Scholarlink

<p>The foundation of solid waste resource utilization and disposal lies in identifying the types and characteristics of solid waste. However, owing to the complexity and mixed storage of solid waste types, classification is difficult, which hinders the development of solid waste recycling and a circular economy. This study, which was based on laser-induced breakdown spectroscopy (LIBS), carried out rapid quantitative detection of 16 heavy metal elements. By using the Kruskal-Wallis H test method, heavy metal element fingerprint factors were selected, and characteristic distribution curves of heavy metal elements in different solid wastes were plotted to construct classification and identification models. The study revealed that the random forest model achieved the highest classification accuracy of 96.8%. Finally, LIBS was used to detect the elemental content in actual samples, and when combined with the model, the classification accuracy reached a maximum of 74.2%. This paper provides a method for identifying solid waste fingerprint features and a recognition model based on fingerprint features for the rapid identification of unknown solid waste, laying the foundation for waste property recognition and resource recovery.</p>

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

Fingerprint feature recognition method for solid waste based on LIBS

  • 期刊论文

  • 发表日期：

    2026年1月5日

查看全文
  • Ruixiao Huang,
  • Yongqi Lu,
  • Jianing Xiao,
  • Yufei Yang,
  • Zhimin Zheng,
  • Jinzhong Yang,
  • Qifei Huang

摘要

The foundation of solid waste resource utilization and disposal lies in identifying the types and characteristics of solid waste. However, owing to the complexity and mixed storage of solid waste types, classification is difficult, which hinders the development of solid waste recycling and a circular economy. This study, which was based on laser-induced breakdown spectroscopy (LIBS), carried out rapid quantitative detection of 16 heavy metal elements. By using the Kruskal-Wallis H test method, heavy metal element fingerprint factors were selected, and characteristic distribution curves of heavy metal elements in different solid wastes were plotted to construct classification and identification models. The study revealed that the random forest model achieved the highest classification accuracy of 96.8%. Finally, LIBS was used to detect the elemental content in actual samples, and when combined with the model, the classification accuracy reached a maximum of 74.2%. This paper provides a method for identifying solid waste fingerprint features and a recognition model based on fingerprint features for the rapid identification of unknown solid waste, laying the foundation for waste property recognition and resource recovery.