Research On Suitability Of Grid Scale Based On Boundary Calculation

At present,administrative units are used as the carrier of spatial data,but the causes of most geographical phenomena are not affected by administrative divisions.At the same time,administrative units can not reflect the internal distribution characteristics of spatial analysis,and it is not convenient to integrate multi-source data.However,grid of appropriate scale can make up for this problem.In order to efficiently and accurately display spatial data with grid cell as the carrier,it is a key problem to conduct statistical research on the scale of spatial data.Most of the studies use specific case areas,specific population data or land use data to explore the appropriate grid scale,with strong pertinence but low generality.Suitable choice for grid size,this paper presents a general higher based on boundary calculation method,defines the concept of bending,the grid within a certain scale,the linear distance between two endpoints curve and curve of the ratio of actual length,according to different scale grid vector boundary curvature in different principle,based on principles of area of various type and shape difference type selected multiple research area,statistics of various types in different scales in the study area and offline ratio of bending area error of each dimension mesh,draw the bending area of curve and error curve.In order to verify the influence of the original scale of the vector data,the bending ratio of the same research area of the vector map in different scales and different scales grids and the efficiency of attribute query were compared.The results show that,when the scale is 1:1 million,the bending ratio curves in each study area have the same general trend,showing a long-tail distribution.Each curve has feature points,but the scale of the feature points in different study areas is different.The larger the scale,the larger the scale of the feature points.The area error is linearly related to the scale and increases with the scale.When the subregions within the study area are taken into account,the bending ratio curve and the law of area error are consistent with that of the outermost boundary alone.Due to the different areas in the inner subregions,the corresponding scales of feature points are different.When the scale is 1:4 million or smaller,the relationship between the proportion of curvature and scale changes from long-tail distribution to linear distribution,and the feature points are not obvious,but the law of area error is the same as that of large scale vector data,and the efficiency curve of attribute query is also highly consistent with that of 1:1 million vector map.Based on the experimental results,the cultivated land data of Hebei Province and the historical place names data of Fujian Province were used for example verification,and the appropriate scale obtained by the experiment was consistent with the results calculated by the method in this paper.Finally,it is concluded that(1)there are feature points in the bending ratio curve of each study area.The larger the area is,the larger the scale corresponding to the feature points will be.The scale corresponding to the feature points is the appropriate scale of the region.(2)The area of the study area is the key factor affecting the suitable size of the grid,and the relationship between the perimeter and shape characteristics of the study area and the suitable size is not obvious.(3)There is a mathematical relationship between the appropriate scale of the grid and the area of the study area and the area error,so the appropriate scale can be calculated according to the formula by the area of the study area or the area error controlled in advance,and the area error at the appropriate scale in a certain study area can also be calculated.(4)If the units within the study area are considered,the minimum unit area in the study area can be selected as the scale selection basis to ensure the suitability of the minimum cell network.