Implementation of Machine Learning using Fastai for Image Classification on the Automatic Waste Sorter Prototype
Muchamad Eris Rizqul Ulum, Joko Hadi Purnomo, Elfitrin Syahrul, Erfiana Wahyuningsih · International Journal of Computer Applications · 2022
Indonesia is one of the countries that contribute to waste in the world.Waste management is one way to reduce waste generated.Sensing the type of waste with a camera using computer vision is one method to take waste images.Using this method can create an automatic waste sorting system.The scattered waste is mixed of many types, including cardboard, glass, metal, paper, and plastic.This research uses the machine learning Convolutional Neural Network (CNN) Model to classify waste.The prototype uses a notebook to process classification and sends the classification to Arduino with Firmata protocol.The error rate value obtained from the train was 0.096708.while the accuracy value of the prototype is 84%. in this research, to be able to apply waste separation technology automatically to the waste, it can use fastbook when using a server computer and Arduino to control the device prototype.