Sparsing algorithm of massive dispersed two-dimensional data based on K-Neighboring

Abstract

With the continuous improvement of exploration equipment, measurement possibly contains more detailed massive data. Usually we need to mine the hidden rules and information of massive data through visualization techniques, but vast data will seriously affect the efficiency of model- reconstruction during the process of visualization. In this paper, the K-Neighboring Sparsing algorithm is proposed, whose core is the division of scattered two-dimensional data, and then establishes data points' K-Neighboring relations, and on the basis of this pump data via the criteria of either residual point counts appointed a forehand or the liminal distance of two points. The plane partition of massive two-dimensional data increases the speed of data sparsing. In practice, the proposed pumping dilute algorithm increases graphics rendering speed while ensuring accuracy of the original model and achieves the desired results.