-
Notifications
You must be signed in to change notification settings - Fork 0
/
euclidean_distance.py
51 lines (41 loc) · 1.53 KB
/
euclidean_distance.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
#!/usr/bin/env python
# =============================================================================
# Date: June, 2019
# Author: Marcelo Villa P.
# Purpose: Creates a raster with the (pixel) distance to the closest target.
# Notes: Own implementation of distance calculation using NumPy.
# =============================================================================
import gdal
import numpy as np
from helper_functions import array_to_tiff
def euclidean_distance(arr, nd_value):
"""
Computes the euclidean distance to all non-NoData values in arr.
:param arr: raster 2D numpy array
:param nd_value: raster´s NoData value
:return: distance 2D numpy array
"""
# create meshgrid
y, x = arr.shape
xx, yy = np.meshgrid(np.arange(x), np.arange(y))
# create indices where arr is different from NoData and reshape them
ind = np.nonzero((arr != nd_value))
ix = ind[1].reshape((-1, 1, 1))
iy = ind[0].reshape((-1, 1, 1))
# compute legs
dx = np.abs(iy - yy)
dy = np.abs(ix - xx)
return np.min(np.hypot(dx, dy), axis=0)
# open raster, get raster information and read data
fn = '../data/raster/COL_rails.tif'
ds = gdal.Open(fn, 0)
gt = ds.GetGeoTransform()
sr = ds.GetProjection()
arr = ds.ReadAsArray()
nd_value = ds.GetRasterBand(1).GetNoDataValue()
# compute euclidean distance
d = euclidean_distance(arr, nd_value)
# create output raster
out_fn = '../data/raster/COL_rails_distance.tif'
nd_value = -99
array_to_tiff(d, out_fn, sr, gt, gdal.GDT_Float32, nd_value)