Using YOLOv5 for Garbage Classification
Ziliang Wu, Duo Zhang, Yanhua Shao, Xiaoqiang Zhang, Xingping Zhang, Yupei Feng, Peng Cui
Abstract
At present, people's daily garbage is increasing day by day. How to intelligently classify garbage can save manpower and improve work efficiency. In this paper, a garbage classification model based on YOLOv5 object detection network named GC-YOLOv5 is designed. First, according to the common daily garbage category, five typical kinds of garbage were selected, data cleaned, labeled, and constructed a garbage dataset. Second, the GC-YOLOv5 was built and trained on our datasets. Third, in view of the convenience of multi-terminal access in the cloud and the reduction of computing pressure on edge devices, we deploy the garbage classification model in the cloud. The experimental results show that GC-YOLOv5 can accurately identify the garbage's types and find out the location of garbage.