-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.py
220 lines (189 loc) · 6.95 KB
/
utils.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
216
217
218
219
220
import os
import cv2
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from mpl_toolkits.mplot3d.art3d import Poly3DCollection
def get_frame(video_name, frame_dir):
os.makedirs(frame_dir, exist_ok=True)
vidcap = cv2.VideoCapture(video_name)
success, image = vidcap.read()
count = 0
while success:
cv2.imwrite("{}/frame{}.png".format(frame_dir, count), image) # save frame as PNG file
success, image = vidcap.read()
print('Read a new frame: ', success)
count += 1
def gen_video(frame_dir, video_name):
names = [name for name in os.listdir(frame_dir) if name.endswith('png')]
names = sorted(names, key=lambda name:int(name[5:].split('.')[0]))
frame = cv2.imread(os.path.join(frame_dir, names[0]))
height, width, _ = frame.shape
video = cv2.VideoWriter(video_name, fourcc=cv2.VideoWriter_fourcc(*'mp4v'), fps=30, frameSize=(width, height))
for name in names:
frame = cv2.imread(os.path.join(frame_dir, name))
video.write(frame)
video.release()
def parse_points3D(txt):
""" Parse points3D.txt from sparse reconstruction.
Return a numpy array of xyz coordinates, (n,3).
"""
xyz = []
with open(txt, 'r') as f:
for line in f.readlines():
if line.startswith('#'):
continue
tokens = line.strip().split()
x = float(tokens[1])
y = float(tokens[2])
z = float(tokens[3])
xyz.append([x, y, z])
return np.array(xyz)
def parse_cameras(txt):
""" Parse cameras.txt from sparse reconstruction.
Assume a single camera model:
SIMPLE_RADIAL params: f, cx, cy, k.
Return a python dict containing camara params.
"""
cameras = {}
with open(txt, 'r') as f:
for line in f.readlines():
if line.startswith('#'):
continue
tokens = line.strip().split()
CAMERA_ID = int(tokens[0])
cameras[CAMERA_ID] = {}
cameras[CAMERA_ID]['MODEL'] = tokens[1]
cameras[CAMERA_ID]['WIDTH'] = int(tokens[2])
cameras[CAMERA_ID]['HEIGHT'] = int(tokens[3])
cameras[CAMERA_ID]['PARAMS'] = [float(tokens[i]) for i in range(4, len(tokens))]
return cameras
def parse_images(txt):
""" Parse images.txt from sparse reconstruction.
Return a python dict containing camera pose for each image.
"""
keys = ['QW', 'QX', 'QY', 'QZ', 'TX', 'TY', 'TZ', 'CAMERA_ID', 'NAME']
images = {}
readline = False
with open(txt, 'r') as f:
for line in f.readlines():
if line.startswith('#') or not readline:
readline = True
else:
readline = False
tokens = line.strip().split()
IMAGE_ID = int(tokens[0])
images[IMAGE_ID] = {key:None for key in keys}
images[IMAGE_ID]['QW'] = float(tokens[1])
images[IMAGE_ID]['QX'] = float(tokens[2])
images[IMAGE_ID]['QY'] = float(tokens[3])
images[IMAGE_ID]['QZ'] = float(tokens[4])
images[IMAGE_ID]['TX'] = float(tokens[5])
images[IMAGE_ID]['TY'] = float(tokens[6])
images[IMAGE_ID]['TZ'] = float(tokens[7])
images[IMAGE_ID]['CAMERA_ID'] = int(tokens[8])
images[IMAGE_ID]['NAME'] = tokens[9]
return images
def detect_outliers_1d(data, threshold):
""" Detect outliers from a vector of data
:param data: 1d vector
:param threshold: determine if a point is outlier
:return: ids of data points
"""
mean = np.mean(data)
std = np.std(data)
ids = []
for i, d in enumerate(data):
z_score = (d - mean) / std
if abs(z_score) > threshold:
ids.append(i)
return np.array(ids)
def detect_outliers_3d(data, threshold):
""" Detect outliers from 3d points cloud
:param data: 3d numpy array, (n,3)
:param threshold: determine if a point is outlier
:return: ids of data points
"""
idx = detect_outliers_1d(data[:, 0], threshold)
idy = detect_outliers_1d(data[:, 1], threshold)
idz = detect_outliers_1d(data[:, 2], threshold)
ids = np.array(list(set(idx).union(set(idy), set(idz))))
return ids
def remove_outliers(data, threshold):
ids = detect_outliers_3d(data, threshold)
data_outliers = data[ids, :]
print('# outlier points:', data_outliers.shape)
mask = np.ones(len(data), np.bool)
mask[ids] = 0
data_inliers = data[mask]
print('# inlier points:', data_inliers.shape)
# # run this in a cell to check
# %matplotlib notebook
# plot3D(data_inliers, data_outliers, plot_plane=False)
return data_inliers
def add_plane(ax, model, x, y):
a, b, c, d = model
xmin, xmax = np.min(x), np.max(x)
xlen = xmax - xmin
xmin -= xlen * 0.2
xmax += xlen * 0.2
ymin, ymax = np.min(y), np.max(y)
ylen = ymax - ymin
ymin -= ylen * 0.2
ymax += ylen * 0.2
X = np.linspace(xmin, xmax, 10)
Y = np.linspace(ymin, ymax, 10)
X, Y = np.meshgrid(X, Y)
Z = -(a*X + b*Y + d) / c
ax.plot_surface(X, Y, Z)
def add_object(ax, corners, combinations):
for comb in combinations:
vertices = corners[comb, :]
ax.add_collection3d(Poly3DCollection([vertices]))
# ax.scatter3D(corners[:, 0], corners[:, 1], corners[:, 2], color='blue')
def plot3D(inplane_points, outplane_points,
plot_plane=False, model=None,
plot_box=False, plot_table=False, corners=None, combinations=None):
fig = plt.figure(1, figsize=(8, 8))
ax = fig.gca(projection='3d')
# plot outplane points
xo = outplane_points[:, 0]
yo = outplane_points[:, 1]
zo = outplane_points[:, 2]
ax.scatter(xo, yo, zo, marker='.', color='gray')
# plot inplane points
xi = inplane_points[:, 0]
yi = inplane_points[:, 1]
zi = inplane_points[:, 2]
zi=np.expand_dims(zi,axis=1)
ax.scatter(xi, yi, zi, marker='o', color='red')
# plot the surface
if plot_plane:
add_plane(ax, model, xi, yi)
# plot the box
if plot_box:
add_object(ax, corners, combinations)
# plot the table
if plot_table:
add_object(ax, corners, combinations)
ax.set_xlabel('X')
ax.set_ylabel('Y')
ax.set_zlabel('Z')
plt.show()
if __name__ == '__main__':
# video_name = './home/home.MOV'
# frame_dir = './home/frames'
# get_frame(video_name, frame_dir)
points3D_txt = './home/points3D.txt'
points3D = parse_points3D(points3D_txt)
print(points3D.shape)
print(np.min(points3D, axis=0))
print(np.max(points3D, axis=0))
print(np.mean(points3D, axis=0))
# cameras_txt = './home/cameras.txt'
# cameras = parse_cameras(cameras_txt)
# print(cameras)
# images_txt = './home/images.txt'
# images = parse_images(images_txt)
# print(images)
# print({k:len(v) for k,v in images.items()})