2018.2.19版本二，多人单目标判断
"""
import cv2
import sys
import gc
from face_train_use_keras import Model
if __name__ == '__main__':
if len(sys.argv) != 2:
print("Usage:%s camera_id\r\n" % (sys.argv[0]))
sys.exit(0)
# 加载模型
model = Model()
model.load_model(
file_path=
'/home/yuwangwang/FaceRecognization_Version1/model/me.face.model.h5')
# 框住人脸的矩形边框颜色,颜色为绿色
color = (0, 255, 0)
# 捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(int(sys.argv[1]))
# 人脸识别分类器本地存储路径
cascade_path = "/usr/local/share/OpenCV/haarcascades/haarcascade_frontalface_alt2.xml"
# 循环检测识别人脸
while True:
#-*- coding: utf-8 -*-
import cv2
import sys
import gc
from face_train_use_keras import Model
if __name__ == '__main__':
if len(sys.argv) != 2:
print("Usage:%s camera_id\r\n" % (sys.argv[0]))
sys.exit(0)
#load model
model = Model()
model.load_model(file_path='me.face.model.h5')
color = (0, 255, 0)
cap = cv2.VideoCapture(int(sys.argv[1]))
cascade_path = "/usr/local/share/OpenCV/haarcascades/haarcascade_frontalface_alt2.xml"
while True:
_, frame = cap.read()
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#load classfacation
cascade = cv2.CascadeClassifier(cascade_path)
#where is face
def start(self):
global string
#加载模型
model = Model()
model.load_model(file_path='./model/me.face.model.h5')
#框住人脸的矩形边框颜色
color = (0, 255, 0)
#捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(0)
#人脸识别分类器本地存储路径
cascade_path = "E:/Python/Lib/site-packages/cv2/data/haarcascade_frontalface_alt2.xml"
#循环检测识别人脸
while True:
_, frame = cap.read() #读取一帧视频
#图像灰化，降低计算复杂度
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#使用人脸识别分类器，读入分类器
cascade = cv2.CascadeClassifier(cascade_path)
#利用分类器识别出哪个区域为人脸
faceRects = cascade.detectMultiScale(frame_gray,
scaleFactor=1.2,
minNeighbors=3,
minSize=(32, 32))
id = ''
if len(faceRects) > 0:
for faceRect in faceRects:
x, y, w, h = faceRect
#截取脸部图像提交给模型识别这是谁
image = frame[y - 10:y + h + 10, x - 10:x + w + 10]
faceID = model.face_predict(image)
#如果是“我”
if faceID == 0:
cv2.rectangle(frame, (x - 10, y - 10),
(x + w + 10, y + h + 10),
color,
thickness=2)
#文字提示是谁
cv2.putText(
frame,
'KK',
(x + 30, y + 30), #坐标
cv2.FONT_HERSHEY_SIMPLEX, #字体
1, #字号
(255, 0, 255), #颜色
2) #字的线宽
id = "KK"
else:
cv2.rectangle(frame, (x - 10, y - 10),
(x + w + 10, y + h + 10),
color,
thickness=2)
#文字提示是谁
cv2.putText(
frame,
'WWL',
(x + 30, y + 30), #坐标
cv2.FONT_HERSHEY_SIMPLEX, #字体
1, #字号
(255, 0, 255), #颜色
2) #字的线宽
id = "wwl"
cv2.imshow("faceRecognition", frame)
#等待10毫秒看是否有按键输入
k = cv2.waitKey(10)
#如果输入q则退出循环
if k & 0xFF == ord('='):
localtime = time.asctime(time.localtime(time.time()))
print("wwl的打卡时间为时间为 :", localtime)
self.onclock(id, localtime)
file = './res/wwl.txt'
with open(file, 'a+') as f:
f.write(localtime + '\n') #加\n换行显示
break
#释放摄像头并销毁所有窗口
cap.release()
cv2.destroyAllWindows()
import gc from face_train_use_keras import Model import os import path import pymysql import time import pandas as pd ''' ''' path = "E:\\python\\face_5.17\\data" l = [] l = os.listdir(path) model = Model() model.load_model(file_path='./model/face_detect607.h5') #框住人脸的矩形边框颜色 color = (0, 255, 0) color2 = (255, 0, 0) color3 = (0, 0, 255) #捕获指定摄像头的实时视频流 cap = cv2.VideoCapture(0) #"rtsp://*****:*****@192.168.1.108/cam/realmonitor?channel=1&subtype=0" #人脸识别分类器本地存储路径 cascade_path = "E:/opencv/data/haarcascades/haarcascade_frontalface_alt2.xml" #conn = pymysql.connect(host='192.168.1.102', port=3306, user='******', passwd='123456',db="dect",charset='utf8') #cursor = conn.cursor()
#-*-coding:utf-8 -*-
__auther__ = 'Alex'
# -*- coding: utf-8 -*-
import cv2
import dlib
import time
from face_train_use_keras import Model
if __name__ == '__main__':
# 加载模型
model = Model()
model.load_model(file_path='./model/jiaYouBG.model.h5')
# 框住人脸的矩形边框颜色
color = (0, 255, 0)
# 捕获指定摄像头的实时视频流
cap = cv2.VideoCapture("./video/2.mp4")
inWidth = 300
inHeight = 300
inScaleFactor = 1.0
net = cv2.dnn.readNetFromCaffe('openCVmodel/deploy.prototxt','openCVmodel/res10_300x300_ssd_iter_140000.caffemodel')
predicterPath = "model/shape_predictor_68_face_landmarks.dat"
sp = dlib.shape_predictor(predicterPath)
# 人脸识别分类器本地存储路径
cascade_path = "/home/alex/桌面/face_recgnition/openCVmodel/haarcascades/haarcascade_frontalface_alt2.xml"
# -*- coding: utf-8 -*-
import cv2
import sys
import gc
from face_train_use_keras import Model
MODEL_PATH = './me.face.model.h5'
if __name__ == '__main__':
if len(sys.argv) != 2:
print("Usage:%s camera_id\r\n" % (sys.argv[0]))
sys.exit(0)
# 加载模型
model = Model()
model.load_model(MODEL_PATH)
faceID = model.face_predict(image)
print(faceID)
def start_face_recog(self, msg):
self.stop_dialog()
model = Model()
model.load_model(file_path='./face_recog/model/face.model.h5')
color = (0, 255, 0)
cascade_path = "/usr/local/share/OpenCV/haarcascades/haarcascade_frontalface_alt2.xml"
num = 0
pos_person = {'jiashi': 0, 'yushizhuo': 0}
while num <= 2:
frame = self.get_img()
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
cascade = cv2.CascadeClassifier(cascade_path)
faceRects = cascade.detectMultiScale(frame_gray,
scaleFactor=1.2,
minNeighbors=3,
minSize=(32, 32))
if len(faceRects) > 0:
if num == 0:
self.show_image("recognising111.gif")
sentence = "Please do not move. I am trying to find out who you \\pau=30\\ are"
self.say(sentence)
num += 1
for faceRect in faceRects:
x, y, w, h = faceRect
image = frame[y - 10:y + h + 10, x - 10:x + w + 10]
faceID = model.face_predict(image)
if faceID == 0:
cv2.rectangle(frame, (x - 10, y - 10),
(x + w + 10, y + h + 10),
color,
thickness=2)
# 文字提示是谁
cv2.putText(
frame,
'jiashi',
(x + 30, y + 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
pos_person['jiashi'] += 1
elif faceID == 1:
cv2.rectangle(frame, (x - 10, y - 10),
(x + w + 10, y + h + 10),
color,
thickness=2)
# 文字提示是谁
cv2.putText(
frame,
'yushizhuo',
(x + 30, y + 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
pos_person['yushizhuo'] += 1
else:
pass
cv2.namedWindow("face_pre", cv2.WINDOW_NORMAL)
cv2.imshow("face_pre", frame)
k = cv2.waitKey(1)
if k & 0xFF == ord('q'):
break
cv2.destroyAllWindows()
# 经过3次取值，获取的最有可能的人
print pos_person
self.current_person = max(pos_person, key=pos_person.get)
this_per = self.get_history_item(self.current_person)
print "========================"
item_list = list(this_per[self.current_person])
print item_list
image_name = self.current_person + '.png'
# 展示人的身份卡
self.show_image(image_name)
sentence = "I see, you must be " + self.current_person + "\\pau=100\\, do you like the " \
+ item_list[len(item_list) - 1] + " you bought last time ?"
self.say(sentence)
self.start_dialog()
print('\033[0;32m [Kamerider I] Your name is ' + self.current_person +
"\033[0m")
# -*- coding: utf-8 -*-
import cv2
from face_train_use_keras import Model
if __name__ == '__main__':
model = Model()
# 加载模型
model.load_model(file_path='my_transfer.h5')
# 框住人脸的矩形边框颜色
color = (0, 255, 0)
# 捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(0)
# 人脸识别分类器本地存储路径
cascade_path = "haarcascade_frontalface_default.xml"
# 循环检测识别人脸
while True:
_, frame = cap.read() # 读取一帧视频
# 图像灰化，降低计算复杂度
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# 使用人脸识别分类器，读入分类器
cascade = cv2.CascadeClassifier(cascade_path)
# 利用分类器识别出哪个区域为人脸
faceRects = cascade.detectMultiScale(frame_gray, scaleFactor=1.2,
def show():
#加载模型
model = Model()
model.load_model(file_path='./model/me.face.model.h5')
#框住人脸的矩形边框颜色
color = (0, 255, 0)
#捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(0)
#人脸识别分类器本地存储路径
cascade_path = "/home/binge/anaconda3/pkgs/opencv3-3.1.0-py36_0/share/OpenCV/haarcascades/haarcascade_frontalface_alt2.xml"
#循环检测识别人脸
while True:
_, frame = cap.read() #读取一帧视频
#图像灰化，降低计算复杂度
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#使用人脸识别分类器，读入分类器
cascade = cv2.CascadeClassifier(cascade_path)
#利用分类器识别出哪个区域为人脸
faceRects = cascade.detectMultiScale(frame_gray,
scaleFactor=1.2,
minNeighbors=3,
minSize=(32, 32))
if len(faceRects) > 0:
for faceRect in faceRects:
x, y, w, h = faceRect
#截取脸部图像提交给模型识别这是谁
image = frame[y - 10:y + h + 10, x - 10:x + w + 10]
faceID = model.face_predict(image)
#如果是“我”
if faceID == 0:
cv2.rectangle(frame, (x - 10, y - 10),
(x + w + 10, y + h + 10),
color,
thickness=2)
#文字提示是谁
cv2.putText(
frame,
'me',
(x + 30, y + 30), #坐标
cv2.FONT_HERSHEY_SIMPLEX, #字体
1, #字号
(255, 0, 255), #颜色
2) #字的线宽
else:
pass
cv2.imshow("识别", frame)
#等待10毫秒看是否有按键输入
k = cv2.waitKey(10)
#如果输入q则退出循环
if k & 0xFF == ord('q'):
break
#释放摄像头并销毁所有窗口
cap.release()
cv2.destroyAllWindows()
#-*- coding: utf-8 -*-
import cv2
import sys
import gc
from face_train_use_keras import Model
from keras import backend as k
if __name__ == '__main__':
#加载模型
model = Model()
model.load_model(file_path='model_weight.h5')
#框住人脸的矩形边框颜色
color = (0, 255, 0)
#捕获指定摄像头的实时视频流
#cap = cv2.VideoCapture('D:/tmp/vedio/123.mp4')
cap = cv2.VideoCapture('D:/tmp/vedio/mengmeng.mp4')
#人脸识别分类器本地存储路径
cascade_path = "C:/projects/opencv-python/opencv/modules/objdetect/src/cascadedetect/haarcascade_frontalface_alt2.xml"
#循环检测识别人脸
while True:
ret, frame = cap.read() #读取一帧视频
#图像灰化，降低计算复杂度
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#使用人脸识别分类器，读入分类器
cascade = cv2.CascadeClassifier(cascade_path)
#利用分类器识别出哪个区域为人脸
gray2 = gray.astype(np.float32)
dst = gray2 - blockImage2
dst = dst.astype(np.uint8)
dst = cv2.GaussianBlur(dst, (3, 3), 0)
dst = cv2.cvtColor(dst, cv2.COLOR_GRAY2BGR)
return dst
############################################################################################
############################################################################################
if __name__ == '__main__':
# 加载模型
model = Model()
model.load_model(file_path='./model/test.face.model.h5')
# 框住人脸的矩形边框颜色
color = (0, 255, 0)
# 捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(0)
# 人脸识别分类器本地存储路径
cascade_path = "C:\\Users\\Administrator\\Desktop\\Graduation design\\code\\csdn\\haarcascade_frontalface_alt2.xml"
# 循环检测识别人脸
while True:
_, frame = cap.read() # 读取一帧视频
# frame=cv2.imread("dupei.jpg")#########################################################################################################测试使用
def read_capture(detector):
# 加载模型
model = Model()
model.load_model(file_path='./model/face.model.h5')
num = 100
video = cv2.VideoCapture(0)
if video.isOpened():
success, frame = video.read()
while success:
im = frame.copy()
# dlib的人脸检测器只能检测80x80和更大的人脸，如果需要检测比它小的人脸，需要对图像上采样，一次上采样图像尺寸放大一倍
# rects = detector(img,1) #1次上采样
rects = detector(im, 0)
print rects
for rect in rects: # rect.left(),rect.top(),rect.right(),rect.bottom()
img2recog = im[rect.top() - 10:rect.bottom() + 10,
rect.left() - 10:rect.right() + 10]
faceID = model.face_predict(img2recog)
if faceID == 0:
print "======================================================", faceID
cv2.rectangle(im, (rect.left(), rect.top()),
(rect.right(), rect.bottom()), (0, 0, 255),
2, 8)
# 文字提示是谁
cv2.putText(
im,
'jiashi',
(rect.left() + 30, rect.top() - 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
elif faceID == 1:
print "======================================================", faceID
cv2.rectangle(im, (rect.left(), rect.top()),
(rect.right(), rect.bottom()), (0, 0, 255),
2, 8)
# 文字提示是谁
cv2.putText(
im,
'renzhongxing',
(rect.left() + 30, rect.top() - 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
elif faceID == 2:
print "======================================================", faceID
cv2.rectangle(im, (rect.left(), rect.top()),
(rect.right(), rect.bottom()), (0, 0, 255),
2, 8)
# 文字提示是谁
cv2.putText(
im,
'jingzhibo',
(rect.left() + 30, rect.top() - 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
elif faceID == 3:
print "======================================================", faceID
cv2.rectangle(im, (rect.left(), rect.top()),
(rect.right(), rect.bottom()), (0, 0, 255),
2, 8)
# 文字提示是谁
cv2.putText(
im,
'zhanghaiyang',
(rect.left() + 30, rect.top() - 30), # 坐标
cv2.FONT_HERSHEY_SIMPLEX, # 字体
1, # 字号
(255, 0, 255), # 颜色
2) # 字的线宽
cv2.imshow('capture face detection', im)
if cv2.waitKey(1) >= 0:
break
success, frame = video.read()
cv2.destroyAllWindows()
video.release()
def __init__(self):
self.model = Model()
self.model.load_model(file_path='/home/yuwangwang/FaceRecognization_Version2/model/me.face.model.h5')
class Camera_reader(object):
# 在初始化camera的时候建立模型，并加载已经训练好的模型
def __init__(self):
self.model = Model()
self.model.load_model(file_path='/home/yuwangwang/FaceRecognization_Version2/model/me.face.model.h5')
def build_camera(self, camera_id, path_name):
# opencv文件中人脸级联文件的位置，用于帮助识别图像或者视频流中的人脸
cascade = cv2.CascadeClassifier('/usr/local/share/OpenCV/haarcascades/haarcascade_frontalface_alt2.xml')
# 读取dataset数据集下的子文件夹名称
name_list = read_name_list(path_name)
# 框住人脸的矩形边框颜色,颜色为绿色
color = (0, 255, 0)
# 打开摄像头并开始读取画面
# 捕获指定摄像头的实时视频流
cap = cv2.VideoCapture(camera_id)
success, frame = cap.read()
while True:
success, frame = cap.read()
# 图像灰化
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# 利用分类器识别出哪个区域为人脸
# 参数1：image--待检测图片，一般为灰度图像加快检测速度；
# 参数2：objects--被检测物体的矩形框向量组；
# 参数3：scaleFactor--表示在前后两次相继的扫描中，搜索窗口的比例系数。默认为1.1即每次搜索窗口依次扩大10%;
# 参数4：minNeighbors--表示构成检测目标的相邻矩形的最小个数(默认为3个)。
# 如果组成检测目标的小矩形的个数和小于 min_neighbors - 1 都会被排除。
# 如果min_neighbors 为 0, 则函数不做任何操作就返回所有的被检候选矩形框，
# 这种设定值一般用在用户自定义对检测结果的组合程序上；
# 参数5：flags--要么使用默认值，要么使用CV_HAAR_DO_CANNY_PRUNING，如果设置为
# CV_HAAR_DO_CANNY_PRUNING，那么函数将会使用Canny边缘检测来排除边缘过多或过少的区域，
# 因此这些区域通常不会是人脸所在区域；
# 参数6、7：minSize和maxSize用来限制得到的目标区域的范围。
faceRects = cascade.detectMultiScale(frame_gray , 1.3, 5) # 识别人脸
#faceRects = cascade.detectMultiScale(frame_gray, scaleFactor = 1.2, minNeighbors = 3, minSize = (32, 32))
for (x, y, w, h) in faceRects:
image = frame[y - 10: y + h + 10, x - 10: x + w + 10]
label,prob = self.model.face_predict(image) # 利用模型对cv2识别出的人脸进行比对
if prob >0.8: # 如果模型认为概率高于70%则显示为模型中已有的label
show_name = name_list[label]
else:
show_name = 'Stranger'
# 显示名字的坐标，字体，字号为1，颜色为粉色，字的线宽为2
cv2.putText(frame, show_name, (x, y - 20), cv2.FONT_HERSHEY_SIMPLEX, 1, (255,0,255), 2)
# 在人脸区域画一个正方形出来
frame = cv2.rectangle(frame, (x, y), (x + w, y + h), color, thickness = 2)
cv2.imshow("Camera", frame)
# 等待10毫秒看是否有按键输入
k = cv2.waitKey(10)
# 如果输入q则退出循环
if k & 0xFF == ord('q'):
break
# 释放摄像头并销毁所有窗口
cap.release()
cv2.destroyAllWindows()
return camera_id, path_name
# -*- coding: utf-8 -*-
"""
Created on Mon Mar 26 21:34:23 2018
@author: 杨玉林
"""
import cv2
from face_train_use_keras import Model
if __name__ == '__main__':
model = Model()
model.load_model(file_path='D:\\face detection\\data\\me.face.model.h5')
color = (0, 255, 0)
cap = cv2.VideoCapture(0)
cascade_path = 'D:\\jiance\\opencv\\build\\etc\\haarcascades\\haarcascade_frontalface_alt2.xml'
while True:
_, frame = cap.read()
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
cascade = cv2.CascadeClassifier(cascade_path)
faceRects = cascade.detectMultiScale(frame_gray,
scaleFactor=1.2,
minNeighbors=3,