LoginSignup
2
9

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

@mix_dvd(きむら しのぶ)

複数のRaspberry PiとMacを使って入室監視システムを作る

Last updated at Posted at 2018-03-24

はじめに

とあるイベントで出展した仕組みに興味を持つ人が多かったので、作り方を以下に整理し、公開します。

材料

  • Raspberry Pi 3 model B
  • USBカメラ
  • MacBook Air

Raspberry Piの設定

OSのインストールは終わっている状態とします。

MJPE-streamer

以下のコマンドでインストール。

cd ~/
sudo apt-get install -y cmake libjpeg8-dev build-essential git
git clone https://github.com/jacksonliam/mjpg-streamer.git
cd mjpg-streamer/mjpg-streamer-experimental
make

以下のコマンドを実行。

./start.sh

同じネットワーク上の別のPCからWebブラウザでアクセスして動作確認。

http://[RaspberryPiのIPアドレス]:8080/?action=stream

カメラの画像が表示されればOK.

起動設定

start.shに以下の行を追加。

cd /home/pi/mjpg-streamer/mjpg-streamer-experimental

rc.localに以下の行を追加。

/home/pi/mjpg-streamer/mjpg-streamer-experimental/start.sh &

Macの環境構築

手前味噌ですが、以下のURLに従ってMacの環境構築を行います。

Macで深層学習の環境をさくっと作る手順 with TensorFlow and OpenCV

Kerasをインストール

pip install keras

スクリプトを設置するフォルダに移動して以下のコマンドによりデータ用フォルダを作成。

mkdir faces
cd faces
mkdir train
mkdir test

顔画像収集スクリプト

Macのカメラを使って20個の顔画像を収集し、19個を学習用、1個をテスト用に分けて保存。

a001_make_face_data.py
# !/usr/bin/env python

import os
import sys
import cv2
import time
from datetime import datetime as dt

if len(sys.argv) < 2:
    print("Usage: a001_make_face_data.py [name]")
    exit()

name = sys.argv[-1]
print(name)

cascade_file = "haarcascade_frontalface_alt2.xml"
cascade = cv2.CascadeClassifier(cascade_file)

cnt_max = 21
cnt_face = 0
faces = []

DEVICE_ID = 0
cap = cv2.VideoCapture(DEVICE_ID)

time.sleep(1)
end_flag, c_frame = cap.read()
height, width, channels = c_frame.shape

while cnt_face < cnt_max:
    
    img = c_frame
    
    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    face_list = cascade.detectMultiScale(img_gray, minSize=(100, 100))
    
    for (pos_x, pos_y, w, h) in face_list:

        img_face = img[pos_y:pos_y+h, pos_x:pos_x+w]
        img_face = cv2.resize(img_face, (100, 100))
        faces.append(img_face)
    
    if len(face_list) > 0:
        cnt_face += 1
        print("\r", cnt_face, end="")
        
    time.sleep(1)
    end_flag, c_frame = cap.read()
    
cap.release()

print(len(faces))

num = 0
tdatetime = dt.now()
tstr = tdatetime.strftime('%Y%m%d%H%M')

path = '{}/faces/train/{}'.format(os.getcwd(), name)
print("学習データ", path)
os.makedirs(path)

for face in faces[:-1]:
    filename = '{}-{}.jpg'.format(tstr, num)
    print("\t", filename)
    cv2.imwrite('{}/{}'.format(path, filename), face)
    num += 1
    
path = '{}/faces/test/{}'.format(os.getcwd(), name)
print("テストデータ", path)
os.makedirs(path)

face = faces[-1]
filename = '{}-{}.jpg'.format(tstr, num)
print("\t", filename)
cv2.imwrite('{}/{}'.format(path, filename), face)

以下のコマンドで顔画像を収集。

chmod a+x a001_make_face_data.py
./a001_make_face_data.py [name]

顔画像学習用スクリプト

取得した顔画像を学習します。

a002_train_face_data.py
# !/usr/bin/env python

import os
import random
import cv2
import numpy as np
import pandas as pd

import gc

import keras
from keras.datasets import cifar10
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.utils import np_utils

path = "./faces/train"
dirs = os.listdir(path)
dirs = [f for f in dirs if os.path.isdir(os.path.join(path, f))]

label_dict = {}
i = 0

for dirname in dirs:
    label_dict[dirname] = i
    i += 1
    
def load_data(data_type):

    filenames, images, labels = [], [], []
    
    walk = filter(lambda _: not len(_[1]) and data_type in _[0], os.walk('faces'))
    
    for root, dirs, files in walk:
        filenames += ['{}/{}'.format(root, _) for _ in files if not _.startswith('.')]
    
    # Shuffle files
    random.shuffle(filenames)
    
    # Read, resize, and reshape images
    images = []
    for file in filenames:
        img = cv2.imread(file)
        img = cv2.resize(img, (32,32))
        images.append(img.astype(np.float32) / 255.0)
    images = np.asarray(images)
    
    for filename in filenames:
        label = np.zeros(len(label_dict))
        for k, v in label_dict.items():
           if k in filename:
                label[v] = 1.
        labels.append(label)
    labels = np.asarray(labels)
    
    return images, labels
    
def make_model(train_shape, label_shape):
    
    model = Sequential()

    model.add(Conv2D(32, (3, 3), padding="same", input_shape=train_shape[1:]))
    model.add(Activation('relu'))
    model.add(Conv2D(32, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.25))

    model.add(Conv2D(64, (3, 3), padding='same'))
    model.add(Activation('relu'))
    model.add(Conv2D(64, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.25))

    model.add(Flatten())
    model.add(Dense(512))
    model.add(Activation('relu'))
    model.add(Dropout(0.5))

    model.add(Dense(label_shape[1]))
    model.add(Activation('softmax'))

    model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])

    model.summary()
    
    return model
    
accuracys = []

train_images, train_labels = load_data('train')
test_images, test_labels = load_data('test')

print("train_images", len(train_images))
print("test_images", len(test_images))

print(train_images.shape, train_labels.shape)

model = make_model(train_images.shape, train_labels.shape)
model.fit(train_images, train_labels, batch_size=32, epochs=10)

# modelのテスト
score = model.evaluate(test_images, test_labels)
print('loss=', score[0])
print('accuracy=', score[1])

model_json_str = model.to_json()
open('./model_tf/face-model.json', 'w').write(model_json_str)

hdf5_file = "./model_tf/face-model.hdf5"
model.save_weights(hdf5_file)

gc.collect()

以下のコマンドで学習。

chmod a+x a002_train_face_data.py
./a002_train_face_data.py

顔識別用スクリプト

以下のスクリプトを作成。

a003_detect_face_data.py
# !/usr/bin/env python

import os
import sys
import cv2
import random
import numpy as np

import requests

from keras.models import model_from_json
import h5py

if len(sys.argv) < 2:
    print("Usage: a003_detect_face_data.py [IP Address]")
    exit()

ip = sys.argv[-1]
print(ip)

model = model_from_json(open('./model_tf/face-model.json').read())
model.load_weights('./model_tf/face-model.hdf5')

model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])

model.summary();

path = "./faces/train"
dirs = os.listdir(path)
dirs = [f for f in dirs if os.path.isdir(os.path.join(path, f))]

label_dict = {}

names = dirs

if __name__ == '__main__':
    
    # 定数定義
    ESC_KEY = 27     # Escキー
    INTERVAL= 33     # 待ち時間

    WINDOW_NAME = "detect"

    DEVICE_ID = 0

    # 分類器の指定
    cascade_file = "haarcascade_frontalface_alt2.xml"
    cascade = cv2.CascadeClassifier(cascade_file)
    
    img_url = 'http://{}:8080/?action=snapshot'.format(ip)

    # ウィンドウの準備
    cv2.namedWindow(WINDOW_NAME)
    
    font = cv2.FONT_HERSHEY_DUPLEX
    font_size = 1.0
    
    # 変換処理ループ
    while True:

        req = requests.get(img_url)
        
        img_buf = np.fromstring(req.content, dtype='uint8')
        img = cv2.imdecode(img_buf, 1)
        
        img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        face_list = cascade.detectMultiScale(img_gray, minSize=(50, 50))
        
        for (pos_x, pos_y, w, h) in face_list:
            
            img_face = img[pos_y:pos_y+h, pos_x:pos_x+w]
            
            img_face = cv2.resize(img_face, (32, 32))
            
            test_images = []
            test_images.append(img_face / 255.0)
            test_images = np.asarray(test_images)
            
            results = model.predict(test_images)
            text = names[np.argmax(results[0])]

            color = (0, 0, 225)
            pen_w = 2
            cv2.putText(img,text,(pos_x,pos_y - 10), font, font_size, color)
            cv2.rectangle(img, (pos_x, pos_y), (pos_x+w, pos_y+h), color, thickness = pen_w)
        
        # フレーム表示
        cv2.imshow(WINDOW_NAME, img)

        # Escキーで終了
        key = cv2.waitKey(INTERVAL)
        if key == ESC_KEY:
            break

    cv2.destroyAllWindows()

以下のコマンドを実行。

chmod a+x a003_detect_face_data.py
./a003_detect_face_data.py [RaspberryPiのIPアドレス] &

複数のRaspberryPiがあれば、上記コマンドの2行目を繰り返し実行することで、複数カメラで監視を実行可能。

さいごに

これで識別した結果をLINEやMessengerなどで通知する仕組みを作れば、色々なところで利用できると考えます。

2
9
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
Sign upLogin
2
9

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?