FYP_0_Phase2-14_11_23 deep learning approach for electronic waste detection , and classification

Final year project on deep learning approach for electronic waste detection , and classification.

1. Internal Project Guide: Dr. JIJESH J J HOD and Professor, Dept. of Electronics and communication SVCE, Bengaluru SRI VENKATESHWARA COLLEGE OF ENGINEERING BENGALURU PROJECT REVIEW-ZERO PID:2023P064 PRANAVA SWAROOP N - 1VE20CS115 PRANAV SASTRY – 1VE20CA015 KAUSTHUB RAO MOHITE – 1VE20CS061 ADITYA – 1VE20CA002 “A DEEP LEARNING APPROACH FOR ELECTRONIC WASTE DETECTION, CLASSIFICATION AND SORTING” PROJECT TITLE PRESENTED BY- Department of CSE and CSE-AI SRI VENKATESHWARA COLLEGE OF ENGINEERING Vidyanagar, Bengaluru- 562157

2. CONTENTS Introduction Problem Statement Significance of the Study Objectives of the Study References

3. introduction The rapid growth of technology, upgradation of technical innovations and a high rate of obsolescence in the electronics industry have led to one of the fastest growing waste streams in the world which consist of end of life electrical and electronic equipment products. These discarded electronic items, which include old computers, mobile phones, electronic accessories, and more, pose a dual challenge: environmental and economic. Improper disposal of e-waste not only contributes to environmental pollution but also results in the loss of valuable resources contained within these electronic devices.

4. This project seeks to address the critical need for a more efficient and precise approach to electronic waste management, By leveraging the power of deep learning and computer vision, it aims to develop a solution that can automatically detect, classify, and sort different types of electronic waste items. Deep learning approaches can be instrumental in addressing the challenges of e-waste detection, classification, and sorting. Deep learning techniques, which are a subset of machine learning, leverage artificial neural networks to automatically learn and extract complex patterns and features from data.

5. PROBLEM STATEMENT This system aims to improve the efficiency and accuracy of e-waste recycling by using advanced image recognition and machine learning techniques, ultimately reducing the environmental impact of improper e-waste disposal . Adapting to evolving e-waste products and materials over time. Meeting the requirements for real-time or high-throughput processing in recycling facilities and minimizing false positives and false negatives in the detection and classification process. Nonetheless, several significant challenges must be tackled for a sustainable e-waste management system, including: Inefficient Recycling Practices Lack of Recycling Accountability

6. SIGNIFICANCE OF STUDY The study's significance lies in its potential to reduce environmental harm from electronic waste (e-waste) by improving recycling, resource recovery, and regulatory compliance. It also enhances safety, supports a circular economy, and raises public awareness about responsible e-waste disposal. This study is significant for the following reasons: Scalability and Speed: E-waste management processes need to be scalable and efficient to handle large volumes of waste quickly. Developing automated systems that can process e-waste at a high rate is crucial. Resource Recovery: To make e-waste management sustainable, it is essential to recover valuable materials from electronic waste. However, this requires efficient sorting and processing to maximize resource recovery. Human Health and Safety: Workers involved in e-waste sorting and recycling face potential health risks from exposure to hazardous materials. Proper safety measures and training are necessary to protect the workforce.

7. Boo!

8. OBJECTIVES Object Detection and Recognition: Deep learning models, particularly convolutional neural networks (CNNs), can be used for object detection and recognition. They can identify and locate various electronic components within e-waste Image Classification: Deep learning models can classify different types of electronic devices, circuit boards, or e-waste components based on images. By training these models on labeled datasets, they can accurately identify and categorize e-waste items. Scalability and Speed: Deep learning models can process large volumes of e-waste data quickly, making them suitable for high-throughput e-waste management processes. Continuous Learning: Deep learning models can continuously adapt to changing e-waste materials and components, ensuring that the system remains effective as electronic technology evolves.

9. Internal Project Guide: Dr. JIJESH J J HOD and Professor, Dept. of Electronics and communication SVCE, Bengaluru SRI VENKATESHWARA COLLEGE OF ENGINEERING BENGALURU PID:2023P064 PRANAVA SWAROOP N - 1VE20CS115 PRANAV SASTRY – 1VE20CA015 KAUSTHUB RAO MOHITE – 1VE20CS061 ADITYA – 1VE20CA002 “A DEEP LEARNING APPROACH FOR ELECTRONIC WASTE DETECTION, CLASSIFICATION AND SORTING” PROJECT TITLE PRESENTED BY- Department of CSE and CSE-AI SRI VENKATESHWARA COLLEGE OF ENGINEERING Vidyanagar, Bengaluru- 562157

10. CONTENTS Literature Survey Frontend – ( Mobile Application)

11. Literature Survey Piotr Nowakowski , Teresa Pamuła (2020) Application of deep learning object classifier to improve e-waste collection planning 11 The proposed image recognition system for e-waste classification works by having individuals take photographs of the waste equipment intended for disposal and uploading the photo to a server running image recognition software that identifies the equipment. Identification and classification of waste electrical and electronic equipment from photos. To improve waste collection planning, photographs of the waste item is uploaded to the waste collection company's server, where it would be recognized and classified automatically. Deep learning convolutional neural network (CNN) was applied to classify the type of e-waste, and a faster region-based convolutional neural network (R-CNN) was used to detect the category and size of the waste equipment in the images. The recognition and classification accuracy of the selected e-waste categories ranged from 90 to 97%.

12. Literature Survey Piotr Nowakowski , Teresa Pamuła (2020) Application of deep learning object classifier to improve e-waste collection planning 12

13. Literature Survey Piotr Nowakowski , Teresa Pamuła (2020) Application of deep learning object classifier to improve e-waste collection planning 13

14. Literature Survey Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 14 Using deep learning and transfer learning techniques to classify waste into different categories (plastic, paper, metal, glass, cardboard, trash, e-waste) to support waste management and recycling. The dataset used contains over 2800 images across the waste categories. Various transfer learning models were compared including Xception , DenseNet121, ResNet-50, MobileNetV2, and EfficientNetB7. DenseNet121 achieved the highest accuracy of around 93.3%. MobileNetV2 and ResNet-50 also performed well with 93% and 92% accuracy respectively. All models except MobileNetV2 achieved above 90% accuracy, showing transfer learning is effective for waste classification. The study demonstrates that deep transfer learning approaches can accurately classify waste with over 90% accuracy to support waste management systems. DenseNet121 emerged as the top performing model.

15. Literature Survey Workflow Xception Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 15

16. Literature Survey Workflow Xception Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 16

17. Literature Survey Resnet-50 Architecture Xception Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 17

18. Literature Survey Resnet-50 Architecture DenseNEt121 Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 18

19. Literature Survey MobileNetV2 Architecture DenseNEt121 Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 19

20. Literature Survey MobileNetV2 Architecture DenseNEt121 Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 20

21. Literature Survey Accuracy DenseNEt121 Architecture Zian Md Afique Amin, Khan Nasik Sami, Raini Hassan (2021) An Approach of Classifying Waste Using Transfer Learning Method 21

22. Literature Survey AV Shreyas Madhav , Raghav Rajaraman, S Harini and Cinu C Kiliroor (2022) Application of artificial intelligence to enhance collection of E-waste: A potential solution for household WEEE collection and segregation in India 22 The paper proposes using a mobile robot with AI-based image recognition to identify and collect electronic waste (e-waste) from households during regular municipal trash collection in India. The CNN model uses transfer learning on a ResNet-50 architecture and achieves 96% accuracy in classifying e-waste items. The application uses deep learning, for object recognition in waste management. It mentions the use of the Trashnet dataset to analyze different neural network models and concludes that DenseNet121 with fine-tuning is the best approach for classifying recyclable materials. Pre-trained deep learning models via transfer learning have been utilized for general e-waste classification tasks. CNN model achieved 96% accuracy in classifying e-waste items.

23. Literature Survey AV Shreyas Madhav , Raghav Rajaraman, S Harini and Cinu C Kiliroor (2022) Application of artificial intelligence to enhance collection of E-waste: A potential solution for household WEEE collection and segregation in India 23 Proposed System

24. Literature Survey AV Shreyas Madhav , Raghav Rajaraman, S Harini and Cinu C Kiliroor (2022) Application of artificial intelligence to enhance collection of E-waste: A potential solution for household WEEE collection and segregation in India 24 VGG16 and Mod-ResNet-50 classification graphs.

25. Mobile Application

26. Overview and progress of the application 20XX Pitch Deck 26

27. SOURCES AND REFERENCES [1] E waste in India – RESEARCH UNIT (LARRDIS) RAJYA SABHA SECRETARIAT NEW DELHI, JUNE, 2011 [2] https://cleanriver.com/blog-what-is-e-waste-electronic-waste-disposal/ [3] https://www.geeksforgeeks.org/what-is-e-waste/

28. Thank you ✌️✌️✌️