At first, the data surface is partitioned into small blocks, and points in each block are reorganized into error-controlled LODs by hierarchical clustering and LOD organization.

• 首先对输入模型进行分块处理，获取的每一分块分别建立误差控制下的多分辨率数据结构。