Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms
Waste management is one of the biggest challenges for the recycling industry, especially because people always put garbage inside the recycling bin without checking its recyclability. The practice of tossing questionable items in the recycling bin and guessing if they are recyclable is known as wis...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/140019 |
| _version_ | 1811697143860166656 |
|---|---|
| author | Kua, Mei San |
| author2 | Chong Yong Kim |
| author_facet | Chong Yong Kim Kua, Mei San |
| author_sort | Kua, Mei San |
| collection | NTU |
| description | Waste management is one of the biggest challenges for the recycling industry, especially because people always put garbage inside the recycling bin without checking its recyclability. The practice of tossing questionable items in the recycling bin and guessing if they are recyclable is known as wish-cycling. (Wish-Cycling) Wish-cycling is a threat for waste management as recycling plants spend time and money to process the garbage, but ends up creating more waste. In this project, the author deployed a deep learning technique and applied it in Raspberry Pi 3 to classify garbage into five recycling categories, which are plastic, paper, glass, metal and non-recyclable item. This project aims to introduce a more effective way to process waste and stop wish-cycling. A dataset that contains 350 images for each recycling category is created to train a convolutional neural network (CNN) classification model. LeNet base network and various hyper-parameters are applied in the neural network to improve the accuracy of the model. By integrating the hardware applications, which are Raspberry Pi Picamera, DC motor, L298N Driver Module, SG90 Micro Servo and ultrasonic sensor, the detected garbage will be delivered to their respective dustbins. At the end of the project, a prototype of smart dustbin is constructed to capture the garbage image, classify and deliver it to its respective dustbin within 45 seconds. Also, Telegram - a communication application is employed to update waste collectors about the garbage level in dustbins and allow Materials Recovery Facility (MRF) administrators to retrieve the classification results from the Raspberry Pi for more meaningful analysis. |
| first_indexed | 2024-10-01T07:50:35Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/140019 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:50:35Z |
| publishDate | 2020 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1400192023-07-07T18:37:47Z Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms Kua, Mei San Chong Yong Kim School of Electrical and Electronic Engineering EYKCHONG@ntu.edu.sg Engineering::Electrical and electronic engineering Waste management is one of the biggest challenges for the recycling industry, especially because people always put garbage inside the recycling bin without checking its recyclability. The practice of tossing questionable items in the recycling bin and guessing if they are recyclable is known as wish-cycling. (Wish-Cycling) Wish-cycling is a threat for waste management as recycling plants spend time and money to process the garbage, but ends up creating more waste. In this project, the author deployed a deep learning technique and applied it in Raspberry Pi 3 to classify garbage into five recycling categories, which are plastic, paper, glass, metal and non-recyclable item. This project aims to introduce a more effective way to process waste and stop wish-cycling. A dataset that contains 350 images for each recycling category is created to train a convolutional neural network (CNN) classification model. LeNet base network and various hyper-parameters are applied in the neural network to improve the accuracy of the model. By integrating the hardware applications, which are Raspberry Pi Picamera, DC motor, L298N Driver Module, SG90 Micro Servo and ultrasonic sensor, the detected garbage will be delivered to their respective dustbins. At the end of the project, a prototype of smart dustbin is constructed to capture the garbage image, classify and deliver it to its respective dustbin within 45 seconds. Also, Telegram - a communication application is employed to update waste collectors about the garbage level in dustbins and allow Materials Recovery Facility (MRF) administrators to retrieve the classification results from the Raspberry Pi for more meaningful analysis. Bachelor of Engineering (Electrical and Electronic Engineering) 2020-05-26T04:35:11Z 2020-05-26T04:35:11Z 2020 Final Year Project (FYP) https://hdl.handle.net/10356/140019 en A3065-191 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Electrical and electronic engineering Kua, Mei San Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title | Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title_full | Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title_fullStr | Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title_full_unstemmed | Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title_short | Embedded system application development on Raspberry Pi 3 : smart dustbin : garbage classification system using deep learning algorithms |
| title_sort | embedded system application development on raspberry pi 3 smart dustbin garbage classification system using deep learning algorithms |
| topic | Engineering::Electrical and electronic engineering |
| url | https://hdl.handle.net/10356/140019 |
| work_keys_str_mv | AT kuameisan embeddedsystemapplicationdevelopmentonraspberrypi3smartdustbingarbageclassificationsystemusingdeeplearningalgorithms |