Design of an urban solid waste classification system based on deep learning algorithm

This paper designs an urban solid waste classification system based on deep learning algorithms, which is applied to the classification of solid waste in urban areas in real life. Through technologies such as deep learning algorithms, machine vision, and computer vision, a Hikvision industrial camera is used to collect images, and the YOLOV8 algorithm network is employed to achieve the detection and classification of 4 major categories and 45 minor categories of common urban solid waste. Meanwhile, in order to improve the classification and positioning accuracy in the waste classification object detection task, an improved YOLOV8 network with an attention mechanism is introduced based on the benchmark YOLOV8 object detection algorithm, and comparative experiments are carried out on the waste classification dataset. This study provides certain practical research significance for urban solid waste classification and the environmental protection field.