Smoking Detection System based on Image Processing with Deep Learning and Cloud Computing

Purpose of the Project

The goal to be achieved in this project is to identify smoking situations in places where smoking is prohibited by using image processing and deep learning approaches. Gas stations and bus stops are the main points where the smoking ban is violated.

General Project Architecture

Web Application architecture

React Application components

Running the Project

Object Detection Section

In a virtual machine, in the directory yolo-object-detection/yolo, start the process for detecting objects:

python3 yolo_vm.py

In a client environment, in the same directory, to start the client's video by opening a camera:

python3 client.py

Then, to upload the detected images to Google Cloud Storage:

python3 toBucket.py

Web Application Section

1. Install the latest node and yarn on your operating system.
2. Change directory: cd web_app.
3. Enter this command in command line:

yarn

4. Change directory to client folder:

cd client

5. Enter this command:

yarn

6. Change directory to project's root directory again:

cd ..

7. To start both the React and the Node.js parts:

yarn dev

8. To start only Node.js:

yarn server

9. To start only React:

yarn client

10. To view React page, open a browser and search localhost:3000.

11. To see Node.js API interactions, open a new tab and search localhost:5000.