You are on page 1of 2

Door Lock System based on Face Recognition using

Arduino

Abstract: a key to access for unlocking the door. This Face Recognition
Technique can keep the assets safe and secure from any
In this paper, we have discussed a door lock system based on unwanted access.
face recognition using Arduino for safety purposes. The most
important feature of any home security is monitoring who
enters and leaves the house. Instead of doing this process with
passwords, we can use unique faces as they are unique for Proposed Model Architecture:
different individuals. We used python and OpenCV libraries Block Diagram:
for software Implementation. We used the Haar-Cascade
algorithm for detecting the faces and the LBPH algorithm to
get clear images of input faces. When the recognition
algorithms from the OpenCV library recognize your face and
if the face matches with the database, then the door will be
unlocked, else the door will be locked. This is how the system
works.

Introduction:

Biometrics are used in many systems that involve security. In


the face recognition algorithm, input face will be compared
with the images stored in the database to verify the person.
The aim is to search for a face in the database, which has the
highest similarity with the input face. This paper deals with Main parts of block diagram:
the idea of a door-unlocking system which provides essential
security to our homes, safes, and everything that needs
security. Arduino controls the camera 1)Arduino: Arduino is a platform based on a single board.
Arduino is used when there are interactions between inputs
for capturing photos for turning on a servo motor for door
and outputs. It is used to control the output with respect to the
unlocking. This module contains a secured face recognizer for
input. The Arduino programming language uses Wiring which
automatic door unlocking. The camera captures the face and
is an integrated development environment (IDE) and a
compares it with the images stored in the database. If the face
microcontroller.
is found in the database, then the door lock will be opened
otherwise it will be locked. Variations in lighting, pose, facial
expression, lighting, etc., are some main factors that influence
face detection and recognition. The face is detected using the 2)Servo Motor: Micro Servo Motor SG90 is a lightweight
Haar-Cascades feature and face recognition is implemented server motor with high output power. The servo motor can
using LBPH. Nowadays door locks can be accessed using rotate 180 degrees (90 in each direction) and works just like
systems with fingerprint sensors, passwords, face recognition, the standard kinds but smaller. You can use any coding
etc. The system comprises a camera to detect the faces and a platforms, hardware, or library to control these servo motors.
door lock for unlocking the door. Every face detected by the It is easy to use as it does not need any motor driver to control.
camera will be verified for similarity with the images stored in
the database of the system. If the face is matched, the door
will open automatically. In this module, we use Arduino as the Proposed Algorithm for face detection:
main control circuit. OpenCV Library & Python are used for
coding the algorithm. The face detection algorithm will detect and highlight the face
from the overall image. Then, it does the necessary cropping
and conversion from color to grayscale. The eigenfaces vector
will be extracted from this process. After that, the
Literature Survey:
classification algorithm will do an analysis to compare the
The goal of our project is to increase security measures and to input with the database, in which it will detect whether the
go one step above conventional door locking. It will not just input-face image is similar to the face image in the database. If
increase the security feature it will also free the User from it is recognized, the system will turn on the servo motor to
worrying about forgetting the password or the key to the door. unlock the door.
A set of images will be stored in the system which will act as

Electronic copy available at: https://ssrn.com/abstract=4236516


The Algorithm works like this:

Step 1- Start

Step 2-The image will be detected and the segmentation takes


place Includes Align then Crop

Step 3-Convert the image into Grayscale

Step 4 - Pre-processing on the Gray scaled image Significant


features Extracted

Step 5- If “User” Classification, Else, Denied

Step 6- Algorithm works on image

Step 7-Identified User.

Step 8-Stop

Conclusion:

In this project, a Secured Door Lock System based on Face


Recognition using Arduino is implemented. We designed the
system which provides secure locks for doors, comfort, and
connivance security for users. In this system, we have
implemented a combination of webcam, Arduino, relay, and
door lock. And we used the Local Binary Pattern Histogram
(LBPH) for recognizing the face. The arrangement of a facial
recognition system using Arduino can make the system lighter
and work successfully using lower control, so it is more
convenient than the pc-based face recognition system.

Experimental Result:
References:
Here in the image below is the captured Trained image of the
User that was stored in the system as the key for the door, the [1] (2017) Comparative Analysis for a Real-Time Face
particular User can now get the access through the door. Recognition System Using

Raspberry Pi Muhammad K Shaikh, Syed Annas Mazhar.

[2] (2017) Secured Room Access Module SuchitShavi.

[3] (2017) Automatic Semantic Face Recognition Mark S.


Nixon University of

Southampton, Southampton, United Kingdom

[4] (2017) Real-Time Implementation of face recognition


system by Neel

Ramakant Borkar and Sonia Kuwelkar, India

[5] (2017) IoT-based home security through Digital Image


Process Algorithms by
When the User wants access to the door again, the System will
detect the image and will recognize its authenticity of it from Beatrice, Dr. S. Britto Ramesh K and J. Jerlin Sharmila from
the database it has and it will provide access. Here around 30 India
images are captured once. The matching accuracy is
configurable depending on the quality of the camera we are
using.

Electronic copy available at: https://ssrn.com/abstract=4236516

You might also like