-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathgenerate_datasets.py
156 lines (134 loc) · 5.88 KB
/
generate_datasets.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
import os
import glob
import h5py
import numpy as np
from sklearn.neighbors import NearestNeighbors
from scipy.spatial.distance import minkowski
from scipy.spatial.transform import Rotation
def load_data(partition):
"""
读取h5文件中的data和label两个数据集到列表中
:param partition: h5文件名
:return: data数据列表和label数据列表
"""
DATA_DIR = 'data'
all_data = []
all_label = []
for h5_name in glob.glob(os.path.join(DATA_DIR, 'ply_data_%s*.h5' % partition)):
f = h5py.File(h5_name)
data = f['data'][:].astype('float32')
label = f['label'][:].astype('int64')
f.close()
all_data.append(data)
all_label.append(label)
all_data = np.concatenate(all_data, axis=0)
all_label = np.concatenate(all_label, axis=0)
return all_data, all_label
def translate_pointcloud(pointcloud):
"""
平移点云
:param pointcloud: 要平移的目标点云
:return: 平移之后的点云
"""
xyz1 = np.random.uniform(low=2. / 3., high=3. / 2., size=[3])
xyz2 = np.random.uniform(low=-0.2, high=0.2, size=[3])
translated_pointcloud = np.add(np.multiply(pointcloud, xyz1), xyz2).astype('float32')
return translated_pointcloud
def jitter_pointcloud(pointcloud, sigma=0.01, clip=0.05):
N, C = pointcloud.shape
pointcloud += np.clip(sigma * np.random.randn(N, C), -1 * clip, clip)
return pointcloud
def transformPointcloud(partial_cloud, rot_factor=4, d=0.5):
"""
:param d:
:param partial_cloud:
:param rot_factor:
:return:
"""
"""
rotation
"""
anglex = np.random.uniform() * np.pi / rot_factor # x轴的旋转角度
angley = np.random.uniform() * np.pi / rot_factor # y轴的旋转角度
anglez = np.random.uniform() * np.pi / rot_factor # z轴的旋转角度
cosx = np.cos(anglex)
cosy = np.cos(angley)
cosz = np.cos(anglez)
sinx = np.sin(anglex)
siny = np.sin(angley)
sinz = np.sin(anglez)
Rx = np.array([[1, 0, 0],
[0, cosx, -sinx],
[0, sinx, cosx]]) # 沿x轴旋转矩阵
Ry = np.array([[cosy, 0, siny],
[0, 1, 0],
[-siny, 0, cosy]]) # 沿y轴旋转矩阵
Rz = np.array([[cosz, -sinz, 0],
[sinz, cosz, 0],
[0, 0, 1]]) # 沿z轴旋转矩阵
R_ab = Rx.dot(Ry).dot(Rz) # 点云P到Q的旋转矩阵
R_ba = R_ab.T # 点云q到P旋转矩阵
"""
translation
"""
translation_ab = np.array([np.random.uniform(-d, d), np.random.uniform(-d, d),
np.random.uniform(-d, d)]) # 沿x， y， z的平移量 由p到q
translation_ba = -R_ba.dot(translation_ab) # 由q到p的平移
partial_cloud = partial_cloud.T
rotation_ab = Rotation.from_euler('zyx', [anglez, angley, anglex])
pointcloud_ = rotation_ab.apply(partial_cloud.T).T + np.expand_dims(translation_ab, axis=1) # 转换后的点云
euler_ab = np.asarray([anglez, angley, anglex]) # 三个轴旋转角度
euler_ba = -euler_ab[::-1]
return pointcloud_, translation_ba, euler_ba
def pairing(start_index, end_index):
data, label = load_data('train')
# transformed_partial_clouds = list()
translation_list = list()
rotation_list = list()
point_clouds_list = list()
transformed_point_clouds_list = list()
for cloud_index in range(start_index, end_index): # data.shape[0]
pointcloud = data[cloud_index][:1024]
pointcloud = translate_pointcloud(pointcloud)
point_clouds_list.append(pointcloud)
transformed_pointcloud, translation_ba, euler_ba = transformPointcloud(pointcloud)
transformed_pointcloud = transformed_pointcloud.T
transformed_point_clouds_list.append(transformed_pointcloud)
translation_list.append(translation_ba)
rotation_list.append(euler_ba)
return point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list#
def saveH5(point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list, file_path):
hdfFile = h5py.File(file_path, 'w')
translation_array = np.array(translation_list)
rotation_array = np.array(rotation_list)
hdfFile.create_dataset('point_clouds', data=np.array(point_clouds_list))
hdfFile.create_dataset('transformed_point_clouds', data=np.array(transformed_point_clouds_list))
hdfFile.create_dataset('translation', data=translation_array)
hdfFile.create_dataset('rotation', data=rotation_array)
hdfFile.close()
def readH5():
h5_name = "./data.h5"
f = h5py.File(h5_name)
partial_data = f.get("partialPointcloud_0")
complete_data = f.get("completePointcloud")
f.close()
def data_preprocess(partition):
# ximin
start_index_list = [0, 6, 11, 17, 23]
end_index_list = [5, 10, 16, 22, 28]
if partition=='train':
# ximin
# start_index_list = [0, 2048, 4096, 6144, 8192]
# end_index_list = [2048, 4096, 6144, 8192, 9840]
for h5_index in range(0, 5):
point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list = pairing(start_index_list[h5_index], end_index_list[h5_index])
# save H5files
saveH5(point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list, "trainData_{}.h5".format(str(h5_index)))
else:
# ximin
point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list = pairing(0,5)
# point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list = pairing(0,2048)
saveH5(point_clouds_list, transformed_point_clouds_list, translation_list, rotation_list,"testData_0.h5")
if __name__ == '__main__':
data_preprocess('train')
#readH5()