python 3利用Dlib 19.7实现摄像头人脸检测特征点标定

时间:2021-08-30 20:03:20

python 3 利用 dlib 19.7 实现摄像头人脸检测特征点标定

0.引言

利用python开发,借助dlib库捕获摄像头中的人脸,进行实时特征点标定;

python 3利用Dlib 19.7实现摄像头人脸检测特征点标定

图1 工程效果示例(gif)

python 3利用Dlib 19.7实现摄像头人脸检测特征点标定

图2 工程效果示例(静态图片)

(实现比较简单,代码量也比较少,适合入门或者兴趣学习。)

1.开发环境

  python:  3.6.3

  dlib:    19.7

  opencv, numpy

?
1
2
3
import dlib     # 人脸识别的库dlib
import numpy as np # 数据处理的库numpy
import cv2     # 图像处理的库opencv

2.源码介绍

  其实实现很简单,主要分为两个部分:摄像头调用+人脸特征点标定

2.1 摄像头调用

  介绍下opencv中摄像头的调用方法;

  利用 cap = cv2.videocapture(0) 创建一个对象;

  (具体可以参考官方文档

?
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
# 2018-2-26
# by timestamp
# cnblogs: http://www.cnblogs.com/adaminxie
 
"""
cv2.videocapture(), 创建cv2摄像头对象/ open the default camera
 
  python: cv2.videocapture() → <videocapture object>
 
  python: cv2.videocapture(filename) → <videocapture object> 
  filename – name of the opened video file (eg. video.avi) or image sequence (eg. img_%02d.jpg, which will read samples like img_00.jpg, img_01.jpg, img_02.jpg, ...)
 
  python: cv2.videocapture(device) → <videocapture object>
  device – id of the opened video capturing device (i.e. a camera index). if there is a single camera connected, just pass 0.
 
"""
cap = cv2.videocapture(0)
 
 
"""
cv2.videocapture.set(propid, value),设置视频参数;
 
  propid:
  cv_cap_prop_pos_msec current position of the video file in milliseconds.
  cv_cap_prop_pos_frames 0-based index of the frame to be decoded/captured next.
  cv_cap_prop_pos_avi_ratio relative position of the video file: 0 - start of the film, 1 - end of the film.
  cv_cap_prop_frame_width width of the frames in the video stream.
  cv_cap_prop_frame_height height of the frames in the video stream.
  cv_cap_prop_fps frame rate.
  cv_cap_prop_fourcc 4-character code of codec.
  cv_cap_prop_frame_count number of frames in the video file.
  cv_cap_prop_format format of the mat objects returned by retrieve() .
  cv_cap_prop_mode backend-specific value indicating the current capture mode.
  cv_cap_prop_brightness brightness of the image (only for cameras).
  cv_cap_prop_contrast contrast of the image (only for cameras).
  cv_cap_prop_saturation saturation of the image (only for cameras).
  cv_cap_prop_hue hue of the image (only for cameras).
  cv_cap_prop_gain gain of the image (only for cameras).
  cv_cap_prop_exposure exposure (only for cameras).
  cv_cap_prop_convert_rgb boolean flags indicating whether images should be converted to rgb.
  cv_cap_prop_white_balance_u the u value of the whitebalance setting (note: only supported by dc1394 v 2.x backend currently)
  cv_cap_prop_white_balance_v the v value of the whitebalance setting (note: only supported by dc1394 v 2.x backend currently)
  cv_cap_prop_rectification rectification flag for stereo cameras (note: only supported by dc1394 v 2.x backend currently)
  cv_cap_prop_iso_speed the iso speed of the camera (note: only supported by dc1394 v 2.x backend currently)
  cv_cap_prop_buffersize amount of frames stored in internal buffer memory (note: only supported by dc1394 v 2.x backend currently)
  
  value: 设置的参数值/ value of the property
"""
cap.set(3, 480)
 
"""
cv2.videocapture.isopened(), 检查摄像头初始化是否成功 / check if we succeeded
返回true或false
"""
cap.isopened()
 
"""
cv2.videocapture.read([imgage]) -> retval,image, 读取视频 / grabs, decodes and returns the next video frame
返回两个值:
  一个是布尔值true/false,用来判断读取视频是否成功/是否到视频末尾
  图像对象,图像的三维矩阵
"""
flag, im_rd = cap.read()

2.2 人脸特征点标定

  调用预测器“shape_predictor_68_face_landmarks.dat”进行68点标定,这是dlib训练好的模型,可以直接调用进行人脸68个人脸特征点的标定;

  具体可以参考我的另一篇博客(python3利用dlib19.7实现人脸68个特征点标定); 

2.3 源码

  实现的方法比较简单:

  利用 cv2.videocapture() 创建摄像头对象,然后利用 flag, im_rd = cv2.videocapture.read() 读取摄像头视频,im_rd就是视频中的一帧帧图像;

  然后就类似于单张图像进行人脸检测,对这一帧帧的图像im_rd利用dlib进行特征点标定,然后绘制特征点;

  你可以按下s键来获取当前截图,或者按下q键来退出摄像头;

# 2018-2-26

?
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
# by timestamp
# cnblogs: http://www.cnblogs.com/adaminxie
# github: https://github.com/coneypo/dlib_face_detection_from_camera
 
import dlib           #人脸识别的库dlib
import numpy as np       #数据处理的库numpy
import cv2           #图像处理的库opencv
 
# dlib预测器
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
 
# 创建cv2摄像头对象
cap = cv2.videocapture(0)
 
# cap.set(propid, value)
# 设置视频参数,propid设置的视频参数,value设置的参数值
cap.set(3, 480)
 
# 截图screenshoot的计数器
cnt = 0
 
# cap.isopened() 返回true/false 检查初始化是否成功
while(cap.isopened()):
 
  # cap.read()
  # 返回两个值:
  #  一个布尔值true/false,用来判断读取视频是否成功/是否到视频末尾
  #  图像对象,图像的三维矩阵
  flag, im_rd = cap.read()
 
  # 每帧数据延时1ms,延时为0读取的是静态帧
  k = cv2.waitkey(1)
 
  # 取灰度
  img_gray = cv2.cvtcolor(im_rd, cv2.color_rgb2gray)
 
  # 人脸数rects
  rects = detector(img_gray, 0)
 
  #print(len(rects))
 
  # 待会要写的字体
  font = cv2.font_hershey_simplex
 
  # 标68个点
  if(len(rects)!=0):
    # 检测到人脸
    for i in range(len(rects)):
      landmarks = np.matrix([[p.x, p.y] for p in predictor(im_rd, rects[i]).parts()])
 
      for idx, point in enumerate(landmarks):
        # 68点的坐标
        pos = (point[0, 0], point[0, 1])
 
        # 利用cv2.circle给每个特征点画一个圈,共68个
        cv2.circle(im_rd, pos, 2, color=(0, 255, 0))
 
        # 利用cv2.puttext输出1-68
        cv2.puttext(im_rd, str(idx + 1), pos, font, 0.2, (0, 0, 255), 1, cv2.line_aa)
    cv2.puttext(im_rd, "faces: "+str(len(rects)), (20,50), font, 1, (0, 0, 255), 1, cv2.line_aa)
  else:
    # 没有检测到人脸
    cv2.puttext(im_rd, "no face", (20, 50), font, 1, (0, 0, 255), 1, cv2.line_aa)
 
  # 添加说明
  im_rd = cv2.puttext(im_rd, "s: screenshot", (20, 400), font, 0.8, (255, 255, 255), 1, cv2.line_aa)
  im_rd = cv2.puttext(im_rd, "q: quit", (20, 450), font, 0.8, (255, 255, 255), 1, cv2.line_aa)
 
  # 按下s键保存
  if (k == ord('s')):
    cnt+=1
    cv2.imwrite("screenshoot"+str(cnt)+".jpg", im_rd)
 
  # 按下q键退出
  if(k==ord('q')):
    break
 
  # 窗口显示
  cv2.imshow("camera", im_rd)
 
# 释放摄像头
cap.release()
 
# 删除建立的窗口
cv2.destroyallwindows()

如果对您有帮助,欢迎在github上star本项目

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:https://www.cnblogs.com/AdaminXie/p/8472743.html