def layer_rasterization(self, raster_head, attribute_r):
"""
Function to rasterize a vector. Complete the gdal pointer empty properties with the layer's information
of the vector and a defined field.
:param raster_head: Raster path that will look like the final raster of the rasterization
:type raster_head: str
:param attribute_r: Field name of the shapefile that contains class names
:type attribute_r: str
:param class_r: Class name corresponding to integer class name
:type class_r: list of str
:param class_out: Value field pixels for the output raster
:type class_out: int
"""
# Export a example of a raster out information
# for the validation shapefile
example_raster = RasterSat_by_date('', '', [0]) # Call the raster class
raster_info = example_raster.raster_data(raster_head)# Extract data info
# Define the validation's vector
valid_raster = self.vector_used[:-3]+'TIF' # Name of the output raster
if os.path.exists(str(valid_raster)):
os.remove(valid_raster)
# Create the empty raster with the same properties
info_out = example_raster.create_raster(valid_raster, raster_info[0], raster_info[1])
self.raster_ds = example_raster.out_ds
# Virtual rasterize the vector
pt_rast = gdal.RasterizeLayer(self.raster_ds, [1], self.data_source.GetLayer(), \
options=["ATTRIBUTE=" + str(attribute_r)])
if pt_rast != 0:
raise Exception("error rasterizing layer: %s" % pt_rast)
new_data = self.raster_ds.ReadAsArray()
# # Expression to find classes in numpy's mask
# np_mask = ''
# # Convert string in a list. For that, it remove
# # space and clip this string with comma (Add everywhere if the script modified
# # because the process work with a input string chain)
# class_r = class_r.replace(' ','').split(',')
# for nm in class_r:
# np_mask = np_mask + '(new_data == ' + str(nm) + ') | '
# np_mask = np_mask[:-3]
# # Replace value data by our own value
# new_data=np.ma.masked_where(eval(np_mask), new_data)
# new_data.fill_value = class_out
#
# new_data = new_data.filled()
self.raster_ds = None
# Complete the raster creation
example_raster.complete_raster(info_out, new_data)
return valid_raster
def i_img_sat(self):
"""
Interface function to processing satellite images:
1. Clip archive images and modify Archive class to integrate clip image path.
With :func:`Toolbox.clip_raster` in ``Toolbox`` module.
2. Search cloud's percentage :func:`RasterSat_by_date.RasterSat_by_date.pourc_cloud`, select
image and compute ndvi index :func:`RasterSat_by_date.RasterSat_by_date.calcul_ndvi`. If cloud's percentage is
greater than 40%, then not select and compute ndvi index.
3. Compute temporal stats on ndvi index [min, max, std, min-max]. With :func:`Toolbox.calc_serie_stats`
in ``Toolbox`` module.
4. Create stats ndvi raster and stats cloud raster.
>>> import RasterSat_by_date
>>> stats_test = RasterSat_by_date(class_archive, big_folder, one_date)
>>> stats_test.complete_raster(stats_test.create_raster(in_raster, stats_data, in_ds), stats_data)
"""
# Clip archive images and modify Archive class to integrate clip image path
for clip in self.check_download.list_img:
clip_index = self.check_download.list_img.index(clip)
self.check_download.list_img[clip_index][3] = clip_raster(clip[3], self.path_area) # Multispectral images
self.check_download.list_img[clip_index][4] = clip_raster(clip[4], self.path_area) # Cloud images
# Images pre-processing
spectral_out = []
for date in self.check_download.single_date:
check_L8 = RasterSat_by_date(self.check_download, self.folder_processing, date)
check_L8.mosaic_by_date()
# Search cloud's percentage, select image and compute ndvi index if > cl
cl = check_L8.pourc_cloud(check_L8._one_date[3], check_L8._one_date[4])
if cl > 0.60:
check_L8.calcul_ndvi(check_L8._one_date[3])
spectral_out.append(check_L8._one_date)
# Compute temporal stats on ndvi index [min, max, std, min-max]
spectral_trans = np.transpose(np.array(spectral_out, dtype=object))
stats_name = ['Min', 'Max', 'Std', 'MaxMin']
stats_ndvi, stats_cloud = calc_serie_stats(spectral_trans)
# Create stats ndvi raster and stats cloud raster
stats_L8 = RasterSat_by_date(self.check_download, self.folder_processing, [int(self.classif_year)])
# Stats cloud raster
out_cloud_folder = stats_L8._class_archive._folder + '/' + stats_L8._big_folder + '/' + self.classif_year + \
'/Cloud_number_' + self.classif_year + '.TIF'
stats_L8.complete_raster(stats_L8.create_raster(out_cloud_folder, stats_cloud, \
stats_L8.raster_data(self.check_download.list_img[0][4])[1]), \
stats_cloud)
# Stats ndvi rasters
for stats_index in range(len(stats_ndvi)):
out_ndvistats_folder = stats_L8._class_archive._folder + '/' + stats_L8._big_folder + '/' + self.classif_year + \
'/' + stats_name[stats_index] + '_' + self.classif_year + '.TIF'
self.out_ndvistats_folder_tab[stats_index] = out_ndvistats_folder
stats_L8.complete_raster(stats_L8.create_raster(out_ndvistats_folder, stats_ndvi[stats_index], \
stats_L8.raster_data(self.check_download.list_img[0][4])[1]), \
stats_ndvi[stats_index])
