Gkit

Gkit is a suite of utilities for processing geo-dataset allowing you to manipulate raster datasets like np.MaskedArray.

Gkit supports Python3 and all OS which could install numpy, matplotlib, gdal.

In Python2, may have unpredictable bugs.

Documents: https://titorx.github.io/gkit/html/

Simple Example

Here is examples of some basic features that Gkit provides.

import numpy as np
import gkit as gk

# Read the first band from .tif
r = gk.read("lst.tif")

# You could also specific point out which band you want to load
r = gk.read("lst.tif", 2)

# gk.read returns a Raster class
type(r)
# Output:
# gkit.core.Raster

# Open an interactive window display raster using matplotlib(call plt.show)
r.show()

The screenshot:

# Only draw raster without calling plt.show.
import matplotlib.pylab as plt
r.plot()
plt.xlabel("Lon")
plt.ylabel("Lat")
plt.title("LST(K)")
plt.colorbar()
plt.savefig("raster_plot.png")

The screenshot:

# Raster class inherits from np.ma.MaskedArray.
# It has all features which MaskedArray has.
r
# Output:
# masked_array(data =
#  [[-- -- -- ..., -- -- --]
#  [-- -- -- ..., -- -- --]
#  [-- -- -- ..., -- -- --]
#  ...,
#  [242.5966339111328 242.6825408935547 242.79612731933594 ...,
#   243.512451171875 243.46498107910156 243.45751953125]
#  [241.1952667236328 241.18592834472656 241.19235229492188 ...,
#   241.02757263183594 241.04196166992188 241.0919189453125]
#  [241.97023010253906 242.03948974609375 242.05393981933594 ...,
#   241.8543243408203 241.85800170898438 241.80813598632812]],
#              mask =
#  [[ True  True  True ...,  True  True  True]
#  [ True  True  True ...,  True  True  True]
#  [ True  True  True ...,  True  True  True]
#  ...,
#  [False False False ..., False False False]
#  [False False False ..., False False False]
#  [False False False ..., False False False]],
#        fill_value = 1e+20)

# Doing operation like common numpy masked array.
tmp = (r - 273.15)**3 / 4
tmp = np.cos(r)
tmp = np.abs(r)
tmp = np.sqrt(r)

print(r.shape)
print(r.mean())
print(r.max())
print(r.min())

# convert data type
tmp = r.astype(np.float64)

# Save to file
r.save("out_file.tif")

# Create a raster from numpy array
import numpy as np
x, y = np.mgrid[-1:1:100j, -2:2:200j]
array = np.sqrt(x**2 + y**2)

print(array.shape)
# Output:
# (100, 200)

transform = [-100, 0.1, 0, 0, 0, -0.1]

raster = gk.Raster(array, transform)
raster.show()

The screenshot: