CleanCampus

Welcome to CleanCampus, an AI-Powered Solution to prevent littering.

Problem

• Littering is a very serious problem, it can lead to an increase in global warming.
• Countries like the US spend an estimated 11.5 Billion Dollars on their efforts to clean litter.

Solution

• We have trained a YOLOv8m model to detect a person and the litter they are throwing. If littering activity is detected, the face of the person is sent to a database, matched with an existing person's face, and then fined accordingly.
• ByteTrack algorithm is used to track the persons and the litter.

Litter Classes

• Paper
• Cardboard
• Can
• Plastic

Training Examples

• About 1500

Technologies

• YOLOv8m Object Detection Model
• ByteTrack Tracking Algorithm
• MySQL for the Database
• image_recognition of Python for the computation of the face embeddings
• Tkinter for the application
• Flask for connecting the model to the Application

Future

• The application paves and proposes promising ways to detect litterbugs; however, more real-world training datasets will be required to bring the model up to mark for real-world use cases.
• Making better UI and application for the admin, and adding additional features.
• Deploying the model on CCTV.

Additional Notes

• If you have an NVIDIA-GPU enabled machine, make sure to install the CUDA library to enable GPU inference.
• Make sure to have dlib installed for the image_recognition library.
• Make sure that you are in the same directory as the script that is going to be executed, otherwise you might get a path error, for example, if you are executing the code for detection and tracking, make sure that you are in the directory DetectionAndTracking, according to your terminal.