Parallel Algorithm And CUDA Implementation Of Vector Data Calculation

Vector data is a data model that uses point,line,surface and their x and Y coordinates to represent the specific features in space,which has been widely concerned and applied in many industries and fields.Vector data is used in government departments and scientific research institutions.High precision vector data has the function of confidentiality.The high-precision geographic information carried by vector data has important value in geographic information system.Vector data with low precision can not meet the requirements of some industries for positioning accuracy,which limits the application of vector data.Spatial overlay analysis is one of the most important analysis functions of GIS.It takes a lot of time to query the spatial relationship of polygon layers.Therefore,combined with the actual vector data production project requirements,the vector data calculation of CUDA is designed and implemented.The main work includes the following contents.(1)By analyzing several normal methods of vector data calibration,the CUDA software architecture of CUDA parallel programming development platform,CUDA heterogeneous computing model and CUDA architecture application are systematically studied.The high-density position coordinate calculation process in the process of vector data correction is optimized,and the vector data correction is completed.The precision control of vector data correction is realized,which makes the technical route of vector data processing more perfect and improves the efficiency of vector data processing.(2)The vector data space correction method used in the system requires high precision of control points.The nonlinear correction of vector data is realized by referring to the normal correction method of vector data space calibration.This paper analyzes the similarity matching technology of feature points of vector line features,systematically tests the accuracy control and efficiency optimization,and realizes the construction of system function module and system development environment.Combined with the actual project requirements analysis,tracking the system test results,comparing the system test results,finally the arbitrary control of vector data position accuracy is corrected,and the vector data production efficiency is improved.(3)The existing spatial query methods of vector data are analyzed.The single loop linked list data structure is used to store polygon related information.The polygon data set is divided into point and edge sets,and the spatial quad tree index of polygons is established.The spatial topological relationship judgment method of nine intersection model is applied to the spatial query between polygons,and the efficiency of polygon spatial relationship judgment is realized.