Skip to content

Overview

Jump to bottom
XIANG edited this page Jun 1, 2022 · 1 revision

📖 Story Behind The Project

This is the introduction of the IoT project. 🎅

"My Little Skynet" is the final project of the IoT course at NCU, Taiwan. This project may still have many bugs or defects to fix and improve. So it is very welcome for you to issue me on my Github.

Why named this project "My Little Skynet" ?

In China, Skynet is a huge social monitor system. This system built with “Face Recognition”、“Big Data”、“IoT” and other technologies successfully forms the social security chain. This inspired me to name the project, so I finally name it “My Little Skynet”.

This project is a simple implementation of the real-world Skynet.

This device will inform the user by sending a Line message if any stranger or weird thing is entering the room or other places.

Due to a lack of time to fix the TensorFlow package installation problem, stranger detection will not be implemented in the IoT project this time.

The model was trained under the newest Tensorflow, but I can't download the newest version of Tensorflow in Rasberry pi 3 model B without enough resources to finish the download.

Each time I download it to 99%, the download process is killed by Raspbian OS.

💡 Possible User Cases

Original (Under Ideal Condition)

  • Assumptions:

    • The image recognition model can only identify two types of objects like family and stranger and does not identify other objects.
    • Do not provide the video live streaming, that is user can't watch the in-time video with the application interface. (User can only see the image captured by the camera during the rotation of servo motor.)
    • While using the device, the room or other places should have a weak light source at least. Do not use it without light, or it will be hard for the device to identify an image.

  • Situation: Check whether there is any stranger.

    • With Stranger

      1. Raspberry Pi will keep getting images from the connected camera.
      2. Raspberry Pi will use the image and recognition model to identify if there exists any stranger.
      3. After identifying by the Raspberry Pi there exists a stranger.
      4. Three things will happen:
        • The buzzer will keep ringing loudly.
        • The light‑emitting diode will keep shining with red light.
        • Raspberry Pi will use Line to send a hint message to inform the user that a stranger has entered the room.
      5. User can interact with Line Bot and press down the "taking photo" button to ask Raspberry Pi for taking pictures and return them.
      6. Raspberry Pi will use Line to send the image gotten from the connected camera every 5 seconds for 1 minute with 12 images in total.
      7. User can press down the "reset" button to reset the status of the device.

    • Without Stranger

      1. Raspberry Pi will keep getting images from the connected camera.
      2. Raspberry Pi will use the image and recognition model to identify if there exists any stranger.
      3. After identifying by the Raspberry Pi there does not exist any stranger.
      4. Only one thing will happen:
        • The light‑emitting diode will keep lighting with green light.
      5. User can interact with Line Bot and press down the "taking photo" button to ask Raspberry Pi for taking pictures and return them.
      6. Raspberry Pi will use Line to send the image gotten from the connected camera every 5 seconds for 1 minute with 12 images in total.

Implementation for Real

  • Assumptions:

    • Do not provide the video live streaming, that is user can't watch the in-time video with the application interface. (User can only see the image captured by the camera during the rotation of servo motor.)
    • While using the device, the room or other places should have a weak light source at least. Do not use it without light, or it will be hard for you to look the image.

  • Normal Situation:

  1. User can change three kinds of the degree to take a different image.
  2. Line Bot will return the image captured by the camera immediately.
  3. User can easily know what time the image was taken.
  4. User can reset the servo motor.

💭 Required Components

Name Quantity
Raspberry Pi 3 model B 1
32G SD card 1
Breadboard 1
5V Traffic Light LED Display Module 1
Dupont Line many
Raspberry Pi Camera Moudule V2 1
Raspberry Pi Camera Module V2 Case 1
MG996R 55g Metal Gear Torque Digital Servo Motor 1
Buzzer 1
Intel® Neural Compute Stick 2 1
Adhesive Tape 1
Carton 1
GPIO Expansion T Board for Raspberry Pi 1
Rainbow Cable 1
Adafruit PCA9685 16-Channel Servo Driver 1
AC Power Adaptor 1
Clone this wiki locally