Survey on Waste Segmentation Using Image Processing

Abstract

This survey investigates Waste Segmentation using Image Processing, crucial for effective waste management amid global urbanization. It reviews diverse techniques, from traditional methods like thresholding to modern approaches involving machine learning and deep neural networks. The survey emphasizes image preprocessing for enhanced accuracy and identifies challenges like limited datasets. These include integrating multispectral imaging for enhanced discrimination and hybrid methodologies combining traditional techniques with deep learning. The survey\'s insights serve as a valuable resource for researchers, practitioners, and policymakers, guiding efforts towards sustainable waste management practices amidst rapid urbanization and population growth.

Introduction

I. INTRODUCTION

India is one of the countries in the world with the greatest rate of trash disposal. Every day, 377 million people in metropolitan India generate over 62 million tonnes of waste, 45 million of which is left untreated and disposed of in an unhygienic manner and posing serious health risks on the environment. Waste segregation, treatment, transportation, and disposal must be handled properly to reduce the risk to the health and safety of patients, the general public, and the environment.

Garbage segregation is the process of detecting, classifying, splitting, and sorting garbage and waste products in order to minimize, reuse, and recycle materials. Despite its substantial development, this approach has some drawbacks. The training picture dataset is one of this model's key drawbacks. To obtain its final prediction level for image classification, it requires a massive training image dataset. This project proposes a residential waste segregation system consisting of hardware and an image-processing-based software system. As a result, the project's purpose is to develop a waste segregation system. When waste is thrown in an open area, it causes pollution by decomposing and spreading odor, creating air and land pollution. When rubbish is deposited near bodies of water, it pollutes the water. In both developed and developing countries, waste is the leading cause of environmental contamination. Domestic garbage is classified into two types: biodegradable waste and non-biodegradable waste. Based on their reusability, these two basic groupings are further separated into two groups.

This research focuses on using the capability of image processing and seamlessly integrating it with a conveyor belt mechanism to revolutionize waste sorting. Using a webcam, the system recognises and categorizes several elements in real time, including metal, paper, plastics, and others.

II. LITERATURE REVIEW

1. This paper addresses India's significant waste management challenges by proposing a smart waste separation system using image processing and a Convolutional Neural Network (CNN)-based classification algorithm. The goal is to categorize waste into paper, food waste, plastics, and metals, addressing the inefficiencies in manual separation and contributing to effective waste management.
2. The "Automated Domestic Waste Segregator using Image processing" project proposes a real-time trash can system integrating Raspberry Pi hardware and machine learning-based image classification. It aims to efficiently segregate domestic waste into biodegradable and non-biodegradable categories, further classifying based on reusability, contributing to effective waste management and resource recovery.
3. This project addresses vital ecological concerns by proposing an automated waste segregation system using image processing. It classifies waste into recyclable and non-recyclable categories, enhancing resource recovery. The hardware includes a conveyor, camera, L-shaped clamp, and Arduino UNO, with a machine learning-based image classification algorithm trained on a municipal solid waste image database.
4. Efficient waste separation and recycling are crucial for reducing landfill waste. Despite potential cost savings and environmental benefits, people often neglect proper waste disposal. A significant portion of waste consists of recyclable materials. Manual sorting contributes to pollution. This survey aims to automate waste separation and implement a waste transport system, minimizing human involvement.
5. Poor solid waste management creates health and environmental problems in the Philippines. This work offers a Smart Garbage Bin Segregation based on Image Processing and Machine Learning, attaining 97.33% accuracy in classifying garbage into biodegradable, non-biodegradable, and unknown categories and providing a practical solution for effective waste segregation.

III. METHODOLOGY

Implementing waste segregation through image processing involves a systematic methodology. Begin by collecting a diverse dataset of waste images and annotating them to categorize waste into groups like paper, food waste, plastics, and metals. Preprocess the images by resizing and normalizing, enhancing quality using techniques like histogram equalization. Choose a suitable deep learning model, either pre-trained and fine-tuned or trained from scratch, for image classification. Split the dataset for training and validation, then monitor and adjust the model to prevent overfitting.

A. Survey of Algorithms

1. Smart waste segregation using image processing employs CNN algorithm with 90.10% accuracy. Requires ample labeled data, intensive computation, and is slower but yields automatic, highly accurate feature extraction.
2. Automated waste segregation system utilizes CNN with 89% accuracy. Requires abundant labeled data, high computation, and is slower but offers precise, automatic feature extraction.
3. Hand-crafted Features for floating plastic detection utilizes CNN with 90% accuracy. Requires ample labeled data, high computation, and is slower but offers precise automatic feature extraction.
4. Garbage classification algorithm based on YOLO achieves 86.9% accuracy. Struggles with close and small objects, requires large datasets. YOLO is fast, generalized, processing frames at 45-150fps.
5. Waste classification employs Random Forest (97.49%), Gaussian Naive Bayes (81.46%), SVM
6. (89.51%), MLP (96.44%). Challenges include computation complexity and tuning hyperparameters. SVM excels in high dimensions.
7. Solid domestic waste classification using image processing and machine learning uses Logistic Regression (78%), K-Nearest Neighbors (77%) and SVM (81%). Challenges include interpretation, high dimensions, and overlapping classes.

B. Block Diagram

The design intends to divide waste into two categories: biodegradable waste and non-biodegradable waste.

Biodegradable garbage is defined as follows:

• Paper waste
• Vegetable waste

Non-biodegradable garbage is defined as follows:

• Plastic waste
• Metallic waste

In order to operate the clamp and transfer the object to the appropriate garbage bin, a servo motor is further attached. A circuit schematic was used to operate the servo and dc motors.

V. FUTURE SCOPE

The future of "Waste Segmentation Using Image Processing" may include smarter machines with robotic arms. These robots could efficiently sort waste in real-time, contributing to improved waste management. Involves making the technology smarter, real-time monitoring, and using it with connected devices. Apps will help people get involved, and better machines will sort waste.

Conclusion

Garbage segregation is an important aspect of the garbage administration chain since it allows for effective reuse and recycling. Automatic waste segregation has received little attention and is practiced informally in many underdeveloped nations, owing to a lack of recognition, a lack of lucrative reasons, and a low first concern in design. The lack of SEGREGATION, collecting, and transportation of unsorted mixed garbage to landfills has an environmental impact. Pollution of the environment can be significantly minimized when wastes are properly segregated and processed. As a result, waste management through waste segregation can be said to play an essential part in environmental protection as well as human health and wellness.

References

[1] Dr.J.Suresh M.E, SMART WASTE SEGREGATION USING IMAGE PROCESSING IN CNN, 2023 IJCRT | Volume 11, Issue 9 September 2023 | ISSN: 2320-2882 file:///C:/Users/hp/Downloads/IJCRT2309288.pdf [2] J Sanjai, V Balaji, K k Pranav B. Aravindan, AUTOMATED DOMESTIC WASTE SEGREGATOR USING IMAGE PROCESSING, Volume: 06 Issue: 04 | Apr 2019 , IRJET , p-ISSN: 2395-072 https://www.irjet.net/archives/V6/i4/IRJET-V6I479.pd f [3] Haritha K N, Gopika S Pillai, Jyothi Krishnan M , AUTOMATED WASTE SEGREGATION SYSTEM USING IMAGE PROCESSING, 2023 IJNRD | Volume 8, Issue 6 June 2023 | ISSN: 2456-4184 | IJNRD.ORG file:///C:/Users/hp/Downloads/IJNRD2306489%20(1) .pdf [4] Prof. Yuvaraj ????, Likhith N Gowda , WASTE SEGREGATION USING IMAGE PROCESSING, JETIR, 2022 JETIR May 2022, Volume 9, Issue 5, ISSN-2349-5162 file:///C:/Users/hp/Downloads/JETIR2205081%20(1). Pdf [5] Froilan N. Jimeno, Briely Jay A. Briz, DEVELOPMENT OF SMART WASTE BIN SEGREGATION USING IMAGE PROCESSING, IEEE,16 March 2022, INSPEC: 21667429, https://ieeexplore.ieee.org/abstract/document/9732038 [6] D. Y. Saurav Kumar, “A NOVEL YOLOV3 ALGORITHM-BASED DEEP LEARNING APPROACH FOR WASTE SEGREGATION: TOWARDS SMART WASTE MANAGEMENT\", MDPI, 2020. https://www.mdpi.com/2079-9292/10/1/14 [7] P. W. Aishwarya Aishwarya, \"A WASTE MANAGEMENT TECHNIQUE TO DETECT AND SEPARATE NON-BIODEGRADABLE WASTE USING MACHINE LEARNING AND YOLO ALGORITHM\", IEEE, 2021. https://ieeexplore.ieee.org/abstract/document/9377163 [8] M. T. Daniel Otero Gomez, \"SOLID DOMESTIC WASTE CLASSIFICATION USING IMAGE PROCESSING AND MACHINE LEARNING\", Research Gate, p. 12, 2021. https://www.researchgate.net/publication/354555430_ Solid_Domestic_Waste_classification_using_Image_ Processing_and_Machine_Learning [9] J. S. I. Ibrahim F. Hanbal, \"CLASSIFYING WASTES USING RANDOM FORESTS, GAUSSIAN NAÏVE BAYES, SUPPORT VECTOR MACHINE AND MULTILAYER PERCEPTRON,\" IOP Conference, p. 6, 2019. https://iopscience.iop.org/article/10.1088/1757-899X/ 803/1/012017 [10] L. Liu and B. Zhou, \"YOLO-BASED MULTI-MODEL ENSEMBLE FOR PLASTIC WASTE DETECTION ALONG RAILWAY LINES\", IEEE, 2022 https://ieeexplore.ieee.org/document/9883308 [11] N. A. Z. M. Azizan, \"AN AUTOMATED SOLID WASTE DETECTION USING THE OPTIMIZED YOLO MODEL FOR RIVERINE MANAGEMENT\", Frontiers, vol. 10, 2022. https://www.frontiersin.org/articles/10.3389/fpubh.2022.907280/full [12] Z. Tian and D. Sun, \"GARBAGE CLASSIFICATION ALGORITHM BASED ON DEEP LEARNING\", IEEE, 2022. https://ieeexplore.ieee.org/document/9549336/citation s?tabFilter=papers#citations [13] G. S. P. Haritha K N, \"AUTOMATED WASTE SEGREGATION SYSTEM USING IMAGE PROCESSING\", IJNRD, vol. 8, no. 6 June, p. 7, 2023. file://C:\\Users\\krnaj\\Downloads\\IJNRD2306489.pdf

Copyright

Copyright © 2024 Prof. Dr. Shilpa Sondkar , Muhib Momin, Sumaiyya Nadaf , Kaushik Najare , Niraj Shevkari . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.