlookup = '{0'
for i in range(1, 3 * (K + 1)):
classlookup += ', ' + str(str(ctable[i]))
classlookup += '}'
hdr['class lookup'] = classlookup
hdr['class names'] = ['class %i' % i for i in range(K + 1)]
f = open(headerfile, 'w')
f.write(str(hdr))
f.close()
print 'classification written to: ' + outfile
print 'elapsed time: ' + str(time.time() - start)
print '--done------------------------'