Citation

Mohd Fadir, Fatimah Balqis and Abdul Karim, Hezerul and AlDahoul, Nour (2024) TrashBot: Innovative Recycling By Utilizing Object Detection. In: 2024 IEEE 8th International Conference on Signal and Image Processing Applications (ICSIPA), 03-05 September 2024, Kuala Lumpur, Malaysia.

Text dd.pdf - Published Version Restricted to Repository staff only Download (660kB)

Abstract

With the ever-growing concern surrounding global warming, efficient waste management is a necessary practice to mitigate climate change. To address the problem, computer-vision-based systems are utilized with the integration of artificial intelligence (AI) and Internet of Things (Iot) technologies to enhance the waste segregation processes. To identify the waste, the object detection system is trained with a dataset consisting of 2401 images which are categorized into either paper, plastic or metal. This research employs the SSD MobileNet and EfficientDet D0 models. With various models trained, a comparative study is executed to examine each model's performance in classifying waste. These models are trained to recognize and classify waste into recycle categories - paper, plastic and metal, items failed to be classified are deemed as trash. The study aimed to incorporate hardware components to automate waste segregation, this is accomplished by designing an IoT hardware that incorporates the use of a servo motor. The servo motors allow the base of the bin to rotate depending on the classification of the object detected. This study was accomplished by loading the SSD MobileNet V2 onto the Raspberry Pi 4 Model B which achieved a precision of 0.714 and a recall of 0.809. With the SSD MobileNet V2 loaded, the python script allows the rotation of the bin. When the model detects metal the servo rotates 90 degrees, when paper is detected the servo motor rotates 180 degrees, when plastic is detected the servo motor rotates 270 degrees and when the item cannot be classified the servo motor will remain at 0 degrees.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Artificial Intelligence Raspberry Pi, Object detection, pre-trained models, Recycling
Subjects:	T Technology > T Technology (General)
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	04 Nov 2024 02:25
Last Modified:	04 Nov 2024 02:25
URII:	http://shdl.mmu.edu.my/id/eprint/13136

Downloads

Edit (login required)