-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathnew.py
212 lines (183 loc) · 6.72 KB
/
new.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
#coding:utf-8
from renderer import Renderer
#from utils import
import numpy as np
import cv2
from PIL import Image
from utils import precompute_projections,create_pose
from os.path import join
from model import Model3D
from os import listdir
from os.path import isfile, join
nb_bg_images = 1500 # JUST FOR TEST
#nb_bg_images = 150000 # USE THIS NUMBER TO TRAIN BB8
# Image window size
def pnp(points_3D, points_2D, cameraMatrix):
try:
distCoeffs = pnp.distCoeffs
except:
distCoeffs = np.zeros((8, 1), dtype='float32')
assert points_2D.shape[0] == points_2D.shape[0], 'points 3D and points 2D must have same number of veritces'
_, R_exp, t = cv2.solvePnP(points_3D,
points_2D,
cameraMatrix,
distCoeffs)
R, _ = cv2.Rodrigues(R_exp)
Rt = np.c_[R, t]
return Rt
#search corner
c = 3
h = 480
w = 640
background_images = np.zeros((nb_bg_images, h, w, c), dtype='uint8')
bg_path = '/home/sky/Project/BB8-J15-master/data/BG/ImageNet'
rgb_sy = np.zeros(( h, w, c), dtype='uint8')
# Print iterations progress
def printProgressBar (iteration, total, prefix = '', suffix = '', decimals = 1, length = 100, fill = '='):
import sys
"""
Call in a loop to create terminal progress bar
@params:
iteration - Required : current iteration (Int)
total - Required : total iterations (Int)
prefix - Optional : prefix string (Str)
suffix - Optional : suffix string (Str)
decimals - Optional : positive number of decimals in percent complete (Int)
length - Optional : character length of bar (Int)
fill - Optional : bar fill character (Str)
"""
percent = ("{0:." + str(decimals) + "f}").format(100 * (iteration / float(total)))
filledLength = int(length * iteration // total)
bar = fill * filledLength + ' ' * (length - filledLength)
sys.stdout.write('\r%s |%s| %s%% %s' % (prefix, bar, percent, suffix))
# Print New Line on Complete
if iteration == total:
print()
def get_all_files(directory):
files = []
for f in listdir(directory):
if isfile(join(directory, f)):
files.append(join(directory, f))
else:
files.extend(get_all_files(join(directory, f)))
return files
def load_bgs_to_memory():
"""
Load background images to the memory.
"""
bg_file_names = get_all_files(bg_path)
print("{0} background images are found".format(len(bg_file_names)))
print('loading background images to memory...')
bg_nb, bg_h, bg_w = background_images.shape[:3]
for bg_idx in range(bg_nb):
random_bg_index = np.random.randint(0, len(bg_file_names) - 1)
bg = cv2.imread(bg_file_names[random_bg_index])
bg = cv2.resize(bg, (int(bg_w), int(bg_h)), interpolation=cv2.INTER_NEAREST)
background_images[bg_idx] = bg
printProgressBar(bg_idx, bg_nb, prefix = 'Progress:', suffix = 'Complete', length = 50)
printProgressBar(bg_nb, bg_nb, prefix = 'Progress:', suffix = 'Complete', length = 50)
#随机的产生两个角度，就可以唯一确定viewpoint,不不不，角度的，在球面不均匀
import math
#step1：
# 随机抽样产生一对均匀分布的随机数 u ，v ；这里u，v 在[-1,1] 范围内
# step2 ：
# 计算 r^2 = u^2+v^2;
# 如果 r^2 > 1 则重新抽样，直到满足 r^2 < 1 .
# step3 ：
# 计算
# x=2*u*sqrt(1-r2);
# y=2*v*sqrt(1-r2);
# z=1-2*r2;
ViewPoint = []
u_v = np.asarray([[0,0,1],[0.3, 0.7416 , 0.6],[1,0,0]])
ViewPoint.append(u_v[0])
ViewPoint.append(u_v[1])
class bench:
def __init__(self):
self.cam = np.identity(3)
self.models = {}
Model_instance = bench()
Model_instance.cam[0, 0] = 680.0
Model_instance.cam[0, 2] = 320.0
Model_instance.cam[1, 1] = 680.0
Model_instance.cam[1, 2] = 240.0
Model_instance.cam[2, 2] = 1.0
#model file name
model_name = '2.ply'
Model_instance.models[model_name] = Model3D()
Model_instance.scale_to_meters = 0.01
#camera information
Model_instance.models[model_name].diameter = 125.64433088
Model_instance.models[model_name].load(model_name,scale = Model_instance.scale_to_meters)
views = [0.6, 0616. , 0.6]
np.set_printoptions(precision=3, suppress=True)
import os
w = 640
h = 480
#主要存储的文件，mask文件，可以根据深度值～=0,pose真值R,T，合成的RGB图像，bb8：8个顶点值
i = 0
ren = Renderer((w, h), Model_instance.cam)
ren1 = Renderer((w, h), Model_instance.cam)
for vp,views in enumerate(ViewPoint):
#print(views)
views = np.asarray(views)
pose = create_pose(views, angle_deg=0)
zr = 0.2*np.random.rand(1,1) + 0.7
#T = np.asarray([0, 0, zr]).T
#Translate = np.matmul(pose[:3,:3],T)
#pose1 = pose
#pose1[:3, 3] = Translate
pose[:3, 3] = [0.0, 0., zr] # zr = 0.5
print('pose{}'.format(pose))
#print('pose1{}'.format(pose1))
ren.draw_model(Model_instance.models[model_name],pose)
ren.draw_boundingbox(Model_instance.models[model_name], pose)
rgb,mask = ren.finish()
ren.clear()
ren1.draw_model(Model_instance.models[model_name], pose)
#ren.cal_bb8(Model_instance.models[model_name],pose)
rgb1, mask1 = ren1.finish()
mask1 = (mask1 > 0)
#cv2.imshow("name",mask*255)
#cv2.imshow("rgb", rgb)
#cv2.waitKey(0)
cv2.imwrite("3.png",rgb*255)
cv2.imwrite("mask.png", mask1 * 255)
ren1.clear()
np.set_printoptions(precision=3, suppress=True)
def get_Rt(filename):
Rt = np.loadtxt(filename,delimiter=',', dtype='float32')[:3, :]
return Rt
#Rt = get_Rt('0001.txt')
#print('Rt:',Rt)
Rt = pose
def get_camera_intrinsic():
K = np.zeros((3, 4), dtype='float32')
K[0, 0], K[0, 2] = 680, 320
K[1, 1], K[1, 2] = 680, 240
K[2, 2] = 1.
return K
K = np.matmul(get_camera_intrinsic(),Rt)
Pt = np.loadtxt('J15_bb.txt', dtype='float32')[:, :]
Pts = np.zeros((8,4),dtype='float32')
Pts[:,0:3] = Pt
print('bb8:',Pts)
print(K)
pt = np.matmul(K,Pts.T)
for idx in range(0,8,1):
pt[0,idx] = pt[0,idx]/pt[2,idx]
pt[1,idx] = pt[1,idx]/pt[2,idx]
pt = pt[0:2,:]
print('pt:',pt)
import cv2
im = cv2.imread("3.png")
for idx in range(0,8,1):
cv2.circle(im,(int(pt[0,idx]),int(pt[1,idx])),4,(0,255,0),-1)
cv2.imshow("projection",im)
cv2.waitKey(0)
def compute_projection(points_3D, transformation, internal_calibration):
projections_2d = np.zeros((2, points_3D.shape[1]), dtype='float32')
camera_projection = (internal_calibration.dot(transformation)).dot(points_3D)
projections_2d[0, :] = camera_projection[0, :]/camera_projection[2, :]
projections_2d[1, :] = camera_projection[1, :]/camera_projection[2, :]
return projections_2d