-
Notifications
You must be signed in to change notification settings - Fork 12
/
point_cloud_overlay_generation.py
215 lines (181 loc) · 10.2 KB
/
point_cloud_overlay_generation.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
'''
Author: Xingtong Liu, Yiping Zheng, Benjamin Killeen, Masaru Ishii, Gregory D. Hager, Russell H. Taylor, and Mathias Unberath
Copyright (C) 2020 Johns Hopkins University - All Rights Reserved
You may use, distribute and modify this code under the
terms of the GNU GENERAL PUBLIC LICENSE Version 3 license for non-commercial usage.
You should have received a copy of the GNU GENERAL PUBLIC LICENSE Version 3 license with
this file. If not, please write to: xliu89@jh.edu or unberath@jhu.edu
'''
import matplotlib
matplotlib.use('agg')
import cv2
import yaml
import numpy as np
from plyfile import PlyData
from pathlib import Path
import utils
import argparse
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Dense Descriptor Learning -- point cloud - video overlay generation',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument("--sequence_root", type=str, required=True, help='root of video sequence')
parser.add_argument("--display_image", action="store_true")
parser.add_argument("--display_visible_only", action="store_true")
parser.add_argument("--write_image", action="store_true")
parser.add_argument("--write_video", action="store_true")
args = parser.parse_args()
display_image = args.display_image
display_visible_only = args.display_visible_only
write_image = args.write_image
write_video = args.write_video
sequence_root = Path(args.sequence_root)
path_list = list(sequence_root.glob("colmap/*"))
path_list.sort()
num_points_per_seq = []
num_points_per_img = []
for prefix_seq in path_list:
print("Processing {}...".format(str(prefix_seq)))
if (prefix_seq / "point_cloud_overlay.avi").exists():
continue
# Read sparse point cloud from SfM
if not (prefix_seq / "structure.ply").exists():
continue
lists_3D_points = []
plydata = PlyData.read(str(prefix_seq / "structure.ply"))
for i in range(plydata['vertex'].count):
temp = list(plydata['vertex'][i])
temp = temp[:3]
temp.append(1.0)
lists_3D_points.append(temp)
lists_colors = [[255, 0, 0] for i in range(len(lists_3D_points))]
# Read camera poses from SfM
stream = open(str(prefix_seq / "motion.yaml"), 'r')
doc = yaml.load(stream)
keys, values = doc.items()
poses = values[1]
# Read indexes of visible views
visible_view_indexes = []
with open(str(prefix_seq / 'visible_view_indexes')) as fp:
for line in fp:
visible_view_indexes.append(int(line))
# Read view indexes per point
view_indexes_per_point = np.zeros((plydata['vertex'].count, len(visible_view_indexes)))
point_count = -1
with open(str(prefix_seq / 'view_indexes_per_point')) as fp:
for line in fp:
if int(line) == -1:
point_count = point_count + 1
else:
view_indexes_per_point[point_count][visible_view_indexes.index(int(line))] = 1
view_indexes_per_point = utils.overlapping_visible_view_indexes_per_point(
view_indexes_per_point, 1)
# Read camera intrinsics used by SfM
camera_intrinsics = []
param_count = 0
temp_camera_intrincis = np.zeros((3, 4))
with open(str(prefix_seq / 'camera_intrinsics_per_view')) as fp:
for line in fp:
if param_count == 0:
temp_camera_intrincis[0][0] = float(line)
param_count += 1
elif param_count == 1:
temp_camera_intrincis[1][1] = float(line)
param_count += 1
elif param_count == 2:
temp_camera_intrincis[0][2] = float(line)
param_count += 1
elif param_count == 3:
temp_camera_intrincis[1][2] = float(line)
temp_camera_intrincis[2][2] = 1.0
camera_intrinsics.append(temp_camera_intrincis)
temp_camera_intrincis = np.zeros((3, 4))
param_count = 0
# Generating projection and extrinsic matrices
projection_matrices = []
extrinsic_matrices = []
projection_matrix = np.zeros((3, 4))
for i in range(len(visible_view_indexes)):
rigid_transform = utils.quaternion_matrix(
[poses["poses[" + str(i) + "]"]['orientation']['w'], poses["poses[" + str(i) + "]"]['orientation']['x'],
poses["poses[" + str(i) + "]"]['orientation']['y'],
poses["poses[" + str(i) + "]"]['orientation']['z']])
rigid_transform[0][3] = poses["poses[" + str(i) + "]"]['position']['x']
rigid_transform[1][3] = poses["poses[" + str(i) + "]"]['position']['y']
rigid_transform[2][3] = poses["poses[" + str(i) + "]"]['position']['z']
transform = np.asmatrix(rigid_transform)
extrinsic_matrices.append(transform)
projection_matrix = np.dot(camera_intrinsics[0], transform)
projection_matrices.append(projection_matrix)
array_3D_points = np.asarray(lists_3D_points).reshape((-1, 4))
# Read mask image
img_mask = cv2.imread(str(prefix_seq / 'undistorted_mask.bmp'), cv2.IMREAD_GRAYSCALE)
img_mask = img_mask.reshape((-1, 1))
overlay_image_list = []
view_indexes_per_point = np.moveaxis(view_indexes_per_point, source=[0, 1], destination=[1, 0])
# Drawing 2D overlay of sparse point cloud onto every image plane
for i in range(len(visible_view_indexes)):
print("Process {}...".format(i))
img = cv2.imread(str(prefix_seq.parents[1] / ("{:08d}.jpg".format(visible_view_indexes[i]))))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
height, width = img.shape[:2]
projection_matrix = projection_matrices[i]
extrinsic_matrix = extrinsic_matrices[i]
points_3D_camera = np.einsum('ij,mj->mi', extrinsic_matrix, array_3D_points)
points_3D_camera = points_3D_camera / points_3D_camera[:, 3].reshape((-1, 1))
points_2D_image = np.einsum('ij,mj->mi', projection_matrix, array_3D_points)
points_2D_image = points_2D_image / points_2D_image[:, 2].reshape((-1, 1))
view_indexes_frame = np.asarray(view_indexes_per_point[i, :]).reshape((-1))
visible_point_indexes = np.where(view_indexes_frame > 0.5)
invisible_point_indexes = np.where(view_indexes_frame <= 0.5)
visible_point_indexes = visible_point_indexes[0]
invisible_point_indexes = invisible_point_indexes[0]
visible_points_2D_image = points_2D_image[visible_point_indexes, :].reshape((-1, 3))
invisible_points_2D_image = points_2D_image[invisible_point_indexes, :].reshape((-1, 3))
visible_points_3D_camera = points_3D_camera[visible_point_indexes, :].reshape((-1, 4))
invisible_points_3D_camera = points_3D_camera[invisible_point_indexes, :].reshape((-1, 4))
indexes = np.where((visible_points_2D_image[:, 0] <= width - 1) & (visible_points_2D_image[:, 0] >= 0) &
(visible_points_2D_image[:, 1] <= height - 1) & (visible_points_2D_image[:, 1] >= 0) &
(visible_points_3D_camera[:, 2] >= 0))
indexes = indexes[0]
in_image_point_1D_locations = (np.round(visible_points_2D_image[indexes, 0]) +
np.round(visible_points_2D_image[indexes, 1]) * width).astype(
np.int32).reshape((-1))
temp_mask = img_mask[in_image_point_1D_locations, :]
indexes_2 = np.where(temp_mask[:, 0] == 255)
indexes_2 = indexes_2[0]
visible_in_mask_point_1D_locations = in_image_point_1D_locations[indexes_2]
indexes = np.where((invisible_points_2D_image[:, 0] <= width - 1) & (invisible_points_2D_image[:, 0] >= 0) &
(invisible_points_2D_image[:, 1] <= height - 1) & (invisible_points_2D_image[:, 1] >= 0)
& (invisible_points_3D_camera[:, 2] > 0))
indexes = indexes[0]
in_image_point_1D_locations = (np.round(invisible_points_2D_image[indexes, 0]) +
np.round(invisible_points_2D_image[indexes, 1]) * width).astype(
np.int32).reshape((-1))
temp_mask = img_mask[in_image_point_1D_locations, :]
indexes_2 = np.where(temp_mask[:, 0] == 255)
indexes_2 = indexes_2[0]
invisible_in_mask_point_1D_locations = in_image_point_1D_locations[indexes_2]
visible_locations_y = list(visible_in_mask_point_1D_locations / width)
visible_locations_x = list(visible_in_mask_point_1D_locations % width)
invisible_locations_y = list(invisible_in_mask_point_1D_locations / width)
invisible_locations_x = list(invisible_in_mask_point_1D_locations % width)
img = utils.scatter_points_to_image(img, visible_locations_x=visible_locations_x,
visible_locations_y=visible_locations_y,
invisible_locations_x=invisible_locations_x,
invisible_locations_y=invisible_locations_y,
only_visible=display_visible_only,
point_size=1)
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
if write_video:
overlay_image_list.append(img)
if display_image:
cv2.imshow("projected spatial points", img)
cv2.waitKey(10)
if write_video:
fourcc = cv2.VideoWriter_fourcc(*'DIVX')
size = (overlay_image_list[0].shape[1], overlay_image_list[0].shape[0])
out = cv2.VideoWriter(str(prefix_seq / "point_cloud_overlay.avi"), fourcc, fps=20.0,
frameSize=size)
for image in overlay_image_list:
out.write(image)
out.release()