Security Management System | SIH 2022

• Developed a smart surveillance system implementing a face recognition model using RaspberryPi which continuously monitors the region in front of the house and alerts the user if an intruderis present along with the number of times he/she has appeared previously
• New person’s image can be added as known by sending a command and the system is retrained immediately

Idea/ Approach Details:

With burglary rate shooting up every day, deploying a strong and affordable security surveillance system at home becomes imperative. In this work, a Raspberry-pi based intruder alert system is designed to accomplish the above goal.

• This system works using the principle of face recognition and continuously monitoring the area in front of the main door of the home.
• When it identifies an intruder/burglar appearing in front of the door, it immediately sends their picture to the owner through mail, in addition to alerting him with an alert-notification.
• This device is intended to provide the best security to all at low cost. It is a design that doesn’t compromise efficiency for cost

Use Cases:

• The front door is under constant surveillance making it difficult for burglars to break-in
• Easily deployable in industries, restaurants, resorts and other places
• Updates the user about security status regularly
• Enables remote monitoring of the properties

Dependencies:

• Requires a constant power supply for operation of RPi
• High speed internet connectivity is recommended
• Relies on mail for alerting the user
• Constant running of the python code is essential

Uniqueness of the proposed system:

• Whenever an unknown person appears in front of the door, the owner is alerted through the eNotify app and the picture of the intruder is sent to him through mail
• In addition to sending an unknown person’s image, the number of times and the dates on which he/she previously visited are also sent to the owner for reference, and take further actions, if necessary
• If the owner decides to add a known person’s image to the ‘known’ list, it just requires him to send a command, specifying the person’s name, to the system. He need not send that person’s image, as it is already present in the device
• For every command sent by the owner, a reply or an acknowledgment mail is sent by the device, to make sure that he knows what is happening inside the system
• This system makes use of multiple cutting edge technologies like IoT, Motion sensors, Face Recognition and make it a robust security system completely not dependent on CCTV

Key Software Tools Used:

• OpenCV: It is used extensively to set the camera to capture the video and do image processing on the captured frames.
• Face Recognition: The face recognition modules are used to load the frames, encode, determine the location of faces in the frame and compare faces.
• SMTP: Simple Mail Transfer Protocol (SMTP) is used to send the mail to the owner whenever a new face is detected or when an acknowledgement needs to be sent or when the owner demands the known images database to be sent to him as a zip file.
• IMAP: Internet Message Access Protocol (IMAP) is used to access the mail server, read mail from the mailbox, add or remove images to/from the device for which the instructions are specified via mail.
• Pickling: Pickling is done here to dump the encoding and names of the images into a byte stream into the pickle file.
• File Handling: It is primarily used here to read from and write into files in binary format compatible with the pickle files.

Snapshots of OUTPUT:

(+) Live running of code

(+) System detecting a known person

(+) System detecting an unknown person

(+) Unknown Person Detected - Email

(+) REMOVE command

(+) REMOVE Acknowledgement -Email

(+) SHOW Command

(+) SHOW Acknowledgement -Email (With Database)