-
Notifications
You must be signed in to change notification settings - Fork 1
/
Main.py
215 lines (145 loc) · 6.39 KB
/
Main.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
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
"""Problem Set 3: Geometry."""
import numpy as np
import cv2
import os
from ps3 import *
input_dir = "input"
output_dir = "output"
def normalize_and_scale(img_in):
"""Maps values in img_in to fit in the range [0, 255]. This will be usually called before displaying or
saving an image.
Args:
img_in (numpy.array): input image.
Returns:
numpy.array: output image with integer pixel values in [0, 255]
"""
return cv2.normalize(img_in, alpha=0, beta=255, norm_type=cv2.NORM_MINMAX).astype(np.uint8)
def part_1a():
l = cv2.imread(os.path.join(input_dir, 'pair0-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair0-R.png'), 0) / 255.
w_size = (5, 5)
dmax = 10
d_l = disparity_ssd(l, r, 0, w_size, dmax)
d_r = disparity_ssd(l, r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-1-a-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-1-a-2.png'), d_r)
def part_1b():
l = cv2.imread(os.path.join(input_dir, 'pair1-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair1-R.png'), 0) / 255.
w_size = (10, 10)
dmax = 100
d_l = disparity_ssd(l, r, 0, w_size, dmax)
d_r = disparity_ssd(l, r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
plt.imshow(d_l,'gray')
plt.imshow(d_r,'gray')
cv2.imwrite(os.path.join(output_dir, 'ps3-1-b-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-1-b-2.png'), d_r)
def part_2a(get_disp=True):
l = cv2.imread(os.path.join(input_dir, 'pair1-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair1-R.png'), 0) / 255.
sigma = 0.1 # You may have to try different values
l_noisy = add_noise(l, sigma)
r_noisy = add_noise(r, sigma)
# Select one or both noisy images and use SSD
image_l = l_noisy
image_r = r
if get_disp:
w_size = (10, 10)
dmax = 100
d_l = disparity_ssd(image_l, image_r, 0, w_size, dmax)
d_r = disparity_ssd(image_l, image_r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-2-a-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-2-a-2.png'), d_r)
return image_l, image_r
def part_2b(get_disp=True):
l = cv2.imread(os.path.join(input_dir, 'pair1-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair1-R.png'), 0) / 255.
value = 10. # percent (%).
image_to_boost = l # Replace None with either l or r
contrast_img = increase_contrast(image_to_boost, value)
# TODO: Change the following two lines accordingly
image_l = contrast_img # Can be either L or contrast_img if l was used
image_r = r # Can be either R or contrast_img if r was used
if get_disp:
w_size = (10, 10) # You may have to try different values
dmax = 100 # You may have to try different values
d_l = disparity_ssd(image_l, image_r, 0, w_size, dmax)
d_r = disparity_ssd(image_l, image_r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-2-b-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-2-b-2.png'), d_r)
return image_l, image_r # These will be used in 3b
def part_3a():
l = cv2.imread(os.path.join(input_dir, 'pair1-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair1-R.png'), 0) / 255.
w_size = (10, 10) # You may have to try different values
dmax = 100 # You may have to try different values
d_l = disparity_ncorr(l, r, 0, w_size, dmax)
d_r = disparity_ncorr(l, r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-a-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-a-2.png'), d_r)
def part_3b_1():
image_l, image_r = part_2a(False) # Here we use the same images selected for 2a
w_size = (10, 10) # You may have to try different values
dmax = 100 # You may have to try different values
d_l = disparity_ncorr(image_l, image_r, 0, w_size, dmax)
d_r = disparity_ncorr(image_l, image_r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-b-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-b-2.png'), d_r)
def part_3b_2():
image_l, image_r = part_2b(False) # Here we use the same images selected for 2b
w_size = (5, 5) # You may have to try different values
dmax = 100 # You may have to try different values
d_l = disparity_ncorr(image_l, image_r, 0, w_size, dmax)
d_r = disparity_ncorr(image_l, image_r, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-b-3.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-3-b-4.png'), d_r)
def part_4():
"""Applies the methods used in previous parts on the images pair2-L and pair2-R to obtain disparity images close
to the ground truth. Here you are encouraged try image processing steps and SSD or normalized correlation.
Use the images pair2-D_L and pair2-D_R as a reference.
The images to save are:
- ps3-4-a-1.png
- ps3-4-a-2.png
Returns:
None.
"""
l = cv2.imread(os.path.join(input_dir, 'pair2-L.png'), 0) / 255.
r = cv2.imread(os.path.join(input_dir, 'pair2-R.png'), 0) / 255.
ls = cv2.GaussianBlur(l, (21,21), 0)
rs = cv2.GaussianBlur(r, (21,21), 0)
kernel_sharpen_1 = np.array([[-1,-1,-1], [-1,9,-1], [-1,-1,-1]])
lsp = cv2.filter2D(l, -1, kernel_sharpen_1)
rsp = cv2.filter2D(r, -1, kernel_sharpen_1)
w_size = (5,5)
dmax = 100
d_l = disparity_ssd(lsp, rsp, 0, w_size, dmax)
d_r = disparity_ssd(lsp, rsp, 1, w_size, dmax)
d_l = normalize_and_scale(d_l)
d_r = normalize_and_scale(d_r)
cv2.imwrite(os.path.join(output_dir, 'ps3-4-a-1.png'), d_l)
cv2.imwrite(os.path.join(output_dir, 'ps3-4-a-2.png'), d_r)
return image_l, image_r # These will be used in 3b
if __name__ == '__main__':
part_1a()
part_1b()
part_2a()
part_2b()
part_3a()
part_3b_1()
part_3b_2()
part_4()
# TODO: Don't forget to answer part 5 in your report.