Grid-Based Encoding QR-Tree K-NN Query

K neighbor search algorithm has a very wide application in complex surface of a large amount of data, aiming at the broad and irregular nature of mass data, this paper will use QR-tree index structure, in accordance with full octree encoding method to code each node, in accordance with the location relationship to establish grid, and nodes intersecting with horizontal and vertical lines are also involved in coding. When find the K-neighbor of a node in the QR-tree, it needs to find the corresponding grid units. Because the attached nature of the data object to the grid not only reflects the positioning in space, but also at the same time describes the relationship location between objects and other objects through the adjacency relationship between the grids, it can quickly find the nearest neighbor points, avoiding the complexity of previous traverse from the root node, raising the query efficiency, reducing the CPU running time and disk visits.