Research On Spatial Precision Optimization Method Of Population Grided Data Based On Land Use

The spatialization of population data can effectively improve the spatial resolution of population data,and has important practical significance for studying the law of population spatial distribution and improving the level of urban fine management.At present,great achievements have been made in the field of population data spatialization,and many representative model methods and population grid data have been formed.On the basis of summarizing the existing research results at home and abroad,this paper puts forward a set of methods to optimize the spatial accuracy of population grid data using land survey data,and carries out spatial accuracy optimization and experimental verification on the basis of two representative population grid data: China Unicom mobile signaling population grid data and World Pop data.This paper uses the man land contradiction detection and repair algorithm combined with statistical means to verify the spatial accuracy of the existing population grid data,and repair the inconsistency between population spatial distribution and land use distribution.The quantitative relationship between population spatial distribution and land use type is explored by using the empirical sampling method and the equal weight combination method based on the average number of people between discrete auxiliary data groups,and the grid resolution of data is improved by combining the zoning density method.Finally,through the simulation experiment of constructing grid scale,the effectiveness of the method in optimizing the spatial accuracy of population grid data is verified by means of statistical indicators and visual comparison.The main conclusions are summarized as follows:1)This paper proposes a method to optimize the spatial accuracy of population grid based on land survey data,which aims at the inconsistency between population spatial distribution and land use distribution and the problem of low grid resolution.The main steps of the method are: optimizing the previous distribution contradiction test and calculating the accuracy of the total number of people;Correction of contradiction between human and land distribution and optimization of grid resolution;After optimization,human land distribution contradiction test and visual comparison test.2)This paper optimizes the spatial accuracy of China Unicom population grid data by using high-precision land survey data.For the problem of inconsistent distribution of people and land,the number of people affected by the problem is reduced by 69% after correction.Secondly,according to the characteristics of China Unicom data,the calculation method of ground class weight is proposed,and the grid resolution is optimized combined with the partition density model.The accuracy is verified by constructing the grid scale simulation experiment.The results show that compared with the averaging method without land class weight,the prediction accuracy of the method described in this paper is higher.The pearson correlation coefficient(R)is 0.87,the average absolute error(MAE)is 74.851,and the root mean square error(RMSE)is 139.83,which are better than the averaging method.3)The method described in this paper is also effective for worldpop data,which has no contradiction between human and land distribution due to the high homogeneity of grid scale.The grid resolution is optimized to 25 meters by empirical sampling method and zoning density model.The results show that the spatial distribution law of population reflected by the optimized data is in line with the objective reality.The optimization method described in this paper has good universality for different population grid data.