You are on page 1of 8

Face Recognition based Home Security System

Using IoT
I. ABSTRACT
The topic illustrates about an intelligent system for home security.
More specifically “internet of things” is utilized in this system for
monitoring the person(s) entering in a home using a smart phone and a
web application. By using this system the user is capable of monitoring
the home at anytime and anywhere using the application. The current
locking devices used by the persons in homes or banks are not capable
of providing a smart security system, as the locking devices are easily
breakable. Face recognition technique is implemented in this system i.e.
the authorization persons are only allowed to enter in the home. The
proposed system embodied with a camera module and ultrasonic sensor
for initiating the face recognition process, further a raspberry pi
installed in the system for generating a command signal for performing
the desired functionalities. The application permits the owner to add
new faces to the data base using the web application.

II. INTRODUCTION
The system's primary purpose is to protect the valuable things of the
home and the family members from burglars. A survey discloses that
burglary occurs in every 13 seconds and thus 240 burglaries occur in an
hour. For avoiding this particular situation, the proposed system is
developed for providing proper security to the homes that is based on
face detection and IOT. IOT allows for detecting a person's motion
before the entrance. The system also capable of detecting the person’s
distance and position of the person from the door. LBP technique is
utilized in this system for obtaining 80% of accuracy. For performing
all the operationsof the system, the person has to stand in front of the
door. The camera identifies the facial characteristics of a person and
decides that a person is allowed or not. After matching the face, door
will automatically open or remains in close position according to face
identification.

III. METHODOLOGY / PROCEDURE


The sensors installed in the system are preferably embedded sensor that
senses any motion or disturbance occurs in its surroundings and also it
consumes very less power. A stepper motor incorporated with the system
for locking/unlocking of the door by means of utilizing the command
signals received from the raspberry pi.
Further the raspberry pi controller comprises of face detection
algorithm that is used to prevent the unauthorized entrance in homes.
The web application and the user interface such as mobiles, tablets used
for making secure connectivity with the internet. The aforementioned
systems working can be distinguished as following parts. Fig 1
illustrates about the block diagram representation of the system.

Fig.1. Block diagram representation of components

Fig.2. Diagrammatic representation of hardware components


IV. IMPLEMENTATION

The system is implemented with the use of the following sensors.

Information generation unit


The Information generation unit comprises multiple sensors to Infrared
sensor HC SR04 ultrasonic sensors for detecting the face of a person.
The sensors further attached to this raspberry pi for transmitting each
collected data of the person’s face. The proposed system includes a
stepper motor, alarm, doorbell unit, camera module, and 6v battery. Fig
2 illustrates the hardware structure.

PIR Sensor
The sensor used in the system is preferably a PIR sensor (passive
infrared sensor) used for extracting the face of the person standing in
front of the sensor. Further the sensor is operated in low power and in
low maintenance. The sensor measures approx. the range between such
door and person, simultaneously The data of the device transmitted to
that raspberry pi controller for detecting each person. Fig 3 discloses
about the diagram of the PIR sensor.
Ultrasonic Sensor
Ultrasonic sensors incorporated with the method for identifying the
distance from person to door. Generally the ultrasonic sensor radiates
ultrasonic waves towards the detected object and the waves collided with
the object[6]. Further the waves reflected after collided with the object,
thus the distance is measured from the sensor to the person as shown in
fig.4.

Stepper Motor
The motor installed with the system is preferably a stepper motor.
Further the stepper motor can be either synchronous or asynchronous
depending up on the current provided to it. The stepper motor
functionalities are controlled by the raspberry pi[7][8][9]. After
recognising the authorized face the raspberry pi transmits a command
signal for actuating the stepper motor. In case if the unauthorized
person tries to breach the door lock, a quick signal will send through
email and SMS. Fig.5. discloses the diagram of the stepper motor.

Information Processing Unit


The Information processing unit comprises the raspberry pi three
models B+ could consist of the face detection algorithm. The
raspberry pi comprising of the below-mentioned features:
1. Bluetooth: 4.1
2. Graphic Processing Unit:300MH

3. Size: 85.60mm × 56.5mm

4. Network: 10/110Mbps Ethernet and 805.11n Cellular LAN

Local Binary Pattern (LBP)

The Taken image processed for face detection using the LBP
technique. Further the technique has been implemented by using
python in system and open cv. Face identification is processed through
extracting the shape and texture information of the face. For face
detection the image of the face is cropped followed by converting the
images in to equivalent sizes. The LBP algorithm is applied for
detecting the face of a person. Each feature of the detected image is
used for providing accuracy. Fig 6 shows the feature extraction process.

IV. PROCESS
All the results of face detection stored in the database by using the mobile
application. The stored data in the data base can be accessible anywhere by
using the smart mobiles and mobile application. If an unauthorized person is
detected by the system the alert signal transmitted to the user through the
mobile application. In the absence of electricity the system is alsocapable of
completing all its functionalities.
Fig.7. Flow chart of the system’s working method

The above mentioned components and methodology illustrates in a flow


diagram mentioned below Fig 7. The proposed system is IOT based and
raspberry pi is used in the system for operating and monitoring the
working of all the components and software’s. The system initiates its
functionalities when a motion is detected in the surrounding of the
ultrasonic sensor. Further the sensor analyses the approx. distance of the
person from the door. The face is detected using the LBP pattern, further a
signal based up on the information of the detected face transmitted to
raspberry pi. This raspberry pi matches the face with the stored face of the
cloud storage. If the face matches the door unlocked through the stepper
motor in accordance to the command signal of the raspberry pi. In case of
unauthorized entrance inside the house the raspberry pi transmits a SMS
regarding the person’s entrance. If the intruder tries to break the lock, then
high alert signal transmits through the buzzer.

VI. RESULT

The LBP algorithm determines the person is authorized or not. In case of


unauthorized person is detected LED‘s are installed in the system that
turns in to red. If the raspberry pi founds any unauthorized entrance the
notifications also transmitted to the owner of the home by means of a SMS
or an email as shown in Fig 9.

Fig.8. Notification via SMS and email

VII. CONCLUSION

The above illustrates about a door locking system based on face


detection technique. The whole system’s functionalities are working by
using IOT. The proposed system provides the best security methods for
detecting the entrance of unauthorized persons inside the homes for
securing valuable things and family members. Also advancement can
be done in to the system for providing more accuracy. The system
utilizes mobile application that makes the system accessible anywhere
and anytime through a user interface.

You might also like