Created by Sha, Kevin, last modified on Oct 26, 2020
NETIDNameClass
ksha3Kevin ShaECE 120
sharma62Siddharth SharmaECE 120




  1. Introduction

    1. Statement of Purpose

      We are creating a door locking system through facial recognition using raspberry pi. Our goal is to create a safe and efficient system for users to use when entering their housing arrangement. While we recognise that this is a software-heavy endeavour, due to the current circumstances, we feel like this is appropriate due to the complex nature of programming required. 

    2. Background Research

      We have researched a few different ways to go about using facial recognition, but the complexity and versatility of the raspberry pi make it the most efficient in order to complete our task. We have researched the actual circuit and how to use it and its OS, and how to go about generally implementing our work. We will be using OpenCV to write, train and implement the function of comparing the face from the camera to the database in the system. 

  2. Design Details

    1. Block Diagram / Flow Chart



    2. System Overview
      Motion Sensor: This will be connected to the lock to indicate whether there is any attempt to enter. This will start the process. 
      Camera: For detecting faces. Used as an input. 
      Raspberry Pi: The central hub. Connects the software to the actual setup. 
      IOT: Used as a database. Will be given an input and will match with entries in database. 
      Power Source: Needed if the solenoid lock will be used. The Raspberry can handle a little over 3, while the lock we plan to use can use anywhere from 7-12V. 

  3. Parts 
    1. 1xRaspberry pi 3 Model B https://www.okdo.com/p/raspberry-pi-3-model-b/
    2. 1xCamera https://www.adafruit.com/wishlists/502823
    3. 1xBreadboard (Given)
    4. 1xSolenoid lock https://www.amazon.com/FC-MXBB-Electronic-Electric-Assembly/dp/B077JXX7QV/ref=sr_1_2?dchild=1&keywords=solenoid+lock&qid=1603688735&sr=8-2
    5. 1xMotion Sensor https://www.amazon.com/Adafruit-LK-918O-SANV-FBACA-PIR-Motion-Sensor/dp/B00JOZTAC6/ref=sr_1_3?dchild=1&keywords=Raspberry+Pi+Motion+Sensor&qid=1603688774&sr=8-3
    6. 1xJumper Wires (Given)

      Challenges:  One of the biggest challenges is definitely calibrating and ensuring that the facial recognition works and is consistent. The coding and camera integration needs to be perfectly executed, otherwise, our project will fail eventually. Furthermore, the biggest challenge would be to eliminate the use of alternative surfaces such as pictures or other devices. This will require even greater detail in the software which will consume additional time. 

      References
    1. "How to Create a Facial Recognition Door Lock With Raspberry Pi | Raspberry Pi", Maker Pro, 2020. [Online]. Available: https://maker.pro/raspberry-pi/projects/how-to-create-a-facial-recognition-door-lock-with-raspberry-pi. [Accessed: 19- Sep- 2020]
    2. "Face Recognition Door Lock System using Raspberry Pi", Iotdesignpro.com, 2020. [Online]. Available: https://iotdesignpro.com/projects/face-recognition-door-lock-system-using-raspberry-pi. [Accessed: 19- Sep- 2020]
    3. M. », "Facial Recognition Door Lock", Instructables, 2020. [Online]. Available: https://www.instructables.com/id/Facial-Recognition-Door-Lock/. [Accessed: 19- Sep- 2020]
    4. D. Team, "Learn How To Use Raspberry Pi for DIY Facial Recognition", Device Plus, 2020. [Online]. Available: https://www.deviceplus.com/raspberry-pi/diy-facial-recognition-with-raspberry-pi/. [Accessed: 19- Sep- 2020]

Attachments:

BLOCK DIAGRAM.png (image/png)
BLOCK DIAGRAM (application/gliffy+json)
BLOCK DIAGRAM (application/gliffy+json)
BLOCK DIAGRAM.png (image/png)
Screenshot 2020-10-25 at 12.50.29 AM.png (image/png)

Comments:

Looks good. Project seems doable, and you seem to understand the challenges. Include a column along with your names and netIDs for which of ECE 110/120 each of you are in, and include pricing for your parts. 
Posted by siva3 at Sep 20, 2020 00:49

Your block diagram embedding seems broken on my end. It shows for a few moments and then disappears; would appreciate it if you could embed it somehow else that doesn't have this issue.

Project feasibility seems good. OpenCV on Raspi is a good, well tested setup. The solenoid may be tricky, those are often hard to power/drive correctly (they need a lot of current). Approved.
Posted by fns2 at Sep 30, 2020 16:48