forked from hou-yz/MultiviewX
-
Notifications
You must be signed in to change notification settings - Fork 0
/
vali.py
218 lines (186 loc) · 10.5 KB
/
vali.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
import os
import warnings
import cv2
import re
import json
import matplotlib.pyplot as plt
import numpy as np
import datasetParameters
intrinsic_camera_matrix_filenames = []
extrinsic_camera_matrix_filenames = []
def get_error(worldcoord2imgcoord_mat,validate_Points, validate_Points_3d,cam, modified = True):
idx = 0
error = 0
for validate_Point_3d in validate_Points_3d:
if(modified):
wcoord = np.array([validate_Point_3d[0], validate_Point_3d[1], validate_Point_3d[2], 1])
else:
wcoord = np.array([validate_Point_3d[0], validate_Point_3d[1], 1])
#wcoord = np.array([validate_Point_3d[0], validate_Point_3d[1], 0])
trueimagecoord = np.array([validate_Points[idx][0], validate_Points[idx][1]])
imgcoord = worldcoord2imgcoord_mat @ wcoord
if(idx % 10 == 0):
print(f"Camera{cam + 1} trueimagecoord: {trueimagecoord[0]},{trueimagecoord[1]}")
print(f"Camera{cam + 1} imgcoordx: {imgcoord[0] / imgcoord[2]},{imgcoord[1] / imgcoord[2]}")
print(f"Camera{cam + 1} raw imgcoord: {imgcoord[0]},{imgcoord[1]},{imgcoord[2]}")
print()
error = error + trueimagecoord[0] - imgcoord[0]/ imgcoord[2] + trueimagecoord[1] - imgcoord[1] / imgcoord[2]
idx = idx + 1
if( abs(error/idx) > 0.1):
print("""
___ .___________.___________. _______ .__ __. .___________. __ ______ .__ __.
/ \ | | || ____|| \ | | | || | / __ \ | \ | |
/ ^ \ `---| |----`---| |----`| |__ | \| | `---| |----`| | | | | | | \| |
/ /_\ \ | | | | | __| | . ` | | | | | | | | | | . ` |
/ _____ \ | | | | | |____ | |\ | | | | | | `--' | | |\ |
/__/ \__\ |__| |__| |_______||__| \__| |__| |__| \______/ |__| \__|
""")
print(
f"Please try to adjust the 'tRandomOffset' field and wait for the end of updates of chessborad in Unity")
print(f"The intrinsic of this Camera{cam + 1} does NOT pass validation")
warnings.warn(f"Failed to validate Calibration",DeprecationWarning)
exit()
return error/idx
def get_imgcoord2worldgrid_matrices(intrinsic_matrices, extrinsic_matrices, worldgrid2worldcoord_mat,modified = True):
projection_matrices = {}
errors = []
# This block can only be used when you have got frames
# for cam in range(datasetParameters.NUM_CAM):
#
# points_2d = np.loadtxt(os.path.join('matchings',f'Camera{cam + 1}.txt'))
# points_3d = np.loadtxt(os.path.join('matchings',f'Camera{cam + 1}_3d.txt'))
# points_2d = points_2d[points_2d[:, 0] == 0, :]
# points_3d = points_3d[points_3d[:, 0] == 0, :]
# visualize_foot_image = points_2d[points_2d[:, 0] == 0, -2:]
# image = cv2.imread(os.path.join(datasetParameters.DATASET_NAME, f'Image_subsets/C{cam + 1}/0000.png'))
#
# for point in visualize_foot_image:
# cv2.circle(image, tuple(point.astype(int)), 10, (0, 0, 255), -1)
# plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
# plt.show()
for cam in range(datasetParameters.NUM_CAM):
print(f"Try to validate Camera {cam+1}")
# print("intrinsic: ",intrinsic_matrices[cam])
# print("extrinsic: ",extrinsic_matrices[cam])
# print("extrinsic: ", extrinsic_matrices[cam][:,3])
# print("extrinsic_dele: ", np.delete(extrinsic_matrices[cam], 2, 1))
# print("extrinsic: \n", extrinsic_matrices[cam])
# print("extrinsic: \n", extrinsic_matrices[cam])
# print("extrinsic_dele \n: ", np.delete(extrinsic_matrices[cam], 2, 1))\
'''np.delete(extrinsic_matrices[cam], 2, 1) deletes the third column (index 2) of the extrinsic matrix,
which corresponds to the translation component.
This is because we are only interested in the rotation component of the extrinsic matrix,
which determines the orientation of the camera in world coordinates.
'''
# print(extrinsic_matrices[cam])
file_validatePoints = f'calib/C{cam + 1}/validatePoints.txt'
validate_Points = np.array(np.loadtxt(file_validatePoints).astype('float32'))
file_validatePoints_3d = f'calib/C{cam + 1}/validatePoints_3d.txt'
validate_Points_3d = np.array(np.loadtxt(file_validatePoints_3d).astype('float32'))
if(modified):
# 3 X 4 = 3 x 3 X 3 x 4
worldcoord2imgcoord_mat = intrinsic_matrices[cam] @ extrinsic_matrices[cam]
else:
extrinsic_matrices[cam][:, 3] = extrinsic_matrices[cam][:, 3] + extrinsic_matrices[cam][:, 2] * 0
# print(np.delete(extrinsic_matrices[cam], 2, 1))
worldcoord2imgcoord_mat = intrinsic_matrices[cam] @ np.delete(extrinsic_matrices[cam], 2, 1)
# Tevc
Tevc = extrinsic_matrices[cam][:, 3]
errors.append(get_error(worldcoord2imgcoord_mat,validate_Points,validate_Points_3d, cam, modified))
# image = cv2.imread(f'Image_subsets/C{cam + 1}/0000.png')
#
#
# cv2.circle(image, tuple(imgcoord[0:2].astype(int)), 20, (255, 0, 0), -1)
# plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
#
# plt.show()
# print("worldcoord2imgcoord_mat: ", worldcoord2imgcoord_mat)
if(not modified):
worldgrid2imgcoord_mat = worldcoord2imgcoord_mat @ worldgrid2worldcoord_mat
# print("worldgrid2imgcoord_mat: ", worldgrid2imgcoord_mat)
imgcoord2worldgrid_mat = np.linalg.inv(worldgrid2imgcoord_mat)
# image of shape C,H,W (C,N_row,N_col); indexed as x,y,w,h (x,y,n_col,n_row)
# matrix of shape N_row, N_col; indexed as x,y,n_row,n_col
# print("imgcoord2worldgrid_mat: ", imgcoord2worldgrid_mat)
permutation_mat = np.array([[0, 1, 0], [1, 0, 0], [0, 0, 1]])
projection_matrices[cam] = permutation_mat @ imgcoord2worldgrid_mat
print()
pass
return projection_matrices, errors
# def get_imgcoord2worldgrid_matrices(intrinsic_matrices, extrinsic_matrices, worldgrid2worldcoord_mat, depth_margin):
# projection_matrices = {}
# for cam in range(4):
#
# # intrinsic_matrices [4,4] @ extrinsic_matrices [4,4] @ worldgrid2worldcoord_mat [4,4]
# # inv: worldgrid2worldcoord_mat-1 @ extrinsic_matrices-1 @ intrinsic_matrices-1
# extrinsic_matrices[cam][:,3] = extrinsic_matrices[cam][:,3]+extrinsic_matrices[cam][:,2]*i*depth_margin
# worldcoord2imgcoord_mat = intrinsic_matrices[cam] @ np.delete(extrinsic_matrices[cam], 2, 1)
# worldgrid2imgcoord_mat = worldcoord2imgcoord_mat @ worldgrid2worldcoord_mat
# imgcoord2worldgrid_mat = np.linalg.inv(worldgrid2imgcoord_mat)
# # image of shape C,H,W (C,N_row,N_col); indexed as x,y,w,h (x,y,n_col,n_row)
# # matrix of shape N_row, N_col; indexed as x,y,n_row,n_col
# permutation_mat = np.array([[0, 1, 0], [1, 0, 0], [0, 0, 1]]) # x->y, y->x
# projection_matrices[(cam,i)] = permutation_mat @ imgcoord2worldgrid_mat
# return projection_matrices
def get_intrinsic_extrinsic_matrix(camIdx):
DATASET_NAME = os.path.join(datasetParameters.DATASET_NAME, "calibrations")
intrinsic_path = os.path.join(DATASET_NAME,f'intrinsic')
extrinsic_path = os.path.join(DATASET_NAME, f'extrinsic')
fp_calibration = cv2.FileStorage(os.path.join(intrinsic_path,
intrinsic_camera_matrix_filenames[camIdx]),
flags=cv2.FILE_STORAGE_READ)
intrinsic_matrix = fp_calibration.getNode('camera_matrix').mat()
fp_calibration.release()
fp_calibration = cv2.FileStorage(os.path.join(extrinsic_path,
extrinsic_camera_matrix_filenames[camIdx]),
flags=cv2.FILE_STORAGE_READ)
rvec, tvec = fp_calibration.getNode('rvec').mat().squeeze(), fp_calibration.getNode('tvec').mat().squeeze()
fp_calibration.release()
rotation_matrix, _ = cv2.Rodrigues(rvec)
translation_matrix = np.array(tvec, dtype=np.float32).reshape(3, 1)
extrinsic_matrix = np.hstack((rotation_matrix, translation_matrix))
# print()
# print(f'Camera{camIdx + 1}\'s intrinsic_matrix is ')
# print(f'{intrinsic_matrix} ')
# print(f'and extrinsic_matrix is ')
# print(f'{extrinsic_matrix}')
return intrinsic_matrix, extrinsic_matrix
def vali(modified = True):
for i in range(datasetParameters.NUM_CAM):
intrinsic_xml = f'intr_Camera{i + 1}.xml'
extrinsic_xml = f'extr_Camera{i + 1}.xml'
intrinsic_camera_matrix_filenames.append(intrinsic_xml)
extrinsic_camera_matrix_filenames.append(extrinsic_xml)
# 2x4
#worldgrid2worldcoord_mat = np.array([[0, 1/datasetParameters.MAP_EXPAND, 0, 1], [1/datasetParameters.MAP_EXPAND, 0, 0, 1]])
# 1x4
#worldgrid2worldcoord_mat = np.array([[1/datasetParameters.MAP_EXPAND,1/datasetParameters.MAP_EXPAND,0,1]])
# 4x2, 此时认为gird为 2x1 的坐标。 WIDTH = 25, HEIGHT = 16 EXPAND = 40 的情况下,例如有格点坐标
# (2,
# 1)
# ,经过此矩阵转换得到了,之所以如此做(4x1举证),是因为需要配合后续乘以worldcoord2imgcoord_mat(一个由3x3 乘以 3x4 矩阵相乘得到的 3x4 矩阵)(3 x 4 乘以 4 x 1 得到3x1)
# (0.05,
# 0.025,
# 0,
# 0,
# )
if(modified):
worldgrid2worldcoord_mat = np.array([[1/(datasetParameters.MAP_EXPAND*datasetParameters.MAP_WIDTH),0],[0,1/(datasetParameters.MAP_EXPAND*datasetParameters.MAP_HEIGHT)],[0,0],[0,0]])
else:
# 3x3,猜想此时grid为2x1 齐次坐标,
worldgrid2worldcoord_mat = np.array([[0, 0.025, 0], [0.025, 0, 0], [0, 0, 1]])
print(worldgrid2worldcoord_mat)
intrinsic_matrices, extrinsic_matrices = zip(*[get_intrinsic_extrinsic_matrix(cam) for cam in range(datasetParameters.NUM_CAM)])
# for intrinsic_matrix in intrinsic_matrices:
# print(intrinsic_matrix)
# print()
#
# for extrinsic_matrix in extrinsic_matrices:
# print(extrinsic_matrix)
# print()
imgcoord2worldgrid_matrices, errors = get_imgcoord2worldgrid_matrices(intrinsic_matrices, extrinsic_matrices, worldgrid2worldcoord_mat,modified)
print("Errors are :")
print(errors)
print()
if __name__ == '__main__':
vali(False)