Heavy metal concentrations prediction of marine sediments by visible-near infrared spectroscopy based on attention mechanism

Xueying Li; Huimin Qiu; Aizhong Ding; Pingping Fan

doi:10.1016/j.jhazmat.2024.136729

Heavy metal concentrations prediction of marine sediments by visible-near infrared spectroscopy based on attention mechanism

J Hazard Mater. 2024 Nov 30:484:136729. doi: 10.1016/j.jhazmat.2024.136729. Online ahead of print.

Authors

Xueying Li¹, Huimin Qiu², Aizhong Ding³, Pingping Fan⁴

Affiliations

¹ School of Data Science, Qingdao University of Science and Technology, Qingdao 266061, China.
² Institute of Oceanographic Instrumentation, Qilu University of Technology (Shandong Academy of Sciences), Qingdao 266061, China.
³ Engineering Research Center of Groundwater Pollution Control and Remediation, Ministry of Education, Beijing Normal University, Beijing 100875, China.
⁴ Institute of Oceanographic Instrumentation, Qilu University of Technology (Shandong Academy of Sciences), Qingdao 266061, China. Electronic address: fanpp_sdioi@126.com.

PMID: 39631203
DOI: 10.1016/j.jhazmat.2024.136729

Abstract

Monitoring heavy metal concentrations in marine sediments is important for assessing marine environmental quality, protecting ecological health, and preventing human health risks. Visible-near infrared spectroscopy technology can overcome the shortcomings of traditional methods. Taking marine sediments from Jiaozhou Bay, Qingdao as an example, this paper collected visible-near infrared spectroscopy of the marine sediments and innovatively proposed a two-scale LSTM with attention mechanism method to predict heavy metal concentrations in marine sediments. By extracting two scale spectral features and combining the attention mechanism to fuse according to the difference in contribution degree of each feature, the method strengthens the effective feature extraction and improves the prediction accuracy of heavy metal concentration in Marine sediments. The results demonstrate that compared to multiple existing model methods, the prediction performance using the two-scale LSTM with attention mechanism method was improved. Specifically, for the sediment Cu concentrations model, R_c² = 0.820, RMSEC= 1.911, R_p² = 0.720, RMSEP= 2.849, RPD= 1.907; for the sediment As concentrations model, R_c² = 0.925, RMSEC= 0.893, R_p² = 0.838, RMSEP= 1.315, RPD= 2.492; and for the sediment Zn concentrations model, R_c² = 0.815, RMSEC= 4.741, R_p² = 0.702, RMSEP= 5.827, RPD= 1.850. This study provides a new tool for rapid analysis of Cu, As, and Zn in sediments of Jiaozhou Bay and other regions using visible-near infrared spectroscopy.

Keywords: Attention mechanism; Deep learning; Heavy metal; Sediment; Visible-near infrared spectroscopy.