python检测吸烟的算法_Openpose驾驶员危险驾驶检测（抽烟打电话）

序

目的：最近在做基于图像的驾驶员疲劳检测系统项目，其中行为检测包含驾驶员打电话和抽烟的危险行为，基于图像用之前的Dlib68特征点和opencv肤色和动态追踪图像处理已经远远达不到理想标准，我们需要获取脸部和手部关键点信息；为了精确识别打电话和抽烟或者是喝水等动作，这里引入了机器视觉中一经典开源模型——Openpose(人体姿态识别模型)。

Openpose简介

正文

(1)配置Openpose环境

Windows10 + python3.7 + anaconda3 + jupyter5.6.0

可以按步骤配置，也可以直接下载训练好的模型

解压，进入..\openpose-master\models目录，双击运行getModels.bat

等待加载模型pose_iter_584000.caffemodel：

等待时间较长，如中途中断，重新双击运行getModels.bat即可，

完整加载模型大小：100M，

保存位置：..\openpose-master\models\pose\body_25\pose_iter_584000.caffemodel

跑了一天完全加载5个模型如图，openpose总大小800M,编译真的太心寒了，这里快速下载入口。

OpenPose 人体姿态模型下载路径：

(2)主要模型介绍

目录

openpose分别检测：人脸、人手、人姿态(又分为三种数据集，mpi为较小的数据集)

一个简单地demo：Windows10 + python3.7 + anaconda3 + jupyter5.6.0

import cv2

import time

import numpy as np

import matplotlib.pyplot as plt

import os

# Load a Caffe Model

if not os.path.isdir('model'):

os.mkdir("model")

protoFile = "D:/myworkspace/JupyterNotebook/openpose/openpose-master/models/pose/mpi/pose_deploy_linevec_faster_4_stages.prototxt"

weightsFile = "D:/myworkspace/JupyterNotebook/openpose/openpose-master/models/pose/mpi/pose_iter_160000.caffemodel"

# Specify number of points in the model

nPoints = 15

POSE_PAIRS = [[0,1], [1,2], [2,3], [3,4], [1,5], [5,6], [6,7], [1,14], [14,8], [8,9], [9,10], [14,11], [11,12], [12,13] ]

net = cv2.dnn.readNetFromCaffe(protoFile, weightsFile)

# Read Image

im = cv2.imread("D:/myworkspace/JupyterNotebook/openpose/tf-pose-estimation-master/images/apink1.jpg")

im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)

inWidth = im.shape[1]

inHeight = im.shape[0]

# Convert image to blob

netInputSize = (368, 368)

inpBlob = cv2.dnn.blobFromImage(im, 1.0 / 255, netInputSize, (0, 0, 0), swapRB=True, crop=False)

net.setInput(inpBlob)

# Run Inference (forward pass)

output = net.forward()

# Display probability maps

plt.figure(figsize=(20,10))

plt.title('Probability Maps of Keypoints')

for i in range(nPoints):

probMap = output[0, i, :, :]

displayMap = cv2.resize(probMap, (inWidth, inHeight), cv2.INTER_LINEAR)

plt.subplot(3, 5, i+1); plt.axis('off'); plt.imshow(displayMap, cmap='jet')

# Extract points

# X and Y Scale

scaleX = float(inWidth) / output.shape[3]

scaleY = float(inHeight) / output.shape[2]

# Empty list to store the detected keypoints

points = []

# Confidence treshold

threshold = 0.1

for i in range(nPoints):

# Obtain probability map

probMap = output[0, i, :, :]

# Find global maxima of the probMap.

minVal, prob, minLoc, point = cv2.minMaxLoc(probMap)

# Scale the point to fit on the original image

x = scaleX * point[0]

y = scaleY * point[1]

if prob > threshold :

# Add the point to the list if the probability is greater than the threshold

points.append((int(x), int(y)))

else :

points.append(None)

# Display Points & Skeleton

imPoints = im.copy()

imSkeleton = im.copy()

# Draw points

for i, p in enumerate(points):

cv2.circle(imPoints, p, 8, (255, 255,0), thickness=-1, lineType=cv2.FILLED)

cv2.putText(imPoints, "{}".format(i), p, cv2.FONT_HERSHEY_SIMPLEX, 1, (255,0,0), 2, lineType=cv2.LINE_AA)

# Draw skeleton

for pair in POSE_PAIRS:

partA = pair[0]

partB = pair[1]

if points[partA] and points[partB]:

cv2.line(imSkeleton, points[partA], points[partB], (255, 255,0), 2)

cv2.circle(imSkeleton, points[partA], 8, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)

plt.figure(figsize=(20,10))

plt.subplot(121); plt.axis('off'); plt.imshow(imPoints);

#plt.title('Displaying Points')

plt.subplot(122); plt.axis('off'); plt.imshow(imSkeleton);

#plt.title('Displaying Skeleton')

plt.show()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

运行效果：

标出了15个人体姿态特征点

mpi15个关键点参考说明：

(3)主要模型参考图

# Body25: 25 points

# COCO: 18 points

# MPI: 15 points1

2

3

左：BODY_25， 右：COCO

Face

Hand

(4)人体姿态各模型检测

#!/usr/bin/python3

#!--*-- coding: utf-8 --*--

from __future__ import division

import cv2

import time

import numpy as np

import matplotlib.pyplot as plt

import os

class general_pose_model(object):

def __init__(self, modelpath, mode="BODY25"):

# 指定采用的模型

# Body25: 25 points

# COCO