Tile Storage And Management For Time Multiple-Versions Image

When the image data is updated locally or globally,a new version of the map image is usually formed.With the massive increase of remote sensing data,the update period of the earth observation data become shorter,not only the map image is frequently updated in the geographic information service,but also the organization and management of multiple time versions of image data has become more arduous.Through the research of multiple time versions image data,this thesis proposes an effective map tile storage method and spatial-temporal index optimization method,which further improves the performance of storage,reading and query response speed of the image tile data,and helps the quality of network geographic information service better meet the needs of large-scale users.The main works and innovations of this paper include:1.Map tile storage technology based on MBTiles(Mapbox Tile)MBTiles is a data storage specification.The image tile data has no uniform format and low I/O performance in organization and storage.The thesis proposes a parallel map tile storage method based on MBTiles.By dividing the parallel tasks,the disadvantage of the SQLite database,not supporting parallel data writing,are effectively avoided.The method can access map tile data more efficiently and shorten the storage time,improve the data reading speed,and make it more convenient to backup and migrate.2.Multi-version spatial-temporal indexing method based on CBB(Clipping Bounding Boxes)CBB is a method of clipping bounding boxes,which records the redundant space position in the minimum bounding boxes of spatial index by calculating the clipping points,so as to eliminate unnecessary sub-node calculation in the query process.Aiming at the problem of large amount of redundant space in the minimum bounding boxes in the index,the thesis proposes a multi-version spatial-temporal indexing method based on CBB,which expands the optimization method from two-dimensional to spatial-temporal three-dimensional.The method further enlarges the clipping volume in the minimum bounding boxes of the node in the spatial-temporal index,and further improves the query performance.3.Multi-version spatial-temporal query optimization method based on IB-CBB(Intersection Based-Clipping Bounding Boxes)IB-CBB method is proposed,aiming at the disadvantage of multi-version spatial-temporal indexing method based on CBB,which calculates all clipping points in the query phases.The subsequent calculation process is judged according to the intersection of the query box and the node bounding boxes,which not only avoids unnecessary sub-node calculations,but also saves the redundant clipping points comparison.The method optimizes the process of range query and further improves the query performance of the spatial-temporal index.4.An prototype system for multi-version image data managementWe implement a multi-version image data management prototype system,which can query,browse,download and compare multi-version image data by timeline to meet efficient multi-version image service and management requirements.