Study On Data Refinement Method Of Soil Heavy Metals In Agricultural Land

The spatial distribution of sampling sites is a key factor for the detection,evaluation and mining analysis of sampling site.By selecting a case study in Shunyi district,Beijing,a data refinement method for the sampling sites of agricultural soil heavy metals was developed in this study.This method can take into account the uniformity of the geographical space of the sampling sites and the representativeness of the characteristic space to a certain extent,not only for big data.Redundancy provides a reference method,and can be used in the design of sample site layout,and has good applications in the soil pollution prevention and control action plan(soil ten articles),detailed investigation of soil pollution status,and point location optimization in other industries prospect.The main research results of this paper are as follows:(1)This paper proposes a method for detecting the uniformity of soil heavy metal sample data on agricultural land.This method constructs Thiessen polygons based on the soil sampling site data of agricultural land,and constructs the uniformity index through the area of Thiessen polygons to achieve uniformity detection.The results of the study show that there is one clustered sampling site and one sparse sampling site in the sample set,and the uniform variation index is 0.429.This method can provide a new idea for detecting the uniformity of spatially distributed sampling sites.(2)A method for de-redundancy and refinement of soil heavy metal sampling site on agricultural land is proposed.According to the test results of the uniformity of the sampling site,different sampling site data de-redundancy and refinement methods are adopted,which mainly include deleting the clustered sampling sites and encrypting the sparse sampling sites to achieve its spatial optimization layout.The research results show that the spatial deviation index of the original sample set is 0.327,and the spatial attribute interpolation error is 6.538;after de-redundancy and refinement,there are no clustered sampling sites and sparse sampling sites,the uniform variation index drops to 0.4,the feature space deviation index decreases slightly,and the spatial attributes The interpolation error drops to 6.357.This method can take into account the uniformity of the sampling sites in the geographical space and the representativeness of the feature space,and it is of great significance to improve the uniformity and representativeness of the spatial distribution of the sampling sites.(3)A method for weight adjustment of sampling site data refinement is proposed.The research results show that there is 1 sampling site whose weight needs to be reduced,and one sampling site whose weight needs to be increased according to the classification of sampling site data types.The spatial attribute interpolation error is 6.538;the weight of the sampling site needs to be reduced from 1 to 0.996 through weight calculation.The sampling sites that need to be weighted are adjusted from 1 to 1.183,and the spatial attribute interpolation error is reduced to 6.5.This method can be used for the optimization of long-term fixed monitoring stations to improve the representativeness and predictive ability of monitoring stations.It can also be applied to the optimized layout of non-point source and heavy metal pollution,site pollution monitoring,marine environment early warning,etc.They can improve the representativeness of sampling sites,reduce data redundancy,and have good application prospects.