Implementasi Computer Vision Dalam Deteksi Dan Klasifikasi Sampah Otomatis Pada Sistem Pengolahan Limbah Perkotaan

Abstract

Waste is a very serious environmental problem commonly faced by Indonesians. According to data from the National Waste Management Information System (SIPSN), Indonesia's waste volume reached 20.02 million tons in 2022. In Indonesia, the amount of waste generated reached 65 million tons per day in 2016 and increased to 66.5 million tons in 2018. The amount of waste in Indonesia continues to increase annually. In large cities, waste management is an increasingly pressing challenge, given the negative impacts caused by improper management, such as waste accumulation in landfills (TPA), water and air pollution, and public health issues. This study aims to design and implement an automatic waste classification system based on Computer Vision technologies as a solution for urban waste management. The system utilizes an Arduino Mega 2560, camera, ultrasonic sensor, servo motor, and conveyor to detect and classify five main types of waste: plastic, paper, glass, metal, and organic materials in real time. The camera captures images of waste, which are then analyzed using a Computer Vision model, while sensors and actuators control the flow and physical sorting process. This research seeks to improve waste processing efficiency by reducing human involvement in hazardous tasks and to promote the application of intelligent technologies in supporting sustainable recycling systems and reducing the burden on final disposal sites (landfills). The system created can detect and classify waste types well.