Description:

👉 Download the dataset here

This dataset contains a collection of 15,150 images, categorized into 12 distinct classes of common household waste. The classes include paper, cardboard, biological waste, metal, plastic, green glass, brown glass, white glass, clothing, shoes, batteries, and general trash. Each category represents a different type of material, contributing to more effective recycling and waste management strategies. Garbage Classification Dataset.

Objective

The purpose of this dataset is to aid in the development of machine learning models designed to automatically classify household waste into its appropriate categories, thus promoting more efficient recycling processes. Proper waste sorting is crucial for maximizing the amount of material that can be recycled, and this dataset is aimed at enhancing automation in this area. The classification of garbage into a broader range of categories, as opposed to the limited classes found in most available datasets (2-6 classes), allows for a more precise recycling process and could significantly improve recycling rates.

Download Dataset

Dataset Composition and Collection Process

The dataset was primarily collected through web scraping, as simulating a real-world garbage collection scenario (such as placing a camera above a conveyor belt) was not feasible at the time of collection. The goal was to obtain images that closely resemble actual garbage. For example, images in the biological waste category include rotten fruits, vegetables, and food remnants. Similarly, categories such as glass and metal consist of images of bottles, cans, and containers typically found in household trash. While the images for some categories, like clothes or shoes, were harder to find specifically as garbage, they still represent the items that may end up in waste streams.

In an ideal setting, a conveyor system could be used to gather real-time data by capturing images of waste in a continuous flow. Such a setup would enhance the dataset by providing authentic waste images for all categories. However, until that setup is available, this dataset serves as a significant step toward automating garbage classification and improving recycling technologies.

Potential for Future Improvements

While this dataset provides a strong foundation for household waste classification, there is potential for further improvements. For example, real-time data collection using conveyor systems or garbage processing plants could provide higher accuracy and more contextual images. Additionally, future datasets could expand to include more specialized categories, such as electronic waste, hazardous materials, or specific types of plastic.

Conclusion

The Garbage Classification dataset offers a broad and diverse collection of household waste images, making it a valuable resource for researchers and developers working in environmental sustainability, machine learning, and recycling automation. By improving the accuracy of waste classification systems, we can contribute to a cleaner, more sustainable future.

This dataset is sourced from Kaggle.

This dataset contains images of 7 waste materials, and the goal is to classify them into different categories. Here’s a brief description of each class:

Cardboard: Images of cardboard materials, such as packaging boxes, cartons, and paperboard. Compost: Images of organic waste that can be composted, including food scraps, plant matter, and biodegradable materials. Glass: Images of glass containers, bottles, and other glass waste. Metal: Images of metallic waste, such as aluminum cans, steel containers, and other metal objects. Paper: Images of paper waste, including newspapers, magazines, office paper, and cardboard. Plastic: Images of plastic materials, such as plastic bottles, bags, and containers. Trash: Images of miscellaneous waste that doesn’t fit into the other categories.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F12145656%2F7096297ab054f505f4d788b86545ce5f%2F42979_2023_1706_Fig1_HTML.png?generation=1709547095589432&alt=media" alt=""> The dataset provides an opportunity to build a deep learning model that can automatically classify waste materials, contributing to better waste management and recycling efforts. You can explore the dataset, preprocess the images, and train a neural network to achieve accurate classification results.

Waste Classification (YOLO)

## Overview

Waste Classification (YOLO) is a dataset for classification tasks - it contains Waste annotations for 300 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

1) Data Introduction • The Garbage Classification Dataset is an image classification dataset consisting of 12 categories of household waste: paper, cardboard, biological waste, metal, plastic, green glass, brown glass, white glass, clothes, shoes, batteries, and general trash.

2) Data Utilization (1) Characteristics of the Garbage Classification Dataset: • Composed of 12 fine-grained waste categories, the dataset is well-suited for training high-precision classification models that reflect real-world waste sorting conditions.

(2) Applications of the Garbage Classification Dataset: • Development of automated household waste classification models: This dataset can be used to train deep learning-based image classifiers capable of sorting various types of waste, and can be applied to the development of smart separation systems that improve recycling efficiency.

Multi Label Classification Of Trash

## Overview

Multi Label Classification Of Trash is a dataset for classification tasks - it contains Materials Logos Objects Shapes annotations for 1,551 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

PROBLEM Waste management is a big problem in our country. Most of the wastes end up in landfills. This leads to many issues like: Increase in landfills, Eutrophication, Consumption of toxic waste by animals, Leachate, Increase in toxins, Land, water and air pollution.

APPROACH Studied white papers on waste management, Analysed the components of household waste, Segregated into two classes (Organic and recyclable), Automated the process by using IOT and machine learning, Reduce toxic waste ending in landfills

IMPLEMENTATION Dataset is divided into train data (85%) and test data (15%)

 Training data - 22564 images
 Test data - 2513 images

The Waste Classification Dataset is a curated collection of pre-labeled images organized into nine distinct categories: Cardboard, Food Organics, Glass, Metal, Miscellaneous Trash, Paper, Plastic, Textile Trash, and Vegetation. Each category contains images representing typical waste materials such as cardboard boxes, leftover food, glass bottles, metal cans, plastic containers, textiles, leaves, and more. The images were sourced and structured in folders named according to their class, enabling straightforward use in supervised machine learning applications.

This dataset is particularly useful for computer vision tasks aimed at sustainability, including training convolutional neural networks (CNNs) for automated waste classification, building smart recycling systems, and conducting research on environmental waste detection. Because the dataset is already organized and labeled, it is well-suited for transfer learning and rapid model prototyping. Overall, it serves as a practical resource for researchers, developers, and data scientists working on solutions in the domains of waste management and environmental conservation.

thomasavare/waste-classification dataset hosted on Hugging Face and contributed by the HF Datasets community

Trash Classify

## Overview

Trash Classify is a dataset for object detection tasks - it contains Recycle annotations for 3,000 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

This dataset is collected by Datacluster Labs. To download full dataset or to submit a request for your new data collection needs, please drop a mail to: sales@datacluster.ai This dataset is an extremely challenging set of over 9000+ original Trash/Garbage images captured and crowdsourced from over 2000+ urban and rural areas, where each image is manually reviewed and verified by computer vision professionals at ****DC Labs.

Dataset Features

Dataset size : 9000+ Captured by : Over 2000+ crowdsource contributors Resolution : 99.9% images HD and above (1920x1080 and above) Location : Captured across 500+ cities Diversity : Various lighting conditions like day, night, varied distances, different material view points etc. Device used : Captured using mobile phones in 2020-2022 Usage : Trash detection, Material classification, Garbage segregation, Trash segregation, etc.

Available Annotation formats COCO, YOLO, PASCAL-VOC, Tf-Record

*To download full datasets or to submit a request for your dataset needs, please drop a mail on sales@datacluster.ai . Visit www.datacluster.ai to know more.

The Garbage Image Dataset consists of images of garbage items collected from nearby localities using smartphones. The dataset is categorized into five different classes. Each category represents a specific type of garbage item commonly found in everyday waste. The purpose of the Garbage Image Dataset is to provide a collection of labelled images of garbage items from different categories. The dataset can be used to train and evaluate deep learning models for garbage classification tasks.

Explore the Garbage Classification Dataset with 15,150 images across 12 categories, designed to improve recycling processes.

The project explores the development of an AI model using computer vision for precise household waste classification, targeting plastics, cardboard, glass, cans, and paper. The literature review underscores the environmental concern in Costa Rica and examines AI techniques in solid waste management for enhanced efficiency and accuracy.The methodology employs a mixed approach, combining quantitative and applied elements. A carefully selected dataset and augmentation techniques are used to develop a robust AI model. Evaluation occurs in real household environments with active user participation. Results indicate good performance in classifying with positive user acceptance.

Smart Garbage Classification Solution Market Outlook

The global Smart Garbage Classification Solution market size was valued at approximately USD 1.2 billion in 2023, and it is projected to grow at a robust CAGR of 15.3% from 2024 to 2032, reaching an estimated USD 3.8 billion by 2032. The considerable growth in this market is primarily driven by increasing urbanization, the need for efficient waste management systems, and the adoption of smart technologies to enhance operational efficiencies.

One of the major growth factors for the Smart Garbage Classification Solution market is the rapid pace of urbanization and industrialization across the globe. As cities expand and industrial activities increase, the generation of waste also escalates, necessitating more efficient and effective waste management solutions. Governments and municipalities are increasingly turning to smart technologies to manage waste more systematically, thereby driving the market growth. For instance, the increasing amount of waste generated from urban centers has led to significant investments in smart waste management technologies, which is a key driver for this market.

Another critical growth factor is the advancement in technology, particularly in the fields of AI, IoT, and RFID, which are being integrated into waste management solutions. These technologies enable real-time monitoring, automated sorting, and efficient tracking of waste, significantly improving the efficiency and effectiveness of waste management processes. The incorporation of AI and machine learning algorithms to identify and classify waste accurately is a game-changer, helping reduce human error and operational costs. This technological advancement is expected to continue propelling the market forward.

Environmental regulations and sustainability goals are also playing a pivotal role in promoting the adoption of smart garbage classification solutions. Governments worldwide are implementing stringent regulations aimed at reducing landfill waste and promoting recycling and waste segregation at the source. These regulations are compelling businesses and municipalities to adopt advanced waste management solutions that are compliant with these guidelines, thereby fueling market growth. Additionally, the increasing awareness among the general public about environmental conservation and waste segregation is further accelerating the market's expansion.

Regionally, the market is witnessing significant growth in Asia Pacific, North America, and Europe, driven by technological advancements and increasing government initiatives. In Asia Pacific, countries like China and India are investing heavily in smart city projects, which include advanced waste management solutions. North America and Europe are also experiencing substantial growth due to the high adoption rate of advanced technologies and stringent environmental regulations. The Middle East & Africa and Latin America are gradually adopting these solutions, with growth primarily driven by urbanization and government initiatives.

Component Analysis

The component segment of the Smart Garbage Classification Solution market is categorized into Hardware, Software, and Services. Each of these components plays a crucial role in the overall functionality and efficiency of the waste management solutions. Hardware components include sensors, smart bins, and compactors, which are essential for the physical execution of waste classification and management tasks. The growing demand for smart bins equipped with sensors that can detect the type and amount of waste is significantly contributing to the hardware segment's growth. Additionally, advancements in sensor technology are enabling more precise and efficient waste sorting, further driving this segment.

Software, on the other hand, forms the backbone of the smart garbage classification solutions, providing the necessary analytics, data processing, and control mechanisms. Platforms that incorporate AI and machine learning algorithms to analyze waste patterns, optimize routes for waste collection, and generate actionable insights are becoming increasingly popular. The software segment is witnessing rapid growth due to the continuous advancements in AI and data analytics, which enhance the efficiency and effectiveness of waste management processes.

Services encompass the various support and maintenance activities required to ensure the smooth functioning of hardware and software components. This includes installation, regular maintenance, system upgrades, and customer support. As the adop

Image Dataset of Domestic Organic Waste and Non-Organic Contaminants for Classification and Segmentation

We developed a https://doi.org/10.1145/3626705.3631881" target="_blank" rel="noopener">Smart Trash Can that allows to unobtrusively take photos of real waste in the producer's home. Over six months, a volunteering family collected a total of 450 raw photos of domestic organic waste and non-organic contaminants, which were then manually labeled and segmented according to the captured waste types. Of the total of 450 images, 119 show pure organic waste while 324 also captured intentionally added contaminants. Another 7 photos show only the side walls or the plastic bag without any added waste, and are hence considered as background.

Context and Methodology

• Image dataset for the training of machine learning-models for computer vision aiming at waste quality management in the producer's home
• Collection of domestic organic waste and non-organic contaminants with a https://doi.org/10.1145/3626705.3631881" target="_blank" rel="noopener">Smart Trash Can
• Images manually labeled and segmented according to the waste types: organic, non-organic, and background
• The dataset was used in a publication demonstrating two machine-learning approaches for computer vision, a binary classification (mean accuracy of 90.35 %) and a impurity segmentation (mean accuracy of 98.24 %, mean intersection over union value of 96.43 %)

Technical Details

• A total of 450 photos: 119 pure organic, 324 with intentionally added contaminants, 7 background
• Domestic waste of a volunteering family (3 male, 2 female; 15-55 yrs; 33.2 ±16.2 yrs)
• Subjects gave written consent to provide the image data for research purposes and publication
• Raw photos provided in a lossless *.png format with compression level 0
• Two folders 'images' and 'masks' containing the *.png images and the *.png masks for semantic segmentation, respectively
• The images are classified and grouped in the three sub-folders 'background', 'organic', and 'non-organic'
• The semantic segmentation labels associated with the colors in the masks are provided in the 'labelmap.csv' file (';' as separator)

About the Dataset: 4 classes of drinking waste: Aluminium Cans, Glass bottles, PET (plastic) bottles and HDPE (plastic) Milk bottles. rawimgs - images of 4 classes of waste YOLO_imgs - images of 4 classes of waste with corresponding txt file (annotations for YOLO framework) labels.txt - labels of the classes

Story This dataset was manually labelled and collected as a part of final year Individual Project at University College London. Pictures were taken with 12 MP phone camera. I created a real-time waste detection and identification system powered by YOLO framework. Use it as you like, if you could cite me in your work, would be much appreciated. Please reach out to me if this dataset actually helped you with your project. Arkadiy Serezhkin - arkadiyhacks@gmail.com

Acknowledgements The dataset used parts of manually collected dataset of Gary Thung and Mindy Yang. I would like to thank them for collecting their dataset as this is not a fun thing to do (from my own experience). You can find their repository here.

Here are a few use cases for this project:

1. Waste Management: The model can be used to automate waste sorting in recycling plants. By correctly identifying and separating different types of trash, recycling process optimization can be achieved, leading to more efficient and sustainable waste management practices.

2. Smart Cities: Municipalities could leverage this model in smart bins installed throughout the city. The bins would be able to identify what type of trash is being thrown away and flag when they're full of a certain type of material, helping to streamline trash collection and recycling.

3. Environmental Research: Researchers could leverage the model to analyze waste composition in different regions or communities. This could provide insights into consumption patterns and inform policy initiatives focused on waste reduction.

4. Consumer Education Apps: This model could power a consumer application designed to educate individuals about the proper way to dispose of trash. The application could identify the type of trash and offer advice on how to recycle or discard it properly.

5. Corporate Sustainability: Businesses could use this model to better manage and reduce their waste, contributing to their corporate social responsibility goals. For example, the model could help categorize and quantify the waste produced in corporate cafeterias, allowing for adjustments in sustainability practices.

Garbage Images Dataset (6 classes)

I used this dataset for my CNN Python project, you can use it by the way 😀

I've collected data from these two datasets into one: - https://www.kaggle.com/datasets/asdasdasasdas/garbage-classification - https://www.kaggle.com/datasets/mostafaabla/garbage-classification

And balanced it by removing unnecessary images and making each class's size equal to 775 images.

Dataset Card for Dataset Name

  Dataset Summary

Dataset used to train a language model to do classification on 50 different waste classes.

  Languages

English

  Dataset Structure





  Data Instances

Phrase Class Index

"I have this apple phone charger to throw, where should I put it ?" PHONE CHARGER 26

"Should I recycle a disposable cup ?" Plastic Cup 32

"I have a milk brick" Tetrapack 45

  Data Fields

Phrase Class… See the full description on the dataset page: https://huggingface.co/datasets/thomasavare/waste-classification-v2.

EcologIANet_3Classes_401: Dataset of 3 classes of solid waste (Usable Waste, Non-Usable Waste, Organic Waste).

EcologIANet_3Classes_906: Dataset of 3 classes of solid waste (Usable Waste, Non-Usable Waste, Organic Waste)

EcologIANet_9Classes.zip: Dataset of 9 classes of solid waste (Cardboard, Electronic Waste, Glass, Hazardous Waste, Metal, Non-Recyclable Waste, Organic, Paper and Plastic).

EcologIANet_Crop_9Classes.zip: Dataset of 9 classes of solid waste (Cardboard, Electronic Waste, Glass, Hazardous Waste, Metal, Non-Recyclable Waste, Organic, Paper and Plastic); this dataset was cropped to emphasise the region of interest and simplify the train of a Convolutional Neural Network.