Research And Application Of Improved Spatial Index On Mass Remote Sensing Data Storage Platform

Spatial index is a kind of data structure which is sorted according to certain rules according to the position and shape of the spatial object or the relationship between the spatial objects.Spatial data indexing technology is the key to improve the performance of spatial query.It studies reasonable and efficient space for the characteristics of spatial data Index structure has long been a research hotspot in the field of Geographic Information System(GIS).Currently,the popular spatial indexing methods mainly include grid indexing,KD-tree,quad tree,R-tree indexing and other algorithms.These algorithms all have some shortcomings when solving the spatial query.For example,the grid index cannot efficiently handle space in the case of uneven data distribution,the KD tree is suitable for point-like spatial data,while the index for other spatial data is inefficient.Quad tree is an unbalanced tree.When the spatial data is unevenly distributed,the depth of different sub-trees is greatly different,which affects the query efficiency.As the most popular and widely used spatial indexing algorithm,R tree also has the problem of overlapped intermediate nodes,invalid query paths and low node space utilization rate.This paper will focus on the R-tree and conduct some research on its existing problems Improve.In this paper,the concept of R-tree and the basic algorithm is introduced in detail.In order to improve the efficiency of R-tree construction processes.When the scheme inserts data into a leaf node that is already saturated,an overflow node is created for the node to store the data inserted into the node at this time and later.When the overflow node reaches saturation,the node and its overflow node Split into two saturated nodes.The simulation results show that the algorithm can reduce the number of splitting in R-tree construction,improve the space utilization of R-tree nodes and improve the efficiency of spatial domain query and k-nearest neighbor query.In this paper we combine spatial indexing technology with distributed system to design and implement a distributed spatial indexing module based on Hadoop.Based on the improved R-tree and MapReduce,a distributed spatial index is generated in parallel,Generated index Spatial data for parallel query,including the regional query and k adjacent queries.The module has the functions of generating distributed index,parallel region query and parallel K-nearest neighbor query function in parallel for spatial data set,which is of certain positive significance for improving index construction and query efficiency of mass spatial data.