forked from PengfeiM/Fatigue-Driven-Detection-Based-on-CNN
-
Notifications
You must be signed in to change notification settings - Fork 0
/
video_detection.py
190 lines (168 loc) · 5.61 KB
/
video_detection.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
# -*- coding: utf-8 -*-
"""
Created on Fri Apr 26 15:58:49 2019
@author: 朋飞
"""
from torch.autograd import Variable
from detection import *
from ssd_net_vgg import *
from voc0712 import *
import torch
import torch.nn as nn
import numpy as np
import cv2
import utils
import torch.backends.cudnn as cudnn
import time
#检测cuda是否可用
if torch.cuda.is_available():
print('-----gpu mode-----')
torch.set_default_tensor_type('torch.cuda.FloatTensor')
else:
print('-----cpu mode-----')
colors_tableau=[ (214, 39, 40),(23, 190, 207),(188, 189, 34),(188,34,188),(205,108,8)]
def Yawn(list_Y,list_y1):
list_cmp=list_Y[:len(list_Y1)]==list_Y1
for flag in list_cmp:
if flag==False:
return False
return True
#初始化网络
net=SSD()
net=torch.nn.DataParallel(net)
net.train(mode=False)
net.load_state_dict(torch.load('./weights/ssd300_VOC_100000.pth',map_location=lambda storage,loc: storage))
if torch.cuda.is_available():
net = net.cuda()
cudnn.benchmark = True
img_mean=(104.0,117.0,123.0)
#打开视频文件，file_name改成0即为打开摄像头
file_name='C:/Users/HP/Desktop/9-FemaleNoGlasses.avi'
cap=cv2.VideoCapture(file_name)
max_fps=0
#保存检测结果的List
#眼睛和嘴巴都是，张开为‘1’，闭合为‘0’
video_fps=20#视频fps=20
list_B=np.ones(video_fps*3)#眼睛状态List,建议根据fps修改，视频fps=20
list_Y=np.zeros(video_fps*10)#嘴巴状态list，10s
list_Y1=np.ones(int(video_fps*1.5))#如果在list_Y中存在list_Y1，则判定一次打哈欠(大约1.5s)，
list_Y1[int(video_fps*1.5)-1]=0#从持续张嘴到闭嘴判定为一次打哈欠
list_blink=np.ones(video_fps*10)#大约是记录10S内信息，眨眼为‘1’，不眨眼为‘0’
list_yawn=np.zeros(video_fps*30)#大约是半分钟内打哈欠记录，打哈欠为‘1’，不打哈欠为‘0’
#blink_count=0#眨眼计数
#yawn_count=0
#blink_start=time.time()#炸眼时间
#yawn_start=time.time()#打哈欠时间
blink_freq=0.5
yawn_freq=0
#开始检测，按‘q’退出
while cap.isOpened():
flag_B=True#是否闭眼的flag
flag_Y=False#张嘴flag
num_rec=0#检测到的眼睛的数量
start=time.time()#计时
ret,img=cap.read()#读取图片
#检测
x=cv2.resize(img,(300,300)).astype(np.float32)
x-=img_mean
x=x.astype(np.float32)
x=x[:,:,::-1].copy()
x=torch.from_numpy(x).permute(2,0,1)
xx=Variable(x.unsqueeze(0))
if torch.cuda.is_available():
xx=xx.cuda()
y=net(xx)
softmax=nn.Softmax(dim=-1)
# detect=Detect(config.class_num,0,200,0.01,0.45)
detect = Detect.apply
priors=utils.default_prior_box()
loc,conf=y
loc=torch.cat([o.view(o.size(0),-1)for o in loc],1)
conf=torch.cat([o.view(o.size(0),-1)for o in conf],1)
detections=detect(
loc.view(loc.size(0),-1,4),
softmax(conf.view(conf.size(0),-1,config.class_num)),
torch.cat([o.view(-1,4) for o in priors],0),
config.class_num,
200,
0.7,
0.45
).data
labels=VOC_CLASSES
top_k=10
#将检测结果放置于图片上
scale=torch.Tensor(img.shape[1::-1]).repeat(2)
for i in range(detections.size(1)):
j=0
while detections[0,i,j,0]>=0.4:
score=detections[0,i,j,0]
label_name=labels[i-1]
if label_name=='closed_eye':
flag_B=False
if label_name=='open_mouth':
flag_Y=True
display_txt='%s:%.2f'%(label_name,score)
pt=(detections[0,i,j,1:]*scale).cpu().numpy()
coords=(pt[0],pt[1]),pt[2]-pt[0]+1,pt[3]-pt[1]+1
color=colors_tableau[i]
cv2.rectangle(img,(pt[0],pt[1]),(pt[2],pt[3]),color,2)
cv2.putText(img,display_txt,(int(pt[0]),int(pt[1])+10),cv2.FONT_HERSHEY_SIMPLEX, 0.4, (255,255,255), 1, 8)
j+=1
num_rec+=1
if num_rec>0:
if flag_B:
#print(' 1:eye-open')
list_B=np.append(list_B,1)#睁眼为‘1’
else:
#print(' 0:eye-closed')
list_B=np.append(list_B,0)#闭眼为‘0’
list_B=np.delete(list_B,0)
if flag_Y:
list_Y=np.append(list_Y,1)
else:
list_Y=np.append(list_Y,0)
list_Y=np.delete(list_Y,0)
else:
print('nothing detected')
#print(list)
if list_B[13]==1 and list_B[14]==0:
#如果上一帧为’1‘，此帧为’0‘则判定为眨眼
print('----------------眨眼----------------------')
list_blink=np.append(list_blink,1)
else:
list_blink=np.append(list_blink,0)
list_blink=np.delete(list_blink,0)
#检测打哈欠
#if Yawn(list_Y,list_Y1):
if (list_Y[len(list_Y)-len(list_Y1):]==list_Y1).all():
print('----------------------打哈欠----------------------')
list_Y=np.zeros(50)#此处是检测到一次打哈欠之后将嘴部状态list全部置‘0’，考虑到打哈欠所用时间较长，所以基本不会出现漏检
list_yawn=np.append(list_yawn,1)
else:
list_yawn=np.append(list_yawn,0)
list_yawn=np.delete(list_yawn,0)
#实时计算PERCLOS perblink,peryawn
#即计算平均闭眼时长百分比，平均眨眼百分比，平均打哈欠百分比
perclos=1-np.average(list_B)
perblink=np.average(list_blink)
peryawn=np.average(list_yawn)
#print('perclos={:f}'.format(perclos))
#此处为判断疲劳部分
#想法1：两个频率计算改为实时的，所以此处不再修改
if(perclos>0.4 or perblink<2.5/(10*video_fps) or peryawn>3/(30*video_fps)):
print('疲劳')
else:
print('清醒')
T=time.time()-start
fps=1/T#实时在视频上显示fps
if fps>max_fps:
max_fps=fps
fps_txt='fps:%.2f'%(fps)
cv2.putText(img,fps_txt,(0,10),cv2.FONT_HERSHEY_SIMPLEX, 0.4, (255,255,255), 1, 8)
cv2.imshow("ssd",img)
if cv2.waitKey(100) & 0xff == ord('q'):
break
#print("-------end-------")
cap.release()
cv2.destroyAllWindows()
#print(max_fps)