-
Notifications
You must be signed in to change notification settings - Fork 9
/
3DRadius_lm.py
234 lines (208 loc) · 9.49 KB
/
3DRadius_lm.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
221
222
223
224
225
226
227
228
229
230
231
232
233
234
import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
import open3d as o3d
import os
from numba import jit, prange
linemod_cls_names = ['ape','benchvise','cam','can','cat','driller','duck','eggbox','glue','holepuncher','iron','lamp','phone']
linemod_K = np.array([[572.4114, 0., 325.2611],
[0., 573.57043, 242.04899],
[0., 0., 1.]])
depthGeneration = False
linemod_path = "datasets/LINEMOD/"
original_linemod_path = "datasets/LINEMOD_ORIG/"
#IO function from PVNet
def project(xyz, K, RT):
"""
xyz: [N, 3]
K: [3, 3]
RT: [3, 4]
"""
#pointc->actual scene
xyz = np.dot(xyz, RT[:, :3].T) + RT[:, 3:].T
actual_xyz=xyz
#np.set_printoptions(threshold=np.inf)
#xyz = xyz[np.where(xyz[:,2]>=1)]
#distance_list = distance_list[np.where(xyz[:,2]>=1)]
#print(xyz)
#np.savetxt('GT.txt', actual_xyz*1000, delimiter=' ')
#os.system("pause")
#scene->image space
xyz = np.dot(xyz, K.T)
#np.set_printoptions(threshold=np.inf)
#print(xyz)
xy = xyz[:, :2] / xyz[:, 2:]
return xy,actual_xyz
def rgbd_to_point_cloud(K, depth):
vs, us = depth.nonzero()
zs = depth[vs, us]
#print(zs.min())
#print(zs.max())
xs = ((us - K[0, 2]) * zs) / float(K[0, 0])
ys = ((vs - K[1, 2]) * zs) / float(K[1, 1])
pts = np.array([xs, ys, zs]).T
return pts, vs, us
@jit(nopython=True, parallel=True)
def fast_for_map(yList, xList, xyz, distance_list, Radius3DMap):
for i in prange(len(xList)):
Radius3DMap[yList[i],xList[i]] = distance_list[i]
return Radius3DMap
#IO function from BOP toolbox
def linemod_pose(path, i):
"""
read a 3x3 rotation and 3x1 translation.
can be done with np.loadtxt, but this is way faster
@return R, t in [cm]
"""
R = open("{}/data/rot{}.rot".format(path, i))
R.readline()
R = np.float32(R.read().split()).reshape((3, 3))
t = open("{}/data/tra{}.tra".format(path, i))
t.readline()
t = np.float32(t.read().split())
return R, t
#IO function from BOP toolbox
def read_depth(path):
if (path[-3:] == 'dpt'):
with open(path) as f:
h,w = np.fromfile(f,dtype=np.uint32,count=2)
data = np.fromfile(f,dtype=np.uint16,count=w*h)
depth = data.reshape((h,w))
else:
depth = np.asarray(Image.open(path)).copy()
return depth
@jit(nopython=True, parallel=True)
def fast_for(pixel_coor, xy, actual_xyz, distance_list, Radius3DMap):
z_mean = np.mean(actual_xyz[:,2])
for coor in pixel_coor:
iter_count=0
z_loc = 0
z_min = 99999999999999999
for xy_single in xy:
if(coor[0]==xy_single[1] and coor[1]==xy_single[0]):
#print(coor)
#print(xy_single)
#print(actual_xyz[iter_count,2])
if(actual_xyz[iter_count,2]<z_min):
z_loc = iter_count
z_min = actual_xyz[iter_count,2]
#Radius3DMap[xy[z_loc][1],xy[z_loc][0]]=distance_list[z_loc]
iter_count+=1
if(z_min<=z_mean):
if depthGeneration:
Radius3DMap[xy[z_loc][1],xy[z_loc][0]]=actual_xyz[z_loc,2]
pre_z_loc = z_loc
else:
Radius3DMap[xy[z_loc][1],xy[z_loc][0]]=distance_list[z_loc]
pre_z_loc = z_loc
else:
if depthGeneration:
Radius3DMap[xy[z_loc][1],xy[z_loc][0]]=actual_xyz[pre_z_loc,2]
else:
Radius3DMap[xy[z_loc][1],xy[z_loc][0]]=distance_list[pre_z_loc]
return Radius3DMap
z_min = 999999999999999999
z_max = 0
depth_list=[]
if __name__=='__main__':
for class_name in linemod_cls_names:
print(class_name)
pcd_load = o3d.io.read_point_cloud(linemod_path+class_name+"/"+class_name+".ply")
xyz_load = np.asarray(pcd_load.points)
print(xyz_load)
#x_mean = np.mean(xyz_load[:,0])
#y_mean = np.mean(xyz_load[:,1])
#z_mean = np.mean(xyz_load[:,2])
keypoints=np.load(linemod_path+class_name+"/"+"Outside9.npy")
points_count = 1
for keypoint in keypoints:
#keypoint = keypoints[1]
print(keypoint)
x_mean = keypoint[0]
y_mean = keypoint[1]
z_mean = keypoint[2]
rootDict = original_linemod_path+class_name+"/"
GTDepthPath = rootDict+'FakeDepth/'
if depthGeneration:
saveDict = original_linemod_path+class_name+"/FakeDepth/"
else:
saveDict = original_linemod_path+class_name+"/Out_pt"+str(points_count)+"_dm/"
if(os.path.exists(saveDict)==False):
os.mkdir(saveDict)
points_count+=1
iter_count = 0
dataDict = rootDict + "data/"
for filename in os.listdir(dataDict):
if filename.endswith(".dpt"):
#and os.path.exists(saveDict+os.path.splitext(filename)[0][5:].zfill(6)+'.npy')==False
print(filename)
#depth = np.load(GTDepthPath+os.path.splitext(filename)[0][5:].zfill(6)+'.npy')*1000
realdepth = read_depth(dataDict+filename)
mask = np.asarray(Image.open(linemod_path+class_name+"/mask/"+os.path.splitext(filename)[0][5:].zfill(4)+".png"), dtype=int)
mask = mask[:,:,0]
#depth[np.where(mask==0)] = 0
realdepth[np.where(mask==0)] = 0
#plt.imshow(realdepth-depth*1000)
#plt.show()
Radius3DMap = np.zeros(mask.shape)
RT = np.load(linemod_path+class_name+"/pose/pose"+os.path.splitext(filename)[0][5:]+".npy")
print(RT)
print(linemod_pose(rootDict,os.path.splitext(filename)[0][5:]))
pixel_coor = np.argwhere(mask==255)
xyz,y,x = rgbd_to_point_cloud(linemod_K, realdepth)
print(xyz)
print(RT)
dump, transformed_kpoint = project(np.array([keypoint]),linemod_K,RT)
transformed_kpoint = transformed_kpoint[0]*1000
print(transformed_kpoint)
distance_list = ((xyz[:,0]-transformed_kpoint[0])**2+(xyz[:,1]-transformed_kpoint[1])**2+(xyz[:,2]-transformed_kpoint[2])**2)**0.5
Radius3DMap = fast_for_map(y, x, xyz, distance_list, Radius3DMap)
#xy, actual_xyz=project(xyz_load,linemod_K,RT)
#dump,kpt = project(np.expand_dims(keypoint,axis=0),linemod_K,RT)
#distance_list = ((actual_xyz[:,0]-kpt[0,0])**2+(actual_xyz[:,1]-kpt[0,1])**2+(actual_xyz[:,2]-kpt[0,2])**2)**0.5
#xy = np.around(xy).astype(int)
#xy = xy.astype(int)
#xy = xy[np.where(xy[:,0]>0)]
#xy = xy[np.where(xy[:,1]>0)]
#distance_list = distance_list[np.where(xy[:,0]>0)]
#distance_list = distance_list[np.where(xy[:,1]>0)]
#Radius3DMap = fast_for(pixel_coor, xy, actual_xyz, distance_list, Radius3DMap)
#if(iter_count==0):
# depth_list = Radius3DMap[Radius3DMap.nonzero()]
#else:
# depth_list = np.append(depth_list,Radius3DMap[Radius3DMap.nonzero()])
iter_count+=1
#print(Radius3DMap[Radius3DMap.nonzero()])
plt.imshow(Radius3DMap)
plt.show()
mean = 0.84241277810665
std = 0.12497967663932731
#Radius3DMap[Radius3DMap.nonzero()] = (Radius3DMap[Radius3DMap.nonzero()]-mean)/std
#depth = depth/1000
#Radius3DMap[Radius3DMap.nonzero()] = (Radius3DMap[Radius3DMap.nonzero()] - min(depth[depth.nonzero()])) / (max(depth[depth.nonzero()])-min(depth[depth.nonzero()]))
#plt.imshow(Radius3DMap)
#plt.show()
#plt.imshow(np.where(Radius3DMap*100>0,1,0))
#plt.show()
if depthGeneration:
np.save(saveDict+os.path.splitext(filename)[0][5:].zfill(6)+'.npy',Radius3DMap)
else:
np.save(saveDict+os.path.splitext(filename)[0][5:].zfill(6)+'.npy',Radius3DMap*10)
#Radius3DMap[xy[:,1],xy[:,0]]=distance_list
#display= depth-Radius3DMap*10
#if(Radius3DMap.min()<z_min):
# z_min = Radius3DMap.min()
if(Radius3DMap.max()>z_max):
z_max = Radius3DMap.max()
#Radius3DMap = (Radius3DMap - Radius3DMap[Radius3DMap.nonzero()].min())/(Radius3DMap.max()-Radius3DMap[Radius3DMap.nonzero()].min())
#Radius3DMap = np.where(Radius3DMap!=0,Radius3DMap*1000,0)
#np.save(os.path.splitext(filename)[0][5:].zfill(6)+'.npy',Radius3DMap)
#plt.imshow(Radius3DMap*10)
#plt.show()
if depthGeneration:
break
#os.system('pause')
#print(class_name+" mean: ", np.mean(np.asarray(depth_list)))
#print(class_name+" std: ", np.std(np.asarray(depth_list)))
#print("min: ", z_min)
#print("max: ", z_max)