基于多模态信息的燃煤电厂排污口污染物泄漏检测研究  

Research on Pollutant Leakage Detection at Pollutant Outlets of Coal-fired Power Plants Based on Multimodal Information

作　　者：杨静 Yang Jing(Ecological and Environmental Monitoring Centor of Xiangyang,Department of Ecology and Environment of Hubei Province,Xiangyang 441021,China)

机构地区：[1]湖北省生态环境厅襄阳生态环境监测中心,湖北襄阳441021

出　　处：《环境科学与管理》2024年第7期122-126,共5页Environmental Science and Management

摘　　要：为解决现有燃煤电厂排污口污染物泄漏检测结果与实际不一致的问题,提出基于多模态信息的燃煤电厂排污口污染物泄漏检测方法。先搭建燃煤电厂排污口污染物泄漏检测框架,应用传感器实时监测排污口的污染数据,然后采集并抽取多模态信息,采用粗粒度算法融合处理多模态信息,获得多模态信息最终特征表示(浓度数值),最后将其与污染物质大气环境标准浓度进行比较,实现排污口污染物泄漏的精准检测。实验数据显示:应用提出方法获得的多模态信息噪声比例最小值为0.8,检测结果与实际保持一致,验证提出方法应用性能更佳。To solve the problem of inconsistency between the leakage detection results of pollutants at the discharge outlets of existing coal-fired power plants and the actual situation,a multi-modal information based pollution leakage detection method for coal-fired power plant discharge outlets is proposed.Firstly,build a framework for detecting pollutant leakage at the discharge outlet of coal-fired power plants,apply sensors to monitor real-time pollution data at the discharge outlet,collect and extract multimodal information,use coarse-grained algorithms to fuse and process multimodal information,obtain the final feature representation(concentration value)of multimodal information,and finally compare it with the atmospheric environmental standard concentration of pollutants to achieve accurate detection of pollutant leakage at the discharge outlet.The experimental data shows that the minimum multimodal information noise ratio obtained by the proposed method is 0.8,and the detection results are consistent with the actual situation,verifying that the proposed method has better application performance.

关 键 词：燃煤电厂 污染物 信息处理 排污口 多模态信息 

分 类 号：X830.7[环境科学与工程—环境工程]