None

Note

This tutorial was generated from an IPython notebook that can be downloaded here.

Usage example

import focal_stats
import rasterio as rio
import matplotlib.pyplot as plt
import numpy as np
import os

Matplotlib is building the font cache; this may take a moment.

os.chdir("../../../")

Loading raster (containing water table depth (Fan et al., 2017)).

with rio.open("data/wtd.tif") as f:
    a = f.read(1).astype(np.float64)
    a[a == -999.9] = np.nan

Inspecting the data

plt.imshow(a, cmap='Blues', vmax=100)
plt.title("Water table depth")
plt.colorbar()

<matplotlib.colorbar.Colorbar at 0x7ff3c4aec550>
../_images/tutorial_6_1.png

Focal statistics

Calculation of the focal mean:

plt.imshow(focal_stats.focal_mean(a, window_size=15), vmax=100, cmap="Blues")

<matplotlib.image.AxesImage at 0x7ff3c4184b50>
../_images/tutorial_9_1.png

This looks quite similar to the input raster, but with smoothing applied. Let’s try a higher window_size, which should increase the smoothing

plt.imshow(focal_stats.focal_mean(a, window_size=101), vmax=100, cmap="Blues")

<matplotlib.image.AxesImage at 0x7ff3c4112c90>
../_images/tutorial_11_1.png

This same functionality can be used to reduce the shape of the raster based on this window_size.

x = focal_stats.focal_mean(a, window_size=108, reduce=True)
plt.imshow(x, vmax=100, cmap="Blues")

<matplotlib.image.AxesImage at 0x7ff3c4084b90>
../_images/tutorial_13_1.png

The shape of this new raster is exactly 108 times smaller than the input raster. Note that for this to work both x and y-axes need to be divisible by the window size.