def __init__(self, iface):
"""Constructor.
:param iface: An interface instance that will be passed to this class
which provides the hook by which you can manipulate the QGIS
application at run time.
:type iface: QgsInterface
"""
# Save reference to the QGIS interface
self.iface = iface
# initialize plugin directory
self.plugin_dir = os.path.dirname(__file__)
# initialize locale
locale = QSettings().value('locale/userLocale')[0:2]
locale_path = os.path.join(self.plugin_dir, 'i18n',
'FFTConvolution_{}.qm'.format(locale))
if os.path.exists(locale_path):
self.translator = QTranslator()
self.translator.load(locale_path)
if qVersion() > '4.3.3':
QCoreApplication.installTranslator(self.translator)
# Create the dialog (after translation) and keep reference
self.dlg = FFTConvolutionDialog()
# Declare instance attributes
self.actions = []
self.menu = self.tr(u'&FFT COnvolution Filters')
# TODO: We are going to let the user set this up in a future iteration
self.toolbar = self.iface.addToolBar(u'FFTConvolution')
self.toolbar.setObjectName(u'FFTConvolution')
self.dlg.output_file.clear()
self.dlg.output_file_button.clicked.connect(self.select_output_file)
def __init__(self, iface):
"""Constructor.
:param iface: An interface instance that will be passed to this class
which provides the hook by which you can manipulate the QGIS
application at run time.
:type iface: QgsInterface
"""
# Save reference to the QGIS interface
self.iface = iface
# initialize plugin directory
self.plugin_dir = os.path.dirname(__file__)
# initialize locale
locale = QSettings().value('locale/userLocale')[0:2]
locale_path = os.path.join(
self.plugin_dir,
'i18n',
'FFTConvolution_{}.qm'.format(locale))
if os.path.exists(locale_path):
self.translator = QTranslator()
self.translator.load(locale_path)
if qVersion() > '4.3.3':
QCoreApplication.installTranslator(self.translator)
# Create the dialog (after translation) and keep reference
self.dlg = FFTConvolutionDialog()
# Declare instance attributes
self.actions = []
self.menu = self.tr(u'&FFT COnvolution Filters')
# TODO: We are going to let the user set this up in a future iteration
self.toolbar = self.iface.addToolBar(u'FFTConvolution')
self.toolbar.setObjectName(u'FFTConvolution')
self.dlg.output_file.clear()
self.dlg.output_file_button.clicked.connect(self.select_output_file)
class FFTConvolution:
"""QGIS Plugin Implementation."""
def __init__(self, iface):
"""Constructor.
:param iface: An interface instance that will be passed to this class
which provides the hook by which you can manipulate the QGIS
application at run time.
:type iface: QgsInterface
"""
# Save reference to the QGIS interface
self.iface = iface
# initialize plugin directory
self.plugin_dir = os.path.dirname(__file__)
# initialize locale
locale = QSettings().value('locale/userLocale')[0:2]
locale_path = os.path.join(self.plugin_dir, 'i18n',
'FFTConvolution_{}.qm'.format(locale))
if os.path.exists(locale_path):
self.translator = QTranslator()
self.translator.load(locale_path)
if qVersion() > '4.3.3':
QCoreApplication.installTranslator(self.translator)
# Create the dialog (after translation) and keep reference
self.dlg = FFTConvolutionDialog()
# Declare instance attributes
self.actions = []
self.menu = self.tr(u'&FFT COnvolution Filters')
# TODO: We are going to let the user set this up in a future iteration
self.toolbar = self.iface.addToolBar(u'FFTConvolution')
self.toolbar.setObjectName(u'FFTConvolution')
self.dlg.output_file.clear()
self.dlg.output_file_button.clicked.connect(self.select_output_file)
# noinspection PyMethodMayBeStatic
def tr(self, message):
"""Get the translation for a string using Qt translation API.
We implement this ourselves since we do not inherit QObject.
:param message: String for translation.
:type message: str, QString
:returns: Translated version of message.
:rtype: QString
"""
# noinspection PyTypeChecker,PyArgumentList,PyCallByClass
return QCoreApplication.translate('FFTConvolution', message)
def add_action(self,
icon_path,
text,
callback,
enabled_flag=True,
add_to_menu=True,
add_to_toolbar=True,
status_tip=None,
whats_this=None,
parent=None):
"""Add a toolbar icon to the toolbar.
:param icon_path: Path to the icon for this action. Can be a resource
path (e.g. ':/plugins/foo/bar.png') or a normal file system path.
:type icon_path: str
:param text: Text that should be shown in menu items for this action.
:type text: str
:param callback: Function to be called when the action is triggered.
:type callback: function
:param enabled_flag: A flag indicating if the action should be enabled
by default. Defaults to True.
:type enabled_flag: bool
:param add_to_menu: Flag indicating whether the action should also
be added to the menu. Defaults to True.
:type add_to_menu: bool
:param add_to_toolbar: Flag indicating whether the action should also
be added to the toolbar. Defaults to True.
:type add_to_toolbar: bool
:param status_tip: Optional text to show in a popup when mouse pointer
hovers over the action.
:type status_tip: str
:param parent: Parent widget for the new action. Defaults None.
:type parent: QWidget
:param whats_this: Optional text to show in the status bar when the
mouse pointer hovers over the action.
:returns: The action that was created. Note that the action is also
added to self.actions list.
:rtype: QAction
"""
icon = QIcon(icon_path)
action = QAction(icon, text, parent)
action.triggered.connect(callback)
action.setEnabled(enabled_flag)
if status_tip is not None:
action.setStatusTip(status_tip)
if whats_this is not None:
action.setWhatsThis(whats_this)
if add_to_toolbar:
self.toolbar.addAction(action)
if add_to_menu:
self.iface.addPluginToRasterMenu(self.menu, action)
self.actions.append(action)
return action
def initGui(self):
"""Create the menu entries and toolbar icons inside the QGIS GUI."""
icon_path = ':/plugins/FFTConvolution/icon.png'
self.add_action(icon_path,
text=self.tr(u'FFT Convolution filters'),
callback=self.run,
parent=self.iface.mainWindow())
def unload(self):
"""Removes the plugin menu item and icon from QGIS GUI."""
for action in self.actions:
self.iface.removePluginRasterMenu(
self.tr(u'&FFT COnvolution Filters'), action)
self.iface.removeToolBarIcon(action)
# remove the toolbar
del self.toolbar
def select_output_file(self):
filename = QFileDialog.getSaveFileName(self.dlg, "Select output file ",
"", '*.tif')
self.dlg.output_file.setText(filename)
def run(self):
"""Run method that performs all the real work"""
# show the dialog
self.dlg.show()
# Run the dialog event loop
result = self.dlg.exec_()
# See if OK was pressed
if result:
if self.dlg.smoothing.isChecked():
edge = False
else:
edge = True
#call the function linking to real work
#the input is translated from the GUI input to correct format here
self.fft_convolution(in_layer=self.dlg.input_layer.currentLayer(),
out_path=self.dlg.output_file.text(),
size=self.dlg.size.text(),
edge=edge,
new_crs=self.dlg.crs.crs(),
tiled=self.dlg.window.isChecked(),
tilerows=self.dlg.window_rows.text(),
tilecols=self.dlg.window_cols.text(),
add_layer=self.dlg.check_add.isChecked())
#this function parses the arguments, calls the appropriate functions and displays the new layer if needed
def fft_convolution(self,
in_layer,
out_path,
size=10,
edge=False,
new_crs=None,
tiled=False,
tilerows=0,
tilecols=0,
add_layer=True):
#if the file has no extension, add '.tif'
ext = os.path.splitext(out_path)[-1].lower()
if ext == '':
out_path = out_path + '.tif'
#if the file already exists, ask the user
if os.path.isfile(out_path):
reply = QMessageBox.question(None, 'File exists!',
'File exists - overwite it?',
QMessageBox.Yes, QMessageBox.No)
if reply == QMessageBox.No:
return False
#we need the CRS as EPSG code, or None if invalid
if new_crs.isValid():
new_crs = new_crs.authid()
else:
new_crs = None
#preprocessing the input layer's path
in_path = in_layer.dataProvider().dataSourceUri()
#QMessageBox.information(None, "DEBUG:", str(in_path))
if in_path.find('=') > -1:
QMessageBox.information(None, "Sorry!",
"WMS support wasn't implemented yet!")
return False
#the main computation
layer = self.gaussian_filter(in_path=in_path,
out_path=out_path,
size=int(re.sub(r"\D", "", size)),
edge=edge,
tiled=tiled,
tilerows=tilerows,
tilecols=tilecols,
new_crs=new_crs)
if add_layer:
QgsMapLayerRegistry.instance().addMapLayers([layer])
qgis.utils.iface.mapCanvas().refresh()
#returns number of array dimensions
def __array_rank(self, arr):
return len(arr.shape)
#loads the newly created layer
def __load_layer(self, path):
fileName = path
fileInfo = QFileInfo(fileName)
baseName = fileInfo.baseName()
rlayer = QgsRasterLayer(fileName, baseName)
if not rlayer.isValid():
raise Exception(
"Computation finished, but layer failed to load. Inspect the path zou specified for the output layer."
)
return rlayer
#pads the window to process its original extend accurately
def __extend_window(self, window, size, height, width):
minrow = max(window[0][0] - size, 0)
maxrow = min(window[0][1] + size, height)
mincol = max(window[1][0] - size, 0)
maxcol = min(window[1][1] + size, width)
return ((minrow, maxrow), (mincol, maxcol))
#creates a window generator, in the same format as it is returned by block_windows method
def __generate_windows(self, height, width, tilerows, tilecols):
rownum = int(math.ceil(float(height) / tilerows))
colnum = int(math.ceil(float(width) / tilecols))
for i in range(rownum):
#last row's and column's dimensions are computed by modulo - they are smaller than regular tiles
if i == rownum - 1:
rowsize = height % tilerows
else:
rowsize = tilerows
for j in range(colnum):
if j == colnum - 1:
colsize = width % tilecols
else:
colsize = tilecols
cell = ((i, j), ((i * tilerows, i * tilerows + rowsize),
(j * tilecols, j * tilecols + colsize)))
yield cell
#like __generate_windows(), but returns an array
def __generate_window_array(self, height, width, tilerows, tilecols):
rownum = int(math.ceil(float(height) / tilerows))
colnum = int(math.ceil(float(width) / tilecols))
windows = np.asarray(np.zeros((height, width), dtype=object))
for i in range(rownum):
#last row's and column's dimensions are computed by modulo - they are smaller than regular tiles
if i == rownum - 1:
rowsize = height % tilerows
else:
rowsize = tilerows
for j in range(colnum):
if j == colnum - 1:
colsize = width % tilecols
else:
colsize = tilecols
windows[i][j] = ((i * tilerows, i * tilerows + rowsize),
(j * tilecols, j * tilecols + colsize))
return windows
#processes the window parameters
#returns the windows as a generator or an array (specified in the generator parameter)
def __compute_windows(self,
in_raster,
height,
width,
size,
tilerows=0,
tilecols=0,
generator=True):
#input validation
size = int(size)
try:
tilerows = int(tilerows)
except ValueError:
tilerows = 0
try:
tilecols = int(tilecols)
except ValueError:
tilecols = 0
#when raster's dimensions are modified due to reprojection, we must adjust the windows as well
#this algorithm is quick'n'dirty - we just sompute one ratio and make all tiles larger/smaller
#reprojecting each tile node would be better - perhaps in some future version
if height != in_raster.height or width != in_raster.width:
hratio = float(height) / in_raster.height
wratio = float(width) / in_raster.width
else:
hratio = 1
wratio = 1
windows = []
#if only one of tile's dimension was set, we assume a square
if tilecols == 0 and tilerows > 0:
tilecols = tilerows
elif tilerows == 0 and tilecols > 0:
tilerows = tilecols
#if tiles are too small (including default 0 length), we make them automatically
#"2*size" is the total padding length and also an arbitrarily chosen minimum
#"size" would be a minimum to get accurate tiles, but this way only 1/9 of the tile would be useful => inefficient
#"2*size" means at least 50% efficiency
if min(tilerows, tilecols) <= 2 * size:
#if the raster has blocks and they are big enough, we use them
blocks = in_raster.block_shapes
block = blocks[0]
if min(block[0], block[1]) > 2 * size:
#if we compute the original raster, use the block as-is
#otherwise use the dimensions and continue
if hratio == 1 and wratio == 1:
return in_raster.block_windows(1)
else:
tilerows = block[0]
tilecols = block[1]
else:
#"2*size + 100" is an arbitrary constant
#it's quite efficient on smaller rasters and shouldn't make any memory issues
#really small rasters shouldn't be computed by tiles anyway
tilerows = 2 * size + 100
tilecols = 2 * size + 100
#we transform the dimensions if needed
tilerows = int(hratio * tilerows)
tilecols = int(wratio * tilecols)
#if the tiles are too big (more than half of any dimension of the raster),
#we switch to the untiled algorithm
if 2 * tilerows >= height or 2 * tilecols >= width:
return False
#if windows are not read from the raster, we must make them
if generator:
windows = self.__generate_windows(height, width, tilerows,
tilecols)
else:
windows = self.__generate_window_array(height, width, tilerows,
tilecols)
return windows
#computes the affine transformation, raster dimensions and metadata for the new raster
def __compute_transform(self, in_raster, new_crs):
affine, width, height = calculate_default_transform(
in_raster.crs, new_crs, in_raster.width, in_raster.height,
*in_raster.bounds)
kwargs = in_raster.meta.copy()
kwargs.update({
'driver': 'GTiff',
'crs': new_crs,
'transform': affine,
'affine': affine,
'width': width,
'height': height
})
return affine, height, width, kwargs
#calls reproject() function for every band
def __reproject(self, in_raster, out_raster, affine, new_crs):
for k in range(1, in_raster.count + 1):
reproject(source=rasterio.band(in_raster, k),
destination=rasterio.band(out_raster, k),
src_transform=affine,
src_crs=in_raster.crs,
dst_transform=affine,
dst_crs=new_crs,
resampling=RESAMPLING.nearest)
return out_raster
#the original MikeT's function from http://gis.stackexchange.com/a/10467
#my addition is the conversion of the padded array to float to avoid errors with integer rasters
#the intended input is a single band raster array
def __gaussian_blur1d(self, in_array, size):
#check validity
try:
if 0 in in_array.shape:
raise Exception("Null array can't be processed!")
except TypeError:
raise Exception("Null array can't be processed!")
# expand in_array to fit edge of kernel
padded_array = np.pad(in_array, size, 'symmetric').astype(float)
# build kernel
x, y = np.mgrid[-size:size + 1, -size:size + 1]
g = np.exp(-(x**2 / float(size) + y**2 / float(size)))
g = (g / g.sum()).astype(float)
# do the Gaussian blur
out_array = fftconvolve(padded_array, g, mode='valid')
return out_array.astype(in_array.dtype)
#the intended input is an array with various number of bands
#returns a numpy array
def __gaussian_blur(self, in_array, size):
#make sure the input is a numpy array
in_array = np.asarray(in_array)
#find number of dimensions - 2 for single-band or 3 for multiband rasters
rank = self.__array_rank(in_array)
if rank == 2:
return self.__gaussian_blur1d(in_array,
size).astype(in_array.dtype)
elif rank > 3 or rank == 1:
raise TypeError("Invalid number of dimensions!")
#continue to multiband
count = in_array.shape[0]
out = []
for i in range(count):
band = in_array[i]
out_band = self.__gaussian_blur1d(band,
size) #.astype(in_array.dtype)
#if out_band != False:
out.append(out_band)
return np.asarray(out)
#the real work is done here
#filters a raster specified by the file's path (in_path) and writes it to another file (out_path)
def gaussian_filter(self,
in_path,
out_path,
size,
edge=False,
tiled=False,
tilerows=0,
tilecols=0,
new_crs=None):
with rasterio.drivers():
with rasterio.open(in_path, 'r') as in_raster:
if new_crs == None:
new_crs = in_raster.crs
affine, height, width, kwargs = self.__compute_transform(
in_raster, new_crs)
if tiled:
#we make two sets of tiles, for the old and the new raster
#this is important in case of reprojection
old_windows = self.__compute_windows(
in_raster=in_raster,
height=in_raster.height,
width=in_raster.width,
size=size,
tilerows=tilerows,
tilecols=tilecols)
#windows for the new raster are made in two steps: generator and array
new_windows = self.__compute_windows(in_raster=in_raster,
height=height,
width=width,
size=size,
tilerows=tilerows,
tilecols=tilecols,
generator=False)
#if windows are too big or invalid, we process the raster without them
try:
iter(old_windows)
iter(new_windows)
except TypeError:
tiled = False
with rasterio.open(out_path, 'w', **kwargs) as out_raster:
out_raster = self.__reproject(in_raster, out_raster,
affine, new_crs)
if tiled:
for index, window in old_windows:
oldbigwindow = self.__extend_window(
window, size, in_raster.height,
in_raster.width)
in_array = in_raster.read(window=oldbigwindow)
out_array = self.__gaussian_blur(in_array, size)
#for edge detection we subtract the output array from the original
#this may produce some artifacts when the raster is reprojected
#or extensive and with degree coordinates
if edge:
out_array = np.subtract(in_array, out_array)
#now compute the window for writing into the new raster
nwindow = new_windows[index[0]][index[1]]
newbigwindow = self.__extend_window(
nwindow, size, height, width)
out_raster.write(out_array, window=newbigwindow)
else:
in_array = in_raster.read()
out_array = self.__gaussian_blur(in_array, size)
if edge:
out_array = out_array = np.subtract(
in_array, out_array)
out_raster.write(out_array)
return self.__load_layer(out_path)
class FFTConvolution:
"""QGIS Plugin Implementation."""
def __init__(self, iface):
"""Constructor.
:param iface: An interface instance that will be passed to this class
which provides the hook by which you can manipulate the QGIS
application at run time.
:type iface: QgsInterface
"""
# Save reference to the QGIS interface
self.iface = iface
# initialize plugin directory
self.plugin_dir = os.path.dirname(__file__)
# initialize locale
locale = QSettings().value('locale/userLocale')[0:2]
locale_path = os.path.join(
self.plugin_dir,
'i18n',
'FFTConvolution_{}.qm'.format(locale))
if os.path.exists(locale_path):
self.translator = QTranslator()
self.translator.load(locale_path)
if qVersion() > '4.3.3':
QCoreApplication.installTranslator(self.translator)
# Create the dialog (after translation) and keep reference
self.dlg = FFTConvolutionDialog()
# Declare instance attributes
self.actions = []
self.menu = self.tr(u'&FFT COnvolution Filters')
# TODO: We are going to let the user set this up in a future iteration
self.toolbar = self.iface.addToolBar(u'FFTConvolution')
self.toolbar.setObjectName(u'FFTConvolution')
self.dlg.output_file.clear()
self.dlg.output_file_button.clicked.connect(self.select_output_file)
# noinspection PyMethodMayBeStatic
def tr(self, message):
"""Get the translation for a string using Qt translation API.
We implement this ourselves since we do not inherit QObject.
:param message: String for translation.
:type message: str, QString
:returns: Translated version of message.
:rtype: QString
"""
# noinspection PyTypeChecker,PyArgumentList,PyCallByClass
return QCoreApplication.translate('FFTConvolution', message)
def add_action(
self,
icon_path,
text,
callback,
enabled_flag=True,
add_to_menu=True,
add_to_toolbar=True,
status_tip=None,
whats_this=None,
parent=None):
"""Add a toolbar icon to the toolbar.
:param icon_path: Path to the icon for this action. Can be a resource
path (e.g. ':/plugins/foo/bar.png') or a normal file system path.
:type icon_path: str
:param text: Text that should be shown in menu items for this action.
:type text: str
:param callback: Function to be called when the action is triggered.
:type callback: function
:param enabled_flag: A flag indicating if the action should be enabled
by default. Defaults to True.
:type enabled_flag: bool
:param add_to_menu: Flag indicating whether the action should also
be added to the menu. Defaults to True.
:type add_to_menu: bool
:param add_to_toolbar: Flag indicating whether the action should also
be added to the toolbar. Defaults to True.
:type add_to_toolbar: bool
:param status_tip: Optional text to show in a popup when mouse pointer
hovers over the action.
:type status_tip: str
:param parent: Parent widget for the new action. Defaults None.
:type parent: QWidget
:param whats_this: Optional text to show in the status bar when the
mouse pointer hovers over the action.
:returns: The action that was created. Note that the action is also
added to self.actions list.
:rtype: QAction
"""
icon = QIcon(icon_path)
action = QAction(icon, text, parent)
action.triggered.connect(callback)
action.setEnabled(enabled_flag)
if status_tip is not None:
action.setStatusTip(status_tip)
if whats_this is not None:
action.setWhatsThis(whats_this)
if add_to_toolbar:
self.toolbar.addAction(action)
if add_to_menu:
self.iface.addPluginToRasterMenu(
self.menu,
action)
self.actions.append(action)
return action
def initGui(self):
"""Create the menu entries and toolbar icons inside the QGIS GUI."""
icon_path = ':/plugins/FFTConvolution/icon.png'
self.add_action(
icon_path,
text=self.tr(u'FFT Convolution filters'),
callback=self.run,
parent=self.iface.mainWindow())
def unload(self):
"""Removes the plugin menu item and icon from QGIS GUI."""
for action in self.actions:
self.iface.removePluginRasterMenu(
self.tr(u'&FFT COnvolution Filters'),
action)
self.iface.removeToolBarIcon(action)
# remove the toolbar
del self.toolbar
def select_output_file(self):
filename = QFileDialog.getSaveFileName(self.dlg, "Select output file ","", '*.tif')
self.dlg.output_file.setText(filename)
def run(self):
"""Run method that performs all the real work"""
# show the dialog
self.dlg.show()
# Run the dialog event loop
result = self.dlg.exec_()
# See if OK was pressed
if result:
if self.dlg.smoothing.isChecked():
edge = False
else:
edge = True
#call the function linking to real work
#the input is translated from the GUI input to correct format here
self.fft_convolution(
in_layer=self.dlg.input_layer.currentLayer(),
out_path=self.dlg.output_file.text(),
size=self.dlg.size.text(),
edge=edge,
new_crs=self.dlg.crs.crs(),
tiled=self.dlg.window.isChecked(),
tilerows=self.dlg.window_rows.text(),
tilecols=self.dlg.window_cols.text(),
add_layer=self.dlg.check_add.isChecked()
)
#this function parses the arguments, calls the appropriate functions and displays the new layer if needed
def fft_convolution( self, in_layer, out_path, size=10, edge=False, new_crs=None,
tiled=False, tilerows=0, tilecols=0, add_layer=True ):
#if the file has no extension, add '.tif'
ext = os.path.splitext(out_path)[-1].lower()
if ext == '':
out_path = out_path + '.tif'
#if the file already exists, ask the user
if os.path.isfile(out_path):
reply = QMessageBox.question(
None,'File exists!','File exists - overwite it?',
QMessageBox.Yes, QMessageBox.No
)
if reply == QMessageBox.No:
return False
#we need the CRS as EPSG code, or None if invalid
if new_crs.isValid():
new_crs = new_crs.authid()
else:
new_crs = None
#preprocessing the input layer's path
in_path = in_layer.dataProvider().dataSourceUri()
#QMessageBox.information(None, "DEBUG:", str(in_path))
if in_path.find('=') > -1:
QMessageBox.information(None, "Sorry!", "WMS support wasn't implemented yet!")
return False
#the main computation
layer = self.gaussian_filter(
in_path = in_path,
out_path=out_path,
size = int(re.sub(r"\D", "", size)),
edge=edge,
tiled = tiled,
tilerows = tilerows,
tilecols = tilecols,
new_crs = new_crs
)
if add_layer:
QgsMapLayerRegistry.instance().addMapLayers([layer])
qgis.utils.iface.mapCanvas().refresh()
#returns number of array dimensions
def __array_rank(self, arr):
return len(arr.shape)
#loads the newly created layer
def __load_layer(self, path):
fileName = path
fileInfo = QFileInfo(fileName)
baseName = fileInfo.baseName()
rlayer = QgsRasterLayer(fileName, baseName)
if not rlayer.isValid():
raise Exception("Computation finished, but layer failed to load. Inspect the path zou specified for the output layer.")
return rlayer
#pads the window to process its original extend accurately
def __extend_window(self, window,size,height,width):
minrow = max(window[0][0]-size,0)
maxrow = min(window[0][1]+size,height)
mincol = max(window[1][0]-size,0)
maxcol = min(window[1][1]+size,width)
return ((minrow,maxrow),(mincol,maxcol))
#creates a window generator, in the same format as it is returned by block_windows method
def __generate_windows(self, height, width, tilerows, tilecols):
rownum = int(math.ceil(float(height)/tilerows))
colnum = int(math.ceil(float(width)/tilecols))
for i in range(rownum):
#last row's and column's dimensions are computed by modulo - they are smaller than regular tiles
if i == rownum-1:
rowsize = height%tilerows
else:
rowsize = tilerows
for j in range(colnum):
if j == colnum-1:
colsize = width%tilecols
else:
colsize = tilecols
cell = ((i,j),((i*tilerows,i*tilerows+rowsize),(j*tilecols,j*tilecols+colsize)))
yield cell
#like __generate_windows(), but returns an array
def __generate_window_array(self, height, width, tilerows, tilecols):
rownum = int(math.ceil(float(height)/tilerows))
colnum = int(math.ceil(float(width)/tilecols))
windows = np.asarray(np.zeros((height, width),dtype=object))
for i in range(rownum):
#last row's and column's dimensions are computed by modulo - they are smaller than regular tiles
if i == rownum-1:
rowsize = height%tilerows
else:
rowsize = tilerows
for j in range(colnum):
if j == colnum-1:
colsize = width%tilecols
else:
colsize = tilecols
windows[i][j]=((i*tilerows,i*tilerows+rowsize),(j*tilecols,j*tilecols+colsize))
return windows
#processes the window parameters
#returns the windows as a generator or an array (specified in the generator parameter)
def __compute_windows(self, in_raster, height, width, size, tilerows=0, tilecols=0, generator=True):
#input validation
size = int(size)
try:
tilerows = int(tilerows)
except ValueError:
tilerows = 0
try:
tilecols = int(tilecols)
except ValueError:
tilecols = 0
#when raster's dimensions are modified due to reprojection, we must adjust the windows as well
#this algorithm is quick'n'dirty - we just sompute one ratio and make all tiles larger/smaller
#reprojecting each tile node would be better - perhaps in some future version
if height != in_raster.height or width != in_raster.width:
hratio = float(height)/in_raster.height
wratio = float(width)/in_raster.width
else:
hratio = 1
wratio = 1
windows = []
#if only one of tile's dimension was set, we assume a square
if tilecols == 0 and tilerows > 0:
tilecols = tilerows
elif tilerows == 0 and tilecols > 0:
tilerows = tilecols
#if tiles are too small (including default 0 length), we make them automatically
#"2*size" is the total padding length and also an arbitrarily chosen minimum
#"size" would be a minimum to get accurate tiles, but this way only 1/9 of the tile would be useful => inefficient
#"2*size" means at least 50% efficiency
if min(tilerows, tilecols) <= 2*size:
#if the raster has blocks and they are big enough, we use them
blocks = in_raster.block_shapes
block = blocks[0]
if min(block[0],block[1]) > 2*size:
#if we compute the original raster, use the block as-is
#otherwise use the dimensions and continue
if hratio==1 and wratio==1:
return in_raster.block_windows(1)
else:
tilerows = block[0]
tilecols = block[1]
else:
#"2*size + 100" is an arbitrary constant
#it's quite efficient on smaller rasters and shouldn't make any memory issues
#really small rasters shouldn't be computed by tiles anyway
tilerows = 2*size + 100
tilecols = 2*size + 100
#we transform the dimensions if needed
tilerows = int(hratio * tilerows)
tilecols = int(wratio * tilecols)
#if the tiles are too big (more than half of any dimension of the raster),
#we switch to the untiled algorithm
if 2*tilerows >= height or 2*tilecols >= width:
return False
#if windows are not read from the raster, we must make them
if generator:
windows = self.__generate_windows(height, width, tilerows, tilecols)
else:
windows = self.__generate_window_array(height, width, tilerows, tilecols)
return windows
#computes the affine transformation, raster dimensions and metadata for the new raster
def __compute_transform(self, in_raster, new_crs):
affine, width, height = calculate_default_transform(
in_raster.crs, new_crs, in_raster.width, in_raster.height, *in_raster.bounds
)
kwargs = in_raster.meta.copy()
kwargs.update({
'driver':'GTiff',
'crs': new_crs,
'transform': affine,
'affine': affine,
'width': width,
'height': height
})
return affine, height, width, kwargs
#calls reproject() function for every band
def __reproject(self, in_raster, out_raster, affine, new_crs):
for k in range(1, in_raster.count + 1):
reproject(
source=rasterio.band(in_raster, k),
destination=rasterio.band(out_raster, k),
src_transform=affine,
src_crs=in_raster.crs,
dst_transform=affine,
dst_crs=new_crs,
resampling=RESAMPLING.nearest)
return out_raster
#the original MikeT's function from http://gis.stackexchange.com/a/10467
#my addition is the conversion of the padded array to float to avoid errors with integer rasters
#the intended input is a single band raster array
def __gaussian_blur1d(self, in_array, size):
#check validity
try:
if 0 in in_array.shape:
raise Exception("Null array can't be processed!")
except TypeError:
raise Exception("Null array can't be processed!")
# expand in_array to fit edge of kernel
padded_array = np.pad(in_array, size, 'symmetric').astype(float)
# build kernel
x, y = np.mgrid[-size:size + 1, -size:size + 1]
g = np.exp(-(x**2 / float(size) + y**2 / float(size)))
g = (g / g.sum()).astype(float)
# do the Gaussian blur
out_array = fftconvolve(padded_array, g, mode='valid')
return out_array.astype(in_array.dtype)
#the intended input is an array with various number of bands
#returns a numpy array
def __gaussian_blur(self, in_array, size):
#make sure the input is a numpy array
in_array = np.asarray(in_array)
#find number of dimensions - 2 for single-band or 3 for multiband rasters
rank = self.__array_rank(in_array)
if rank == 2:
return self.__gaussian_blur1d(in_array, size).astype(in_array.dtype)
elif rank > 3 or rank == 1:
raise TypeError("Invalid number of dimensions!")
#continue to multiband
count = in_array.shape[0]
out = []
for i in range(count):
band = in_array[i]
out_band = self.__gaussian_blur1d(band, size)#.astype(in_array.dtype)
#if out_band != False:
out.append( out_band )
return np.asarray(out)
#the real work is done here
#filters a raster specified by the file's path (in_path) and writes it to another file (out_path)
def gaussian_filter(self, in_path, out_path, size, edge=False, tiled=False, tilerows=0, tilecols=0, new_crs=None):
with rasterio.drivers():
with rasterio.open(in_path,'r') as in_raster:
if new_crs == None:
new_crs = in_raster.crs
affine, height, width, kwargs = self.__compute_transform(in_raster, new_crs)
if tiled:
#we make two sets of tiles, for the old and the new raster
#this is important in case of reprojection
old_windows = self.__compute_windows(
in_raster=in_raster,
height=in_raster.height,
width=in_raster.width,
size=size,
tilerows=tilerows,
tilecols=tilecols
)
#windows for the new raster are made in two steps: generator and array
new_windows = self.__compute_windows(
in_raster=in_raster,
height=height,
width=width,
size=size,
tilerows=tilerows,
tilecols=tilecols,
generator=False
)
#if windows are too big or invalid, we process the raster without them
try:
iter(old_windows)
iter(new_windows)
except TypeError:
tiled = False
with rasterio.open(out_path,'w',**kwargs) as out_raster:
out_raster = self.__reproject(in_raster, out_raster, affine, new_crs)
if tiled:
for index, window in old_windows:
oldbigwindow = self.__extend_window(window,size,in_raster.height,in_raster.width)
in_array = in_raster.read(window=oldbigwindow)
out_array = self.__gaussian_blur(in_array, size)
#for edge detection we subtract the output array from the original
#this may produce some artifacts when the raster is reprojected
#or extensive and with degree coordinates
if edge:
out_array = np.subtract(in_array, out_array)
#now compute the window for writing into the new raster
nwindow = new_windows[index[0]][index[1]]
newbigwindow = self.__extend_window(nwindow,size,height,width)
out_raster.write(out_array,window=newbigwindow)
else:
in_array = in_raster.read()
out_array = self.__gaussian_blur(in_array, size)
if edge:
out_array = out_array = np.subtract(in_array, out_array)
out_raster.write(out_array)
return self.__load_layer(out_path)
