Scholarlink

Classification of municipal solid waste (MSW) at the source is essential in environmental protection and recycling. Most studies focus on the visual characteristics of objects in classification algorithms. However, many objects have similar visual characteristics but very different material properties, so distinguishing them using visual features is impossible. This paper proposes an advanced deep learning fusion model for object classification that adaptively combines visual features and collision sounds of objects. Experimental results demonstrate that the model achieves competitive classification performance, attaining an accuracy of 98.92%. In addition, the proposed model is also deployed in a prototyped smart bin and yields promising results.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Solid-Waste Classification Using Deep Learning Fusion Model

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Pham Doan Tinh,
  • Le Van Minh

摘要

Classification of municipal solid waste (MSW) at the source is essential in environmental protection and recycling. Most studies focus on the visual characteristics of objects in classification algorithms. However, many objects have similar visual characteristics but very different material properties, so distinguishing them using visual features is impossible. This paper proposes an advanced deep learning fusion model for object classification that adaptively combines visual features and collision sounds of objects. Experimental results demonstrate that the model achieves competitive classification performance, attaining an accuracy of 98.92%. In addition, the proposed model is also deployed in a prototyped smart bin and yields promising results.