Computer Vision Based Waste segregation System

A computer vision approach to classifying garbage into recycling categories could be an efficient way to process waste. The objective of this project is to take images of a single piece of recycling or garbage and classify it into six classes consisting of glass, paper, metal, plastic, cardboard, an

Project Title

Computer Vision Based Waste segregation System

Project Area of Specialization

Artificial Intelligence

Project Summary

A computer vision approach to classifying garbage into recycling categories could be an efficient way to process waste. The objective of this project is to take images of a single piece of recycling or garbage and classify it into six classes consisting of glass, paper, metal, plastic, cardboard, and trash. In order to provide the most efficient approach, we experimented on well-known deep convolutional neural network architectures. For training without any pre-trained weights. Training and testing will be performed with image data consisting of several classes on different garbage types. The data set used during training and testing will be generated from original frames taken from garbage images. The data set used for deep learning structures has a total of 2527 images with 6 different classes. Half of these images in the data set were used for the training process and the remaining part was used for the testing procedure. Also, transfer learning was used to obtain shorter training and test procedures with and higher accuracy.

Project Objectives

The main purpose of our project is to classify wastes with high performance. In addition, our aim is to achieve this with high speed. Today, one of the main problems of countries is a waste management and waste classification. In many countries, people dispose of their waste without classifying it, so countries need to establish facilities to classify waste. It is important for the economy of the country that these wastes are classified with high performance. Because the more accurate the classification of waste, the more they can contribute to their economy. More importantly, they can leave a greener world for tomorrow. So why do we want to do it fast? According to World Bank data¹, we produce 1.3 billion tons of solid waste annually, which means 1.2 kg of waste per person per day. The World Bank estimates that by 2025 the amount of waste per person per day will be 1.42 kg. Based on these data, we can understand that we produce very fast waste. In order to achieve this speed, waste sorting facilities should also be fast.

Project Implementation Method

We will be using computer vision to tackle this issue. We will take live camera footage from some camera device. Our Model would classify the object present in the image.  We will only create the software part in our FYP that would generate output that can be used later to guide any robotic arm or some other hardware to segregate waste.

For creating the best possible model we will be utilizing the power of Deep Learning which is relatively new but is so much powerful that we cannot ignore it. We will create a model based on Masked RCNN which will not only classify the object but it will also tell us the position and shape of the object present in the image as output. Such output can be very helpful for controlling the hardware.

Benefits of the Project

The biggest advantages are stated below

1. Currently in Pakistan most of the segregation that is done is done by Human hands. This puts workers at risk of injury and disease. Furthermore, it is a very slow process. With our software guiding the hardware, the human cost can be reduced to the bare minimum, speed of the process can be increased by many folds, and risking important human lives can be avoided. 

2. There are very few waste segregation solutions used in Pakistan. The solutions that are being used are very slow and are not capable of handling the amount of waste that is being produced in our country. Because of which most of the waste ends up in landfills unsegregated which occupies thousands of acres of land, and causes insect flies flying, sewage overflowing, smelly, serious pollution of the environment. Pakistan is in dire need of a system like ours that would help to reduce the size of those landfills and avoid further polluting our land, water, and air.

3. Waste is a treasure if we can use it properly. We can save a lot of money in the production of many products that can be made from recycled products. We can also save money on energy generation by using Waste to Energy Plants.  Waste to Energy Plants burns municipal solid waste(MSW) to produce energy. But for all this, we need to separate waste into different categories and for that, we first need a proper, fast, low cost, efficient and effective waste segregation system. 

Technical Details of Final Deliverable

FYP 1 Includes:

1. Gathering Dataset

2. Preprocessing Dataset and Annotating Dataset

4. Finding the best possible Architecture for Model

5. Training the Model to best fit the Dataset 

6. Detect, Mask, and classify an object present in an image.

7. Calculating accuracy using different approaches of the machine and deep learning.

FYP 2 includes:

1. GUI application that utilizes the model.

2. Improve model Accuracy as much as possible.

3. Detection, Masks, and Classification on the live camera video feed.

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Good Health and Well-Being for People, Clean Water and Sanitation, Affordable and Clean Energy, Climate Action

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com