Efficient solid waste inspection through drone‐based aerial imagery and TinyML vision model

• Published in: Transactions on emerging telecommunications technologies Vol. 35; no. 4
• Main Authors: Malche, Timothy, Maheshwary, Priti, Tiwari, Pradeep Kumar, Alkhayyat, Ahmed Hussein, Bansal, Abhinav, Kumar, Raghvendra
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 01.04.2024
• ISSN: 2161-3915; 2161-3915
• DOI: 10.1002/ett.4878

Solid waste management is a significant challenge in the development of smart cities. Existing approaches for solid waste monitoring are often time‐consuming and resource intensive. Therefore, this study proposes a novel approach to solid waste monitoring that utilizes drone technology. The proposed method enables the efficient identification and classification of waste objects in the garbage discovered by the drone. This system can inspect every part of a smart city from a remote location, allowing for the timely and effective management of solid waste. Thus, the proposed system can be easily integrated in the existing waste management system for smart city. The drone‐based solid waste monitoring system comprises a drone equipped with a computer vision model for resource‐constrained devices and a software application that operates the drone and analyzes the captured image or video. The system utilizes the Internet of Things (IoT) to upload the collected data to the cloud, making it easily accessible whenever necessary. The proposed drone‐based solid waste monitoring system is a promising solution for the efficient and cost‐effective management of solid waste in smart cities. The system's innovative use of drone technology and IoT provides a scalable and adaptable solution that can be customized to meet the needs of any city. This research presents a drone‐based solid waste monitoring system designed to enhance smart city waste management. Using IoT and computer vision models, the system efficiently detects and classifies waste objects in real time, allowing for timely waste management. Quantized TinyML models enable the system to run on resource‐constrained devices, achieving high accuracy. The findings demonstrate its potential to revolutionize waste management in smart cities, offering a cost‐effective, scalable, and environmentally friendly solution.