-
Notifications
You must be signed in to change notification settings - Fork 124
/
camera_calibration.py
170 lines (142 loc) · 6.66 KB
/
camera_calibration.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
__author__ = 'Douglas and Iacopo'
import numpy as np
import cv2
import math
def estimate_camera(model3D, fidu_XY, pose_db_on=False):
if pose_db_on:
rmat, tvec = calib_camera(model3D, fidu_XY, pose_db_on=True)
tvec = tvec.reshape(3,1)
else:
rmat, tvec = calib_camera(model3D, fidu_XY)
RT = np.hstack((rmat, tvec))
projection_matrix = model3D.out_A * RT
return projection_matrix, model3D.out_A, rmat, tvec
def calib_camera(model3D, fidu_XY, pose_db_on=False):
#compute pose using refrence 3D points + query 2D point
## np.arange(68)+1 since matlab starts from 1
if pose_db_on:
rvecs = fidu_XY[0:3]
tvec = fidu_XY[3:6]
else:
goodind = np.setdiff1d(np.arange(68)+1, model3D.indbad)
goodind=goodind-1
fidu_XY = fidu_XY[goodind,:]
ret, rvecs, tvec = cv2.solvePnP(model3D.model_TD, fidu_XY, model3D.out_A, None, None, None, False)
rmat, jacobian = cv2.Rodrigues(rvecs, None)
inside = calc_inside(model3D.out_A, rmat, tvec, model3D.size_U[1], model3D.size_U[0], model3D.model_TD)
if(inside == 0):
tvec = -tvec
t = np.pi
RRz180 = np.asmatrix([np.cos(t), -np.sin(t), 0, np.sin(t), np.cos(t), 0, 0, 0, 1]).reshape((3, 3))
rmat = RRz180*rmat
return rmat, tvec
def get_yaw(rmat):
modelview = rmat
modelview = np.zeros( (3,4 ))
modelview[0:3,0:3] = rmat.transpose()
modelview = modelview.reshape(12)
# Code converted from function: getEulerFromRot()
angle_y = -math.asin( modelview[2] ) # Calculate Y-axis angle
C = math.cos( angle_y)
angle_y = math.degrees(angle_y)
if np.absolute(C) > 0.005: # Gimball lock?
trX = modelview[10] / C # No, so get X-axis angle
trY = -modelview[6] / C
angle_x = math.degrees( math.atan2( trY, trX ) )
trX = modelview[0] / C # Get z-axis angle
trY = - modelview[1] / C
angle_z = math.degrees( math.atan2( trY, trX) )
else:
# Gimball lock has occured
angle_x = 0
trX = modelview[5]
trY = modelview[4]
angle_z = math.degrees( math.atan2( trY, trX) )
# Adjust to current mesh setting
angle_x = 180 - angle_x
angle_y = angle_y
angle_z = -angle_z
out_pitch = angle_x
out_yaw = angle_y
out_roll = angle_z
return out_yaw
def get_opengl_matrices(camera_matrix, rmat, tvec, width, height):
projection_matrix = np.asmatrix(np.zeros((4,4)))
near_plane = 0.0001
far_plane = 10000
fx = camera_matrix[0,0]
fy = camera_matrix[1,1]
px = camera_matrix[0,2]
py = camera_matrix[1,2]
projection_matrix[0, 0] = 2.0 * fx / width
projection_matrix[1, 1] = 2.0 * fy / height
projection_matrix[0, 2] = 2.0 * (px / width) - 1.0
projection_matrix[1, 2] = 2.0 * (py / height) - 1.0
projection_matrix[2, 2] = -(far_plane + near_plane) / (far_plane - near_plane)
projection_matrix[3, 2] = -1
projection_matrix[2, 3] = -2.0 * far_plane * near_plane / (far_plane - near_plane)
deg = 180
t = deg*np.pi/180.
RRz=np.asmatrix([np.cos(t), -np.sin(t), 0, np.sin(t), np.cos(t), 0, 0, 0, 1]).reshape((3, 3))
RRy=np.asmatrix([np.cos(t), 0, np.sin(t), 0, 1, 0, -np.sin(t), 0, np.cos(t)]).reshape((3, 3))
rmat=RRz*RRy*rmat
mv = np.asmatrix(np.zeros((4, 4)))
mv[0:3, 0:3] = rmat
mv[0, 3] = tvec[0]
mv[1, 3] = -tvec[1]
mv[2, 3] = -tvec[2]
mv[3, 3] = 1.
return mv, projection_matrix
def extract_frustum(camera_matrix, rmat, tvec, width, height):
mv, proj = get_opengl_matrices(camera_matrix, rmat, tvec, width, height)
clip = proj * mv
frustum = np.asmatrix(np.zeros((6 ,4)))
#/* Extract the numbers for the RIGHT plane */
frustum[0, :] = clip[3, :] - clip[0, :]
#/* Normalize the result */
v = frustum[0, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[0, :] = frustum[0, :]/t
#/* Extract the numbers for the LEFT plane */
frustum[1, :] = clip[3, :] + clip[0, :]
#/* Normalize the result */
v = frustum[1, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[1, :] = frustum[1, :]/t
#/* Extract the BOTTOM plane */
frustum[2, :] = clip[3, :] + clip[1, :]
#/* Normalize the result */
v = frustum[2, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[2, :] = frustum[2, :]/t
#/* Extract the TOP plane */
frustum[3, :] = clip[3, :] - clip[1, :]
#/* Normalize the result */
v = frustum[3, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[3, :] = frustum[3, :]/t
#/* Extract the FAR plane */
frustum[4, :] = clip[3, :] - clip[2, :]
#/* Normalize the result */
v = frustum[4, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[4, :] = frustum[4, :]/t
#/* Extract the NEAR plane */
frustum[5, :] = clip[3, :] + clip[2, :]
#/* Normalize the result */
v = frustum[5, :3]
t = np.sqrt(np.sum(np.multiply(v, v)))
frustum[5, :] = frustum[5, :]/t
return frustum
def calc_inside(camera_matrix, rmat, tvec, width, height, obj_points):
frustum = extract_frustum(camera_matrix, rmat, tvec, width, height)
inside = 0
for point in obj_points:
if(point_in_frustum(point[0], point[1], point[2], frustum) > 0):
inside += 1
return inside
def point_in_frustum(x, y, z, frustum):
for p in range(0, 3):
if(frustum[p, 0] * x + frustum[p, 1] * y + frustum[p, 2] + z + frustum[p, 3] <= 0):
return False
return True