基于环境监测大数据的污染源动态识别与追踪研究

打开文本图片集

关键词：污染源；动态识别；时空关联分析；机器学习

中图分类号：X83 文献标识码：A 文章编号：1008-9500（2025）12-0195-03

DOI： 10.3969/j.issn.1008-9500.2025.12.057

Abstract： Adynamic identification method for environmental polution sources integrating spatiotemporal statisticsand machinelearingisproposed，stablishingamultistagefeatureengineering modelcoupledwithspatiotemporaloelatio. Bysynergisticallyextracting spatiotemporal transmsion features ofpolutants through bidirectionallong sort-term memory networks and graph convolution，combined withroling horizon control forsource tracing optimization，dynamic decoupling of polutionsourcecontributionsand cross-regional trackingareachieved.Empirical results demonstrate that this method performs wellacrossvarious scenarios，meeting operational system requirementsand providing inteligent decision-making support for environmental regulation.

Keywords： pollution sources; dynamic identification; spatiotemporal correlation analysis; machine learning

工业污染源的动态识别与追踪是环境精准治理的关键环节，传统方法在应对多源异构数据融合与复杂时空动态演变方面存在显著局限。（剩余2972字）