-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathcamera_detection.py
213 lines (202 loc) · 8.33 KB
/
camera_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
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
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
import tkinter as tk
class FatigueDetection:
def __init__(self, model = './weights/ssd300_VOC_100000.pth'):
# 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-----')
self.show_result = 0
self.show_keypoint = 1
self.text_input_frame = None
self.img_mean = (104.0, 117.0, 123.0)
self.colors_tableau = [
(214, 39, 40), (23, 190, 207), (188, 189, 34), (188, 34, 188), (205, 108, 8)]
# Eye state list, suggest modifying based on fps
self.list_B = np.ones(15)
# Mouth state list, suggest modifying based on fps
self.list_Y = np.zeros(50)
# If list_Y has list_Y1, it is judged as yawning. Same as above, suggest modifying length
self.list_Y1 = np.ones(5)
self.blink_count = 0
self.yawn_count = 0
self.blink_start = time.time()
self.yawn_start = time.time()
self.blink_freq = 0.5
self.yawn_freq = 0
self.net = SSD()
self.net = torch.nn.DataParallel(self.net)
self.net.train(mode=False)
self.net.load_state_dict(torch.load(
model, map_location=lambda storage, loc: storage))
self.detect = Detect.apply
self.priors = utils.default_prior_box()
self.labels = VOC_CLASSES
# 调用摄像头
self.cap = cv2.VideoCapture(0)
self.max_fps = 0
self.detect_fatigue()
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
def get_frame_size(self):
return int(self.cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(self.cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
def detect_fatigue(self):
flag_B = True # 是否闭眼的flag
flag_Y = False
num_rec = 0 # 检测到的眼睛的数量
start = time.time() # 计时
ret, img = self.cap.read() # 读取图片
# 检测
x = cv2.resize(img, (300, 300)).astype(np.float32)
x -= self.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 = self.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 = self.colors_tableau[i]
if self.show_keypoint:
cv2.rectangle(img, (int(pt[0]), int(
pt[1])), (int(pt[2]), int(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')
self.list_B = np.append(self.list_B, 1) # 睁眼为‘1’
else:
# print(' 0:eye-closed')
self.list_B = np.append(self.list_B, 0) # 闭眼为‘0’
self.list_B = np.delete(self.list_B, 0)
if flag_Y:
self.list_Y = np.append(self.list_Y, 1)
else:
self.list_Y = np.append(self.list_Y, 0)
self.list_Y = np.delete(self.list_Y, 0)
# else:
# print('nothing detected')
# print(list)
# 实时计算PERCLOS
perclos = 1-np.average(self.list_B)
# print('perclos={:f}'.format(perclos))
if self.list_B[13] == 1 and self.list_B[14] == 0:
# 如果上一帧为’1‘，此帧为’0‘则判定为眨眼
print('----------------眨眼----------------------')
if self.show_result:
self.text_input_frame.textbox.tag_config("debug", foreground="grey")
self.text_input_frame.textbox.insert(tk.END, '<debug> 眨眼\n', 'debug')
self.blink_count += 1
blink_T = time.time()-self.blink_start
if blink_T > 10:
# 每10秒计算一次眨眼频率
self.blink_freq = self.blink_count/blink_T
self.blink_start = time.time()
self.blink_count = 0
print('blink_freq={:f}'.format(self.blink_freq))
# 检测打哈欠
# if Yawn(list_Y,list_Y1):
if (self.list_Y[len(self.list_Y)-len(self.list_Y1):] == self.list_Y1).all():
print('----------------------打哈欠----------------------')
if self.show_result:
self.text_input_frame.textbox.tag_config("debug", foreground="black", background="white")
self.text_input_frame.textbox.insert(tk.END, '<debug> ----------------------打哈欠----------------------\n', 'debug')
self.yawn_count += 1
self.list_Y = np.zeros(50)
# 计算打哈欠频率
yawn_T = time.time()-self.yawn_start
if yawn_T > 60:
self.yawn_freq = self.yawn_count/yawn_T
self.yawn_start = time.time()
self.yawn_count = 0
print('yawn_freq={:f}'.format(self.yawn_freq))
# 此处为判断疲劳部分
'''
想法1：最简单，但是太影响实时性
if(perclos>0.4 or blink_freq<0.25 or yawn_freq>5/60):
print('疲劳')
if(blink_freq<0.25)
else:
print('清醒')
'''
# 想法2：
if (perclos > 0.4):
print('疲劳')
if self.show_result:
self.text_input_frame.textbox.tag_config("debug", foreground="grey")
self.text_input_frame.textbox.insert(tk.END, '<debug> 疲劳\n', 'debug')
elif (self.blink_freq < 0.25):
print('疲劳')
if self.show_result:
self.text_input_frame.textbox.tag_config("debug", foreground="grey")
self.text_input_frame.textbox.insert(tk.END, '<debug> 疲劳\n', 'debug')
self.blink_freq = 0.5 # 如果因为眨眼频率判断疲劳，则初始化眨眼频率
elif (self.yawn_freq > 5.0/60):
print("疲劳")
if self.show_result:
self.text_input_frame.textbox.tag_config("debug", foreground="grey")
self.text_input_frame.textbox.insert(tk.END, '<debug> 疲劳\n', 'debug')
self.yawn_freq = 0 # 初始化，同上
# else:
# print('清醒')f
if self.show_keypoint:
T = time.time()-start
fps = 1/T # 实时在视频上显示fps
if fps > self.max_fps:
self.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)
self.image = img