基于OpenCV 和 Dlib 进行头部姿态估计

庐山烟雨浙江潮,未到千般恨不消。到得还来别无事,庐山烟雨浙江潮。—-《庐山烟雨浙江潮》苏轼

写在前面


  • 工作中遇到,简单整理
  • 博文内容涉及基于 OpenCV 和 Dlib头部姿态评估的简单Demo
  • 理解不足小伙伴帮忙指正

庐山烟雨浙江潮,未到千般恨不消。到得还来别无事,庐山烟雨浙江潮。 —-《庐山烟雨浙江潮》苏轼


https://github.com/LIRUILONGS/Head-posture-detection-dlib-opencv-.git

实验项目以上传,只需 git 克隆,安装需要的 pytohn 包,就可以开始使用了,但是需要说明的是 Dlib 的基于 HOG特征和SVM分类器的人脸检测器很一般,很多脸都检测不到,实际情况中可以考虑使用深度学习模型来做关键点检测,然后评估姿态。可以查看文章末尾大佬的开源项目

实现效果

Demo
原图
原图
原图
特征点标记后
特征点标记后
姿态标记
姿态标记
姿态对应的Yaw,Pitch,Roll 度数
姿态对应的Yaw,Pitch,Roll
姿态对应的Yaw,Pitch,Roll

步骤

三个主要步骤

人脸检测

人脸检测:引入人脸检测器 dlib.get_frontal_face_detector() 以检测包含人脸的图片,多个人脸会选择面积最大的人脸。

dlib.get_frontal_face_detector()dlib 库中的一个函数,用于获取一个基于HOG特征和SVM分类器的人脸检测器。该函数返回一个可以用于检测图像中人脸的对象。

具体来说,HOG(Histogram of Oriented Gradients,梯度方向直方图)是一种常用于图像识别中的特征描述子,SVM(Support Vector Machine,支持向量机)是一种常用的分类器。将HOG特征与SVM分类器结合起来,可以得到一个有效的人脸检测器。

在使用 dlib.get_frontal_face_detector()函数时,只需将待检测的图像作为参数传入,即可得到一个用于检测人脸的对象。一个Demo

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
import dlib
import cv2

# 读取图像
img = cv2.imread('image.jpg')

# 获取人脸检测器
detector = dlib.get_frontal_face_detector()

# 在图像中检测人脸
faces = detector(img)

# 输出检测到的人脸数
print("检测到的人脸数为:", len(faces))

面部特征点检测

面部特征点检测,利用预训练模型 shape_predictor_68_face_landmarks.dat 以人脸图像为输入,输出68个人脸特征点

shape_predictor_68_face_landmarks.dat 是基于 dlib 库中的人脸特征点检测模型,该模型使用了基于 HOG 特征和 SVM 分类器的人脸检测器来检测图像中的人脸,并使用回归算法来预测人脸的 68 个关键点位置。这些关键点包括眼睛、鼻子、嘴巴等部位,可以用于进行人脸识别、表情识别、姿态估计等应用。

这个模型文件可以在dlib的官方网站上下载。在使用它之前,需要安装dlib库并将模型文件加载到程序中。

1
predictor = dlib.shape_predictor(r".\shape_predictor_68_face_landmarks.dat")

姿势估计

姿势估计。在获得 68 个面部特征点后,选择部分特征点,通过 PnP算法计算姿势 Yaw、Pitch、Roll 度数

1
2
(success, rotation_vector, translation_vector) = cv2.solvePnP(model_points, image_points, camera_matrix,
dist_coeffs, flags=cv2.SOLVEPNP_ITERATIVE)

Yaw、Pitch、Roll 是用于描述物体或相机在三维空间中的旋转角度的术语,常用于姿态估计和姿态控制中。

  • Yaw(左右):绕垂直于物体或相机的轴旋转的角度,也称为偏航角。通常以 z 轴为轴进行旋转,正值表示逆时针旋转,负值表示顺时针旋转。
  • Pitch(上下):绕物体或相机的横轴旋转的角度,也称为俯仰角。通常以 x 轴为轴进行旋转,正值表示向上旋转,负值表示向下旋转。
  • Roll(弯曲):绕物体或相机的纵轴旋转的角度,也称为翻滚角。通常以 y 轴为轴进行旋转,正值表示向右旋转,负值表示向左旋转。

这三个角度通常以欧拉角的形式表示,可以用于描述物体或相机的姿态信息。在计算机视觉中,常用于人脸识别、动作捕捉、机器人控制等应用场景。

完整 Demo 代码

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
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
#!/usr/bin/env python
# -*- encoding: utf-8 -*-
"""
@File : face_ypr_demo.py
@Time : 2023/06/05 21:32:45
@Author : Li Ruilong
@Version : 1.0
@Contact : liruilonger@gmail.com
@Desc : 根据68个人脸关键点,获取人头部姿态评估
"""

# here put the import lib

import cv2
import numpy as np
import dlib
import math
import uuid

# 头部姿态检测(dlib+opencv)

detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(r".\shape_predictor_68_face_landmarks.dat")
POINTS_NUM_LANDMARK = 68


# shape_predictor_68_face_landmarks.dat 是一个预训练的人脸关键点检测模型,可以用于识别人脸的68个关键点,如眼睛、鼻子、嘴巴等。这个模型可以被用于人脸识别、人脸表情分析、面部姿势估计等领域。
# 它是由dlib库提供的,可以在Python中使用。如果你想使用它,可以在dlib的官方网站上下载。

# 获取最大的人脸
def _largest_face(dets):
"""
@Time : 2023/06/05 21:30:37
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 从一个由 dlib 库检测到的人脸框列表中,找到最大的人脸框,并返回该框在列表中的索
如果只有一个人脸,直接返回
Args:
dets: 一个由 `dlib.rectangle` 类型的对象组成的列表,每个对象表示一个人脸框
Returns:
人脸索引
"""
# 如果列表长度为1,则直接返回
if len(dets) == 1:
return 0
# 计算每个人脸框的面积
face_areas = [(det.right() - det.left()) * (det.bottom() - det.top()) for det in dets]
import heapq
# 找到面积最大的人脸框的索引
largest_area = face_areas[0]
largest_index = 0
for index in range(1, len(dets)):
if face_areas[index] > largest_area:
largest_index = index
largest_area = face_areas[index]
# 打印最大人脸框的索引和总人脸数
print("largest_face index is {} in {} faces".format(largest_index, len(dets)))

return largest_index


def get_image_points_from_landmark_shape(landmark_shape):
"""
@Time : 2023/06/05 22:30:02
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 从dlib的检测结果抽取姿态估计需要的点坐标
Args:
landmark_shape: 所有的位置点
Returns:
void
"""

if landmark_shape.num_parts != POINTS_NUM_LANDMARK:
print("ERROR:landmark_shape.num_parts-{}".format(landmark_shape.num_parts))
return -1, None

# 2D image points. If you change the image, you need to change vector

image_points = np.array([
(landmark_shape.part(17).x, landmark_shape.part(17).y), # 17 left brow left corner
(landmark_shape.part(21).x, landmark_shape.part(21).y), # 21 left brow right corner
(landmark_shape.part(22).x, landmark_shape.part(22).y), # 22 right brow left corner
(landmark_shape.part(26).x, landmark_shape.part(26).y), # 26 right brow right corner
(landmark_shape.part(36).x, landmark_shape.part(36).y), # 36 left eye left corner
(landmark_shape.part(39).x, landmark_shape.part(39).y), # 39 left eye right corner
(landmark_shape.part(42).x, landmark_shape.part(42).y), # 42 right eye left corner
(landmark_shape.part(45).x, landmark_shape.part(45).y), # 45 right eye right corner
(landmark_shape.part(31).x, landmark_shape.part(31).y), # 31 nose left corner
(landmark_shape.part(35).x, landmark_shape.part(35).y), # 35 nose right corner
(landmark_shape.part(48).x, landmark_shape.part(48).y), # 48 mouth left corner
(landmark_shape.part(54).x, landmark_shape.part(54).y), # 54 mouth right corner
(landmark_shape.part(57).x, landmark_shape.part(57).y), # 57 mouth central bottom corner
(landmark_shape.part(8).x, landmark_shape.part(8).y), # 8 chin corner
], dtype="double")
return 0, image_points


def get_image_points(img):
"""
@Time : 2023/06/05 22:30:43
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 用dlib检测关键点,返回姿态估计需要的几个点坐标
Args:

Returns:
void
"""

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # 图片调整为灰色

dets = detector(img, 0)

if 0 == len(dets):
print("ERROR: found no face")
return -1, None
largest_index = _largest_face(dets)
face_rectangle = dets[largest_index]

landmark_shape = predictor(img, face_rectangle)
draw = im.copy()
cv2.circle(draw, (landmark_shape.part(0).x, landmark_shape.part(0).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(1).x, landmark_shape.part(1).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(2).x, landmark_shape.part(2).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(3).x, landmark_shape.part(3).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(4).x, landmark_shape.part(4).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(5).x, landmark_shape.part(5).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(6).x, landmark_shape.part(6).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(7).x, landmark_shape.part(7).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(8).x, landmark_shape.part(8).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(9).x, landmark_shape.part(9).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(10).x, landmark_shape.part(10).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(11).x, landmark_shape.part(11).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(12).x, landmark_shape.part(12).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(13).x, landmark_shape.part(13).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(14).x, landmark_shape.part(14).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(15).x, landmark_shape.part(15).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(16).x, landmark_shape.part(16).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(17).x, landmark_shape.part(17).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(18).x, landmark_shape.part(18).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(19).x, landmark_shape.part(19).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(20).x, landmark_shape.part(20).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(21).x, landmark_shape.part(21).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(22).x, landmark_shape.part(22).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(23).x, landmark_shape.part(23).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(24).x, landmark_shape.part(24).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(25).x, landmark_shape.part(25).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(26).x, landmark_shape.part(26).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(27).x, landmark_shape.part(27).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(28).x, landmark_shape.part(28).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(29).x, landmark_shape.part(29).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(30).x, landmark_shape.part(30).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(31).x, landmark_shape.part(31).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(32).x, landmark_shape.part(32).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(33).x, landmark_shape.part(33).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(34).x, landmark_shape.part(34).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(35).x, landmark_shape.part(35).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(36).x, landmark_shape.part(36).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(37).x, landmark_shape.part(37).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(38).x, landmark_shape.part(38).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(39).x, landmark_shape.part(39).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(40).x, landmark_shape.part(40).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(41).x, landmark_shape.part(41).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(42).x, landmark_shape.part(42).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(43).x, landmark_shape.part(43).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(44).x, landmark_shape.part(44).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(45).x, landmark_shape.part(45).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(46).x, landmark_shape.part(46).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(47).x, landmark_shape.part(47).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(48).x, landmark_shape.part(48).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(49).x, landmark_shape.part(49).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(50).x, landmark_shape.part(50).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(51).x, landmark_shape.part(51).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(52).x, landmark_shape.part(52).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(53).x, landmark_shape.part(53).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(54).x, landmark_shape.part(54).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(55).x, landmark_shape.part(55).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(56).x, landmark_shape.part(56).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(57).x, landmark_shape.part(57).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(58).x, landmark_shape.part(58).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(59).x, landmark_shape.part(59).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(60).x, landmark_shape.part(60).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(61).x, landmark_shape.part(61).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(62).x, landmark_shape.part(62).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(63).x, landmark_shape.part(63).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(64).x, landmark_shape.part(64).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(65).x, landmark_shape.part(65).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(66).x, landmark_shape.part(66).y), 2, (0, 255, 0), -1)
cv2.circle(draw, (landmark_shape.part(67).x, landmark_shape.part(67).y), 2, (0, 255, 0), -1)

# 部分关键点特殊标记
cv2.circle(draw, (landmark_shape.part(17).x, landmark_shape.part(17).y), 2, (0, 165, 255),
-1) # 17 left brow left corner
cv2.circle(draw, (landmark_shape.part(21).x, landmark_shape.part(21).y), 2, (0, 165, 255),
-1) # 21 left brow right corner
cv2.circle(draw, (landmark_shape.part(22).x, landmark_shape.part(22).y), 2, (0, 165, 255),
-1) # 22 right brow left corner
cv2.circle(draw, (landmark_shape.part(26).x, landmark_shape.part(26).y), 2, (0, 165, 255),
-1) # 26 right brow right corner
cv2.circle(draw, (landmark_shape.part(36).x, landmark_shape.part(36).y), 2, (0, 165, 255),
-1) # 36 left eye left corner
cv2.circle(draw, (landmark_shape.part(39).x, landmark_shape.part(39).y), 2, (0, 165, 255),
-1) # 39 left eye right corner
cv2.circle(draw, (landmark_shape.part(42).x, landmark_shape.part(42).y), 2, (0, 165, 255),
-1) # 42 right eye left corner
cv2.circle(draw, (landmark_shape.part(45).x, landmark_shape.part(45).y), 2, (0, 165, 255),
-1) # 45 right eye right corner
cv2.circle(draw, (landmark_shape.part(31).x, landmark_shape.part(31).y), 2, (0, 165, 255),
-1) # 31 nose left corner
cv2.circle(draw, (landmark_shape.part(35).x, landmark_shape.part(35).y), 2, (0, 165, 255),
-1) # 35 nose right corner
cv2.circle(draw, (landmark_shape.part(48).x, landmark_shape.part(48).y), 2, (0, 165, 255),
-1) # 48 mouth left corner
cv2.circle(draw, (landmark_shape.part(54).x, landmark_shape.part(54).y), 2, (0, 165, 255),
-1) # 54 mouth right corner
cv2.circle(draw, (landmark_shape.part(57).x, landmark_shape.part(57).y), 2, (0, 165, 255),
-1) # 57 mouth central bottom corner
cv2.circle(draw, (landmark_shape.part(8).x, landmark_shape.part(8).y), 2, (0, 165, 255), -1)

# 保存关键点标记后的图片
cv2.imwrite('new_' + "KeyPointDetection.jpg", draw)

return get_image_points_from_landmark_shape(landmark_shape)


def get_pose_estimation(img_size, image_points):
"""
@Time : 2023/06/05 22:31:31
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 获取旋转向量和平移向量
Args:

Returns:
void
"""

# 3D model points.
model_points = np.array([
(6.825897, 6.760612, 4.402142), # 33 left brow left corner
(1.330353, 7.122144, 6.903745), # 29 left brow right corner
(-1.330353, 7.122144, 6.903745), # 34 right brow left corner
(-6.825897, 6.760612, 4.402142), # 38 right brow right corner
(5.311432, 5.485328, 3.987654), # 13 left eye left corner
(1.789930, 5.393625, 4.413414), # 17 left eye right corner
(-1.789930, 5.393625, 4.413414), # 25 right eye left corner
(-5.311432, 5.485328, 3.987654), # 21 right eye right corner
(2.005628, 1.409845, 6.165652), # 55 nose left corner
(-2.005628, 1.409845, 6.165652), # 49 nose right corner
(2.774015, -2.080775, 5.048531), # 43 mouth left corner
(-2.774015, -2.080775, 5.048531), # 39 mouth right corner
(0.000000, -3.116408, 6.097667), # 45 mouth central bottom corner
(0.000000, -7.415691, 4.070434) # 6 chin corner
])
# Camera internals

focal_length = img_size[1]
center = (img_size[1] / 2, img_size[0] / 2)
camera_matrix = np.array(
[[focal_length, 0, center[0]],
[0, focal_length, center[1]],
[0, 0, 1]], dtype="double"
)

dist_coeffs = np.array([7.0834633684407095e-002, 6.9140193737175351e-002, 0.0, 0.0, -1.3073460323689292e+000],
dtype="double") # Assuming no lens distortion

(success, rotation_vector, translation_vector) = cv2.solvePnP(model_points, image_points, camera_matrix,
dist_coeffs, flags=cv2.SOLVEPNP_ITERATIVE)

# print("Rotation Vector:\n {}".format(rotation_vector))
# print("Translation Vector:\n {}".format(translation_vector))
return success, rotation_vector, translation_vector, camera_matrix, dist_coeffs


def draw_annotation_box(image, rotation_vector, translation_vector, camera_matrix, dist_coeefs, color=(0, 255, 0),
line_width=2):
"""
@Time : 2023/06/05 22:09:14
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 标记一个人脸朝向的3D框
Args:

Returns:
void
"""

"""Draw a 3D box as annotation of pose"""
point_3d = []
rear_size = 10
rear_depth = 0
point_3d.append((-rear_size, -rear_size, rear_depth))
point_3d.append((-rear_size, rear_size, rear_depth))
point_3d.append((rear_size, rear_size, rear_depth))
point_3d.append((rear_size, -rear_size, rear_depth))
point_3d.append((-rear_size, -rear_size, rear_depth))

front_size = 10
# 高度
front_depth = 10
point_3d.append((-front_size, -front_size, front_depth))
point_3d.append((-front_size, front_size, front_depth))
point_3d.append((front_size, front_size, front_depth))
point_3d.append((front_size, -front_size, front_depth))
point_3d.append((-front_size, -front_size, front_depth))
point_3d = np.array(point_3d, dtype=np.float32).reshape(-1, 3)

# Map to 2d image points
(point_2d, _) = cv2.projectPoints(point_3d,
rotation_vector,
translation_vector,
camera_matrix,
dist_coeefs)
point_2d = np.int32(point_2d.reshape(-1, 2))

# Draw all the lines
cv2.polylines(image, [point_2d], True, color, line_width, cv2.LINE_AA)
cv2.line(image, tuple(point_2d[1]), tuple(
point_2d[6]), color, line_width, cv2.LINE_AA)
cv2.line(image, tuple(point_2d[2]), tuple(
point_2d[7]), color, line_width, cv2.LINE_AA)
cv2.line(image, tuple(point_2d[3]), tuple(
point_2d[8]), color, line_width, cv2.LINE_AA)


# 从旋转向量转换为欧拉角
def get_euler_angle(rotation_vector):
"""
@Time : 2023/06/05 22:31:52
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 从旋转向量转换为欧拉角
Args:

Returns:
void
"""

# calculate rotation angles
theta = cv2.norm(rotation_vector, cv2.NORM_L2)

# transformed to quaterniond
w = math.cos(theta / 2)
x = math.sin(theta / 2) * rotation_vector[0][0] / theta
y = math.sin(theta / 2) * rotation_vector[1][0] / theta
z = math.sin(theta / 2) * rotation_vector[2][0] / theta

ysqr = y * y
# pitch (x-axis rotation)
t0 = 2.0 * (w * x + y * z)
t1 = 1.0 - 2.0 * (x * x + ysqr)

# print('t0:{}, t1:{}'.format(t0, t1))
pitch = math.atan2(t0, t1)

# yaw (y-axis rotation)
t2 = 2.0 * (w * y - z * x)
if t2 > 1.0:
t2 = 1.0
if t2 < -1.0:
t2 = -1.0
yaw = math.asin(t2)

# roll (z-axis rotation)
t3 = 2.0 * (w * z + x * y)
t4 = 1.0 - 2.0 * (ysqr + z * z)
roll = math.atan2(t3, t4)

print('pitch:{}, yaw:{}, roll:{}'.format(pitch, yaw, roll))

# 单位转换:将弧度转换为度
pitch_degree = int((pitch / math.pi) * 180)
yaw_degree = int((yaw / math.pi) * 180)
roll_degree = int((roll / math.pi) * 180)

return 0, pitch, yaw, roll, pitch_degree, yaw_degree, roll_degree


def get_pose_estimation_in_euler_angle(landmark_shape, im_szie):
try:
ret, image_points = get_image_points_from_landmark_shape(landmark_shape)
if ret != 0:
print('get_image_points failed')
return -1, None, None, None

ret, rotation_vector, translation_vector, camera_matrix, dist_coeffs = get_pose_estimation(im_szie,
image_points)
if ret != True:
print('get_pose_estimation failed')
return -1, None, None, None

ret, pitch, yaw, roll = get_euler_angle(rotation_vector)
if ret != 0:
print('get_euler_angle failed')
return -1, None, None, None

euler_angle_str = 'Pitch:{}, Yaw:{}, Roll:{}'.format(pitch, yaw, roll)
print(euler_angle_str)
return 0, pitch, yaw, roll

except Exception as e:
print('get_pose_estimation_in_euler_angle exception:{}'.format(e))
return -1, None, None, None


def build_img_text_marge(img_, text, height):
"""
@Time : 2023/06/01 05:29:09
@Author : liruilonger@gmail.com
@Version : 1.0
@Desc : 生成文字图片拼接到 img 对象
Args:

Returns:
void
"""
import cv2
from PIL import Image, ImageDraw, ImageFont

# 定义图片大小和背景颜色
width = img_.shape[1]
background_color = (255, 255, 255)

# 定义字体、字号和颜色
font_path = 'arial.ttf'
font_size = 26
font_color = (0, 0, 0)

# 创建空白图片
image = Image.new('RGB', (width, height), background_color)

# 创建画笔
draw = ImageDraw.Draw(image)

# 加载字体
font = ImageFont.truetype(font_path, font_size)

# 写入文字
text_width, text_height = draw.textsize(text, font)
text_x = (width - text_width) // 2
text_y = (height - text_height) // 2
draw.text((text_x, text_y), text, font=font, fill=font_color)

# 将Pillow图片转换为OpenCV图片
image_cv = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)

montage_size = (width, img_.shape[0])
import imutils
montages = imutils.build_montages([img_, image_cv], montage_size, (1, 2))

# 保存图片
return montages[0]


if __name__ == '__main__':
from imutils import paths

# for imagePath in paths.list_images("W:\\python_code\\deepface\\huge_1.jpg"):
for imagePath in range(1):
print(f"处理的图片路径为: {imagePath}")
# Read Image
im = cv2.imread("image.jpg")
size = im.shape
# 对图像进行缩放的操作
if size[0] > 700:
h = size[0] / 3
w = size[1] / 3
# 如果图像的高度大于700,就将其高度和宽度分别缩小为原来的1/3,然后使用双三次插值的方法进行缩放。最后返回缩放后的图像的大小。
im = cv2.resize(im, (int(w), int(h)), interpolation=cv2.INTER_CUBIC)
size = im.shape
# 获取坐标点
ret, image_points = get_image_points(im)
if ret != 0:
print('get_image_points failed')
continue

ret, rotation_vector, translation_vector, camera_matrix, dist_coeffs = get_pose_estimation(size, image_points)

if ret != True:
print('get_pose_estimation failed')
continue
draw_annotation_box(im, rotation_vector, translation_vector, camera_matrix, dist_coeffs)
cv2.imwrite('new_' + "draw_annotation_box.jpg", im)

ret, pitch, yaw, roll, pitch_degree, yaw_degree, roll_degree = get_euler_angle(rotation_vector)

draw = im.copy()
# Yaw:

if yaw_degree < 0:
output_yaw = "left : " + str(abs(yaw_degree)) + " degrees"
elif yaw_degree > 0:
output_yaw = "right :" + str(abs(yaw_degree)) + " degrees"
else:
output_yaw = "No left or right"
print(output_yaw)

# Pitch:
if pitch_degree > 0:
output_pitch = "dow :" + str(abs(pitch_degree)) + " degrees"
elif pitch_degree < 0:
output_pitch = "up :" + str(abs(pitch_degree)) + " degrees"
else:
output_pitch = "No downwards or upwards"
print(output_pitch)

# Roll:
if roll_degree < 0:
output_roll = "bends to the right: " + str(abs(roll_degree)) + " degrees"
elif roll_degree > 0:
output_roll = "bends to the left: " + str(abs(roll_degree)) + " degrees"
else:
output_roll = "No bend right or left."
print(output_roll)

# Initial status:
if abs(yaw) < 0.00001 and abs(pitch) < 0.00001 and abs(roll) < 0.00001:
cv2.putText(draw, "Initial ststus", (20, 40), cv2.FONT_HERSHEY_SIMPLEX, .5, (0, 255, 0))
print("Initial ststus")

# 姿态检测完的数据写在对应的照片
imgss = build_img_text_marge(im, output_yaw + "\n" + output_pitch + "\n" + output_roll, 200)
cv2.imwrite('new_' + str(uuid.uuid4()).replace('-', '') + ".jpg", imgss)

博文部分内容参考

© 文中涉及参考链接内容版权归原作者所有,如有侵权请告知,这是一个开源项目,如果你认可它,不要吝啬星星哦 :)


https://blog.csdn.net/zhang2gongzi/article/details/124520896

https://github.com/JuneoXIE/

https://github.com/yinguobing/head-pose-estimation


© 2018-2023 liruilonger@gmail.com, All rights reserved. 保持署名-非商用-相同方式共享(CC BY-NC-SA 4.0)

发布于

2023-06-02

更新于

2024-01-22

许可协议

评论
Your browser is out-of-date!

Update your browser to view this website correctly.&npsb;Update my browser now

×