Dustbin Waste Classifier

Project Overview

The Dustbin Waste Classifier is a computer vision-based solution designed to assist in waste management and recycling efforts. It can classify various types of waste into five categories and provides recommendations for the appropriate bins for disposal.

Classification Categories

• List of waste types that can be classified: 0. Plastic

  1. Glass
  2. Cardboard
  3. Biological
  4. Paper

• List of bins for waste collection:

  1. Recyclable
  2. Hazardous
  3. Residual
  4. Food Waste

Table of Contents

• Getting Started
  • Prerequisites
  • Dataset
• Usage
  • Training
  • Running the Classifier
  • Interpreting the Results

Getting Started

Prerequisites

Before using the Dustbin Waste Classifier, ensure you have the following prerequisites installed:

• Python
• OpenCV
• cvzone
• TensorFlow/Keras

Dataset

The images used to train the Dustbin Waste Classifier were obtained from Kaggle. The original dataset includes 12 classes with nearly 500 images in each class. However, for our specific project, we selected a subset of 5 classes to train the model.

Usage

Training

• We used Google Teachable Machine, an online tool, to create and train our custom machine learning model. Teachable Machine allows for easy uploading of images, labeling, and training of models.
• We provided labeled examples of waste images for each waste type to Teachable Machine. The platform automatically generated a model based on these examples, making it accessible and efficient for our project.
• After successful training, we exported the trained model in Keras format (keras_model.h5) for integration into our Dustbin Waste Classifier.

Running the Classifier

run the file main2.py Point your webcam at the waste item you want to classify.

Interpreting the Results

The classifier will provide the following results:

The type of waste (e.g., Plastic, Glass) detected in the image. A recommendation for the appropriate bin (e.g., Recyclable, Hazardous) for disposal.