Application Research Of Clustering Analysis Algorithm In Surveying And Mapping Data Quality Inspection System

With the further development of geographic information industry, combined with the development of a new generation of mapping technology and equipment, the accumulation of GIS data is more and more abundant. In the face of massive GIS data, mining useful information from the massive amounts of data to improve the utilization efficiency of data is a key issue of geographic information industry. Besides, improving the efficiency of data quality inspection and management not only the requirements of the development of surveying and mapping industry, but also the guarantee of the GIS data used in relevant industry. This thesis studies the feasibility and effectiveness of applying clustering analysis algorithm to the quality inspection system based on the development of surveying and mapping data quality inspection system. The main research content falls in following aspects:1. Research on the sampling inspection plan based on surveying and mapping data quality inspection system. According to the analysis of the quality inspection system, sampling inspection method for surveying and mapping data has been studied. And this thesis elaborates the stratified sampling method advantages in the sampling inspection of the lot.2. Research on the stratified sampling scheme based on clustering analysis algorithm. Due to the lot selection problem in quality inspection system, combing with spatial correlation characteristics of GIS data, the stratified sampling scheme based on cluster analysis algorithm has been proposed. This thesis selects K-Means clustering algorithm to group the lot, and uses Matlab to compare the simple random sampling and cluster stratified sampling. The result shows that cluster stratified sampling is better than simple random sampling.3. Development of inspection point sampling program based on ArcEngine. Using ArcEngine, the point sampling program has been developed with c# language and cluster stratified sampling. And the program has been tested by instance data. The result shows that the program runs well and has achieved the expected experimental results.