You are on page 1of 1

EXPOLATORY PROJECT

Waste Segrigation Using M.L.


Department of Chemical Engineering and Technology , IIT (BHU) VARANASI

INTRODUCTION Fulcrum and Arduino Design :


Waste segregation is a crucial aspect of effective waste
management for environmental sustainability. In this The waste segregation process involves the use of a
project, we address the need for automated waste fulcrum designed to separate plastic and nonplastic
segregation, specifically focusing on classifying waste into waste.
plastic and non-plastic categories. Leveraging machine The provided Arduino code successfully operates the
learning, a model was trained to distinguish between these fulcrum, tilting it to separate waste based on the machine
two types of waste, offering a scalable and efficient learning model's predictions.
solution for smarter waste management practices.
Challenges Faced:
OBJECTIVE - Despite the successful functionality of the Arduino code,
challenges arose in constructing a sturdy fulcrum due to
The primary objective of this project is to develop an limited materials.
automated waste segregation system using a machine
learning model. The system aims to accurately categorize
waste into plastic and non-plastic types, streamlining the
waste management process. By implementing this
RESULT AND CONCLUSION
- The trained model demonstrated high accuracy and
solution, we strive to contribute to environmental
effectiveness in segregating waste into plastic and non-
sustainability, reduce manual sorting efforts, and enhance
plastic categories.
the efficiency of waste disposal.
METAL WASTE PAPER WASTE PLASTIC WASTE
- In-depth performance analysis revealed metrics such
as precision, recall, and accuracy, ensuring the model's
MATERIAL AND METHODS reliability.
- Real-world testing using a separate set of waste images
Dataset Preparation confirmed the model's practical applicability. - The
A diverse dataset was compiled, consisting of images of Arduino Uno code effectively controlled the fulcrum
mechanism, showcasing successful integration with the
waste items such as plastic waste, metallic waste and
waste segregation system.
paper waste.

Model Architecture:
TEAM MEMBERS
A convolutional neural network (CNN) model was designed
using TensorFlow and Keras. PATIL ABHISHEK VASANAT
- The model comprises multiple convolutional and pooling
ABHISHEK KUMAR SINGH
layers, followed by fully connected layers for
accurecy

ANIKET BHUI
classification.ste.
SAATVIK PANDEY
Training Process: AMIT KUMAR
- The dataset was split into training, validation, and test
sets.
Name Of The Supervisor : DR. R.S. SINGH
- Model training involved multiple epochs, with the
training process monitored using TensorBoard for number of epochs Signature With Date:
visualizing loss and accuracy trends.

You might also like