Autonomous Trash Collector Based on Object Detection Using Deep Neural Network

• Published in: TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON) pp. 1406 - 1410
• Main Authors: Hossain, Shamima, Debnath, Bidya, Anika, Adrita, Junaed-Al-Hossain, Md, Biswas, Sabyasachi, Shahnaz, Celia
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-10-2019
• Subjects: Autonomous; Cameras; DC motors; Deep Learning; Pins; Plastics; Raspberry Pi; Robot sensing systems; Servomotors; Trash detection; Ultrasonic Sensor
• ISSN: 2159-3450
• DOI: 10.1109/TENCON.2019.8929270

Non-biodegradable product usage and ignorance towards proper disposal are creating the problem of ever-growing trash stacks. An autonomous mobile trash collector which collects trash lying on ground in a trash-container attached to it can be a feasible solution to the problem. In this procedure trash detection is done via deep learning algorithm. An ultrasonic sonar sensor on the robot detects object along the path and a camera module sends pictures of the object to raspberry pi for classification into trash or not trash. The prototype robot is of low-cost and can detect a wide range of trash with high accuracy. Therefore has good environmental as well as economic impact.