def apply_shdi_300(in_fn, out_fn300):
in_ds = gdal.Open(in_fn)
in_band = in_ds.GetRasterBand(1)
in_data = in_band.ReadAsArray()
slices = pb.make_slices(in_data, (21, 21))
stacked_data = np.ma.dstack(slices)
rows, cols = in_band.YSize, in_band.XSize
out_data = np.ones((rows, cols), np.int32) * -99
out_data[10:-10, 10:-10] = np.mean(stacked_data, 2)
pb.make_raster(in_ds, out_fn300, out_data, gdal.GDT_Int32, -99)
del in_ds
print("shdi 300 executed")
# follow the method shown in listing 10.2.
os.chdir(os.path.join(data_dir, 'Massachusetts'))
in_fn = 'm_4207162_ne_19_1_20140718_20140923.tif'
out_fn = 'ndvi2.tif'
ds = gdal.Open(in_fn)
red = ds.GetRasterBand(1).ReadAsArray().astype(np.float)
nir = ds.GetRasterBand(4).ReadAsArray()
red = np.ma.masked_where(nir + red == 0, red)
# This is the first method shown in the text.
ndvi = (nir - red) / (nir + red)
ndvi = np.where(np.isnan(ndvi), -99, ndvi)
ndvi = np.where(np.isinf(ndvi), -99, ndvi)
pb.make_raster(ds, 'ndvi2.tif', ndvi, gdal.GDT_Float32, -99)
# This is the second method shown in the text.
ndvi = np.where(nir + red > 0, (nir - red) / (nir + red), -99)
pb.make_raster(ds, 'ndvi3.tif', ndvi, gdal.GDT_Float32, -99)
del ds
########################### 10.2.2 Focal Analyses ###########################
indata = np.array([
[3, 5, 6, 4, 4, 3],
[4, 5, 8, 9, 6, 5],
[2, 2, 5, 7, 6, 4],
[5, 7, 9, 8, 9, 7],
# Script to smooth an elevation dataset. import numpy as np from osgeo import gdal import ospybook as pb in_fn = "../dati/osgeopy-data/Nepal/everest.tif" out_fn = 'everest_smoothed_edges.tif' in_ds = gdal.Open(in_fn) in_band = in_ds.GetRasterBand(1) in_data = in_band.ReadAsArray() # Stack the slices slices = pb.make_slices(in_data, (3, 3)) stacked_data = np.ma.dstack(slices) rows, cols = in_band.YSize, in_band.XSize # Initialize an output array to the NoData value (-99) out_data = np.ones((rows, cols), np.int32) * -99 # Put the result into the middle of the output, leaving the # outside rows and columns alone, so they still have -99. out_data[1:-1, 1:-1] = np.mean(stacked_data, 2) pb.make_raster(in_ds, out_fn, out_data, gdal.GDT_Int32, -99) del in_ds
(8 * cell_height)
run = ((data[2] + (2 * data[5]) + data[8]) -
(data[0] + (2 * data[3]) + data[6])) / \
(8 * cell_width)
dist = np.sqrt(np.square(rise) + np.square(run))
return np.arctan(dist) * 180 / np.pi
# Read the data as floating point.
in_ds = gdal.Open(in_fn)
in_band = in_ds.GetRasterBand(1)
in_data = in_band.ReadAsArray().astype(np.float32)
cell_width = in_ds.GetGeoTransform()[1]
cell_height = in_ds.GetGeoTransform()[5]
# Pass your slope function, along with the pixel size, to the
# generic_filter function. This will apply your function to
# each window, and each time it will pass the extra arguments
# in to the function along with the data.
out_data = scipy.ndimage.filters.generic_filter(in_data,
slope,
size=3,
mode='nearest',
extra_arguments=(cell_width,
cell_height))
# Save the output
pb.make_raster(in_ds, out_fn, out_data, gdal.GDT_Float32)
del in_ds
# Script to compute NDVI.
import os
import numpy as np
from osgeo import gdal
import ospybook as pb
os.chdir(r'D:\osgeopy-data\Massachusetts')
in_fn = 'm_4207162_ne_19_1_20140718_20140923.tif'
out_fn = 'ndvi2.tif'
ds = gdal.Open(in_fn)
red = ds.GetRasterBand(1).ReadAsArray().astype(np.float)
nir = ds.GetRasterBand(4).ReadAsArray()
# Mask the red band.
red = np.ma.masked_where(nir + red == 0, red)
# Do the calculation.
ndvi = (nir - red) / (nir + red)
# Fill the empty cells.
ndvi = ndvi.filled(-99)
# Set NoData to the fill value when creating the new raster.
out_ds = pb.make_raster(ds, out_fn, ndvi, gdal.GDT_Float32, -99)
overviews = pb.compute_overview_levels(out_ds.GetRasterBand(1))
out_ds.BuildOverviews('average', overviews)
del ds, out_ds
data - 1D array containing the 9 pixel values, starting
in the upper left and going left to right and down
cell_width - pixel width in the same units as the data
cell_height - pixel height in the same units as the data
"""
rise = ((data[6] + (2 * data[7]) + data[8]) - (data[0] + (2 * data[1]) + data[2])) / (8 * cell_height)
run = ((data[2] + (2 * data[5]) + data[8]) - (data[0] + (2 * data[3]) + data[6])) / (8 * cell_width)
dist = np.sqrt(np.square(rise) + np.square(run))
return np.arctan(dist) * 180 / np.pi
# Read the data as floating point.
in_ds = gdal.Open(in_fn)
in_band = in_ds.GetRasterBand(1)
in_data = in_band.ReadAsArray().astype(np.float32)
cell_width = in_ds.GetGeoTransform()[1]
cell_height = in_ds.GetGeoTransform()[5]
# Pass your slope function, along with the pixel size, to the
# generic_filter function. This will apply your function to
# each window, and each time it will pass the extra arguments
# in to the function along with the data.
out_data = scipy.ndimage.filters.generic_filter(
in_data, slope, size=3, mode="nearest", extra_arguments=(cell_width, cell_height)
)
# Save the output
pb.make_raster(in_ds, out_fn, out_data, gdal.GDT_Float32)
del in_ds
folder = r"D:\osgeopy-data\Landsat\Utah"
raster_fns = ["LE70380322000181EDC02_60m.tif", "LE70380322000181EDC02_TIR_60m.tif"]
out_fn = "kmeans_prediction_60m.tif"
# Function from list 10.10.
def stack_bands(filenames):
"""Returns a 3D array containing all band data from all files."""
bands = []
for fn in raster_fns:
ds = gdal.Open(fn)
for i in range(1, ds.RasterCount + 1):
bands.append(ds.GetRasterBand(i).ReadAsArray())
return np.dstack(bands)
# Stack the bands and run the analysis.
os.chdir(folder)
data = pb.stack_bands(raster_fns)
classes, centers = spectral.kmeans(data)
# Save the output.
ds = gdal.Open(raster_fns[0])
out_ds = pb.make_raster(ds, out_fn, classes, gdal.GDT_Byte)
levels = pb.compute_overview_levels(out_ds.GetRasterBand(1))
out_ds.BuildOverviews("NEAREST", levels)
out_ds.FlushCache()
out_ds.GetRasterBand(1).ComputeStatistics(False)
del out_ds, ds
'LE70380322000181EDC02_60m.tif', 'LE70380322000181EDC02_TIR_60m.tif'
]
out_fn = 'kmeans_prediction_60m.tif'
# Function from list 10.10.
def stack_bands(filenames):
"""Returns a 3D array containing all band data from all files."""
bands = []
for fn in raster_fns:
ds = gdal.Open(fn)
for i in range(1, ds.RasterCount + 1):
bands.append(ds.GetRasterBand(i).ReadAsArray())
return np.dstack(bands)
# Stack the bands and run the analysis.
os.chdir(folder)
data = pb.stack_bands(raster_fns)
classes, centers = spectral.kmeans(data)
# Save the output.
ds = gdal.Open(raster_fns[0])
out_ds = pb.make_raster(ds, out_fn, classes, gdal.GDT_Byte)
levels = pb.compute_overview_levels(out_ds.GetRasterBand(1))
out_ds.BuildOverviews('NEAREST', levels)
out_ds.FlushCache()
out_ds.GetRasterBand(1).ComputeStatistics(False)
del out_ds, ds
data = pb.stack_bands(raster_fns)
# Sample the satellite data at the coordinates from the csv.
sample = data[rows, cols, :]
# Fit the classification tree.
clf = tree.DecisionTreeClassifier(max_depth=5)
clf = clf.fit(sample, classes)
# Apply the new classification tree model to the satellite data.
rows, cols, bands = data.shape
data2d = np.reshape(data, (rows * cols, bands))
prediction = clf.predict(data2d)
prediction = np.reshape(prediction, (rows, cols))
# Set the pixels with no valid satellite data to 0.
prediction[np.sum(data, 2) == 0] = 0
# Save the output.
predict_ds = pb.make_raster(ds, out_fn, prediction, gdal.GDT_Byte, 0)
predict_ds.FlushCache()
levels = pb.compute_overview_levels(predict_ds.GetRasterBand(1))
predict_ds.BuildOverviews("NEAREST", levels)
# Apply the color table from the SWReGAP landcover raster.
gap_ds = gdal.Open(gap_fn)
colors = gap_ds.GetRasterBand(1).GetRasterColorTable()
predict_ds.GetRasterBand(1).SetRasterColorTable(colors)
del ds
import os
import numpy as np
from osgeo import gdal
import ospybook as pb
os.chdir(r'D:\osgeopy-data\Massachusetts')
in_fn = 'm_4207162_ne_19_1_20140718_20140923_clip.tif'
out_fn = 'ndvi.tif'
ds = gdal.Open(in_fn)
red = ds.GetRasterBand(1).ReadAsArray().astype(np.float)
nir = ds.GetRasterBand(4).ReadAsArray()
# Mask the red band.
red = np.ma.masked_where(nir + red == 0, red)
# Do the calculation.
ndvi = (nir - red) / (nir + red)
# Fill the empty cells.
ndvi = ndvi.filled(-99)
# Set NoData to the fill value when creating the new raster.
out_ds = pb.make_raster(
ds, out_fn, ndvi, gdal.GDT_Float32, -99)
overviews = pb.compute_overview_levels(out_ds.GetRasterBand(1))
out_ds.BuildOverviews('average', overviews)
del ds, out_ds
# Script to smooth an elevation dataset. import os import numpy as np from osgeo import gdal import ospybook as pb in_fn = r"D:\osgeopy-data\Nepal\everest.tif" out_fn = r'D:\Temp\everest_smoothed_edges.tif' in_ds = gdal.Open(in_fn) in_band = in_ds.GetRasterBand(1) in_data = in_band.ReadAsArray() # Stack the slices slices = pb.make_slices(in_data, (3, 3)) stacked_data = np.ma.dstack(slices) rows, cols = in_band.YSize, in_band.XSize # Initialize an output array to the NoData value (-99) out_data = np.ones((rows, cols), np.int32) * -99 # Put the result into the middle of the output, leaving the # outside rows and columns alone, so they still have -99. out_data[1:-1, 1:-1] = np.mean(stacked_data, 2) pb.make_raster(in_ds, out_fn, out_data, gdal.GDT_Int32, -99) del in_ds
data = pb.stack_bands(raster_fns)
# Sample the satellite data at the coordinates from the csv.
sample = data[rows, cols, :]
# Fit the classification tree.
clf = tree.DecisionTreeClassifier(max_depth=5)
clf = clf.fit(sample, classes)
# Apply the new classification tree model to the satellite data.
rows, cols, bands = data.shape
data2d = np.reshape(data, (rows * cols, bands))
prediction = clf.predict(data2d)
prediction = np.reshape(prediction, (rows, cols))
# Set the pixels with no valid satellite data to 0.
prediction[np.sum(data, 2) == 0] = 0
# Save the output.
predict_ds = pb.make_raster(ds, out_fn, prediction, gdal.GDT_Byte, 0)
predict_ds.FlushCache()
levels = pb.compute_overview_levels(predict_ds.GetRasterBand(1))
predict_ds.BuildOverviews('NEAREST', levels)
# Apply the color table from the SWReGAP landcover raster.
gap_ds = gdal.Open(gap_fn)
colors = gap_ds.GetRasterBand(1).GetRasterColorTable()
predict_ds.GetRasterBand(1).SetRasterColorTable(colors)
del ds
