-
Notifications
You must be signed in to change notification settings - Fork 1
/
find_cliffs.py
192 lines (161 loc) · 7.51 KB
/
find_cliffs.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
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
import os
import sys
import glob
import numpy as np
import rasterio
import multiprocessing
import cv2
from tqdm import tqdm
from random import shuffle
from pyproj import Geod
import shutil
import matplotlib
import matplotlib.pyplot as plt
from matplotlib.colors import LightSource
# lat, lon, blackout area radius
blacklist = [[27.822736740112305, 92.57917785644531, 20000],
[27.879862, 93.0008, 20000],
[28.443410873413086, 85.11734771728516, 20000],
[29.648374557495117, 101.93683624267578, 15000],
[5.212995529174805, -76.18224334716797, 10000],
[59.947605, 6.6277957, 2000],
[71.83742, -52.464878, 10000],
[71.418915, -51.742565, 10000],
[75.57478, -21.901924, 10000],
[-12.0806055, -74.27191, 2000],
[-49_05573, -74_76408, 10000],
[-49.784317 , -74.62307, 5000],
[-61.090355, -54.72234, 10000],
[-66_79399, 163.20633, 10000],
[29.65848541, 95.11452484,1000],
[29.63715744, 95.06532288, 1000],
[71.56819153, -52.0521698, 1000],
[-43.61115265, 170.07957458, 1000],
[27.70921326, 97.11391449, 1000],
[-67.37149811, 164.75134277, 10000]]
METERS_PER_PIXEL = 30
MAX_LAT = 85
MIN_DROP = 400
GRADE = 3
ROTATE_ANGLE_DEGREES = 30
IMAGE_SAVE_PAD = 200
MIN_SPACING_METERS = 3000
def in_blacklist(lat, lon):
for black in blacklist:
if calculate_distance(lat, lon, black[0], black[1]) < black[2]:
return True
return False
def save_topo(file_name, topo_array, cmap=matplotlib.cm.gist_earth, azdeg=315, altdeg=45):
ls = LightSource(azdeg=azdeg, altdeg=altdeg)
shaded = ls.shade(topo_array, cmap=cmap, vert_exag=0.1, blend_mode='overlay')
plt.figure()
plt.imshow(shaded, aspect='auto')
plt.axis('off') # Hide axes and ticks
plt.savefig(file_name, bbox_inches='tight', pad_inches=0)
plt.close()
def pad_array(array, num_rows, num_cols):
padded_array = np.pad(array, ((num_rows, num_rows), (num_cols, num_cols)), mode='constant', constant_values=0)
return padded_array
def calculate_distance(lat1, lon1, lat2, lon2):
geod = Geod(ellps='WGS84')
_, _, dist = geod.inv(lon1, lat1, lon2, lat2)
return dist
def rotate(image, angle):
(h, w) = image.shape[:2]
center = (w / 2, h / 2)
M = cv2.getRotationMatrix2D(center, angle, scale=1.0)
rotated = cv2.warpAffine(image, M, (w, h), borderValue=np.nan)
return rotated
def get_image(lat, lon, array, coord_array):
min_dist = np.nanargmin(np.linalg.norm(coord_array - np.array([lat, lon]), axis=-1))
row, col = np.unravel_index(min_dist, array.shape)
array = pad_array(array, IMAGE_SAVE_PAD, IMAGE_SAVE_PAD)
row = row + IMAGE_SAVE_PAD
col = col + IMAGE_SAVE_PAD
return array[row-IMAGE_SAVE_PAD:row+IMAGE_SAVE_PAD, col-IMAGE_SAVE_PAD:col+IMAGE_SAVE_PAD]
def find_max_diff(file_path):
repo_dir = os.path.dirname(__file__)
output_dir = repo_dir + "/cliffs_found/"
with rasterio.open(file_path) as src:
original_array = src.read(1)
bounds = src.bounds
# High latitudes have problems
if abs(bounds.top) > MAX_LAT or abs(bounds.bottom) > MAX_LAT:
return None
if np.nanmax(original_array) - np.nanmin(original_array) < MIN_DROP:
return None
w_meters = calculate_distance(bounds.bottom, bounds.left, bounds.bottom, bounds.right)
h_meters = calculate_distance(bounds.bottom, bounds.left, bounds.top, bounds.left)
width_pixels = int(w_meters / METERS_PER_PIXEL)
height_pixels = int(h_meters / METERS_PER_PIXEL)
array = cv2.resize(np.copy(original_array), (width_pixels, height_pixels), interpolation=cv2.INTER_LINEAR)
grid = np.meshgrid(np.linspace(bounds.top, bounds.bottom, original_array.shape[0], dtype=np.float32),
np.linspace(bounds.left, bounds.right, original_array.shape[1], dtype=np.float32), indexing='ij')
original_coord_array = np.stack(grid, axis=-1)
grid = np.meshgrid(np.linspace(bounds.top, bounds.bottom, array.shape[0], dtype=np.float32),
np.linspace(bounds.left, bounds.right, array.shape[1], dtype=np.float32), indexing='ij')
coord_array = np.stack(grid, axis=-1)
desired_size = max(width_pixels, height_pixels) * 1.45
pad_rows = int((desired_size - height_pixels) / 2)
pad_cols = int((desired_size - width_pixels) / 2)
array = np.pad(array, ((pad_rows, pad_rows), (pad_cols, pad_cols)), mode='constant', constant_values=np.nan)
coord_array = np.pad(coord_array, ((pad_rows, pad_rows), (pad_cols, pad_cols), (0,0)), mode='constant', constant_values=np.nan)
cliff_height = MIN_DROP
test_runs = 0
all_cliffs = []
qualifying_cliffs = None
while qualifying_cliffs is None or len(qualifying_cliffs) > 0:
qualifying_cliffs = []
for angle in range(0,180,ROTATE_ANGLE_DEGREES):
array_rot = rotate(array.copy(), angle)
coord_array_rot = rotate(coord_array.copy(), angle)
LAT_DISTANCE = cliff_height / GRADE
LAT_PIXELS = int(LAT_DISTANCE / METERS_PER_PIXEL)
diff = abs(np.column_stack([0.0*array_rot[:,:LAT_PIXELS//2], array_rot[:,LAT_PIXELS:] - array_rot[:,:-LAT_PIXELS], 0.0*array_rot[:,-(LAT_PIXELS-LAT_PIXELS//2):]]))
while np.nanmax(diff) > cliff_height:
argmax_flat = np.nanargmax(diff)
row, col = np.unravel_index(argmax_flat, diff.shape)
lat, lon = coord_array_rot[row, col]
qualifying_cliffs.append([diff[row, col], lat, lon, angle, original_array, original_coord_array])
blackout_pixels = (MIN_SPACING_METERS / METERS_PER_PIXEL) //2
diff[ min(0, row - blackout_pixels): max(row + blackout_pixels, diff.shape[0]),
min(0, col - blackout_pixels): max(col + blackout_pixels, diff.shape[0])] = 0.0
test_runs += 1
if test_runs > 100:
print(f"Test runs: {test_runs} on {file_path}, {cliff_height}m {len(all_cliffs)} cliffs found, might be hitting infinite loop")
all_cliffs.extend(qualifying_cliffs)
cliff_height += 100
all_cliffs.sort(key=lambda x: x[0], reverse=True)
for i, cliff in enumerate(all_cliffs):
drama_factor, lat, lon, angle, original_array, original_coord_array = cliff
# check if not too close to other cliffs
too_close = False
for other_cliff in all_cliffs[:i]:
_, other_lat, other_lon, _, _, _ = other_cliff
if calculate_distance(lat, lon, other_lat, other_lon) < MIN_SPACING_METERS:
too_close = True
if too_close or in_blacklist(lat, lon):
continue
else:
print(f'Found {drama_factor}m diff at {lat}, {lon} in {file_path} at angle {angle}')
lat_str = str(lat).replace(".", "_")
lon_str = str(lon).replace(".", "_")
image = get_image(lat, lon, original_array, original_coord_array)
save_topo(f"{output_dir}lat_{lat_str}_lon_{lon_str}.png", np.nan_to_num(image))
with open(f'{output_dir}full_list.txt', 'a') as the_file:
the_file.write(f'{drama_factor}, {lat}, {lon}\n')
return
def main():
repo_dir = os.path.dirname(__file__)
output_dir = repo_dir + "/cliffs_found/"
topo_database_dir = "/home/batman/Documents/global_topo_copernicus/"
shutil.rmtree(output_dir, ignore_errors=True)
os.makedirs(output_dir, exist_ok=True)
print(f"Starting to find cliffs in {output_dir}")
all_tifs = glob.glob(topo_database_dir + "*/*.tif")
shuffle(all_tifs)
num_processes = 20
with multiprocessing.Pool(processes=num_processes) as pool:
results = list(tqdm(pool.imap(find_max_diff, all_tifs), total=len(all_tifs)))
if __name__ == "__main__":
main()