Construction Of Big Data Platform For Heavy Metal Pollution Prevention And Control System In Karst Areas

With the continuous development of mining and metallurgy in the Karst area of Guizhou Province,the compound pollution of heavy metals has become more serious,which has a great impact on people’s living standards and the local ecological environment.How to conduct scientific and systematic management and restoration has become the first priority.Therefore,the big data platform of the heavy metal pollution prevention and control system in the karst area is of great significance.In addition,the update and replacement of map collection tools and the increase in data storage volume have led to the accumulation of massive amounts of geographic data by pollution monitoring departments at all levels.However,the use of traditional geographic information visualization software cannot realize the import,analysis and timely interaction of big data,and it will also waste human and material resources,which is not conducive to management and decision-making.In response to the above-mentioned problems,this article combines the powerful geographic visualization of Web GIS and Postgre SQL database storage functions to build an integrated geographic information big data platform for in-depth data analysis and mining.The main tasks are as follows:(1)Research different WebGIS visualization software and design front-end Web interface.Under the django framework,folium is used to display two-dimensional administrative division maps and topographic maps,cesium builds a three-dimensional space model and pyecharts draws statistical charts of heavy metal pollution.On the basis of fully understanding the user’s business needs,web pages for uploading,analysis,retrieval and management have been developed to help users monitor and predict changes in areas contaminated by heavy metals.(2)Construction and implementation of Hadoop big data platform.Under the Ubuntu system,three high-performance computers are used to build a distributed computing cluster.According to different geographic data,Postgre SQL is selected as the back-end database,and professional plug-ins are loaded to import vector data and raster data in batches.Under the django framework,the pandas and numpy functions in the python environment are used to clean and transform data,thereby converting different types of vector data into a standard json format.After extracting the different characteristics of the pollution data,we can export it locally or display it on the front-end Web.In addition to adding,deleting,modifying and checking postgresql databases,we also designed functions such as indexing,security authentication,logical replication,and monitoring.(3)Classification prediction of heavy metal pollution pictures,regression filling of some missing geographic data,and new ecological environment assessment methods.In the case of a large amount of data,an artificial neural network model is established to realize the classification of different heavy metal pollution environments.Secondly,when the missing data of some data sets is not serious,the random forest regression method is adopted to fill in the missing data in turn,and finally proposed a new method to further improve the risk assessment methods of heavy metal point-source and area-source source pollution.