Abstract: Point-cloud simplifications often adopt single clustering or iterative schemes.In this paper，combines the advantages of both.First，a processing of uniform clustering is performed for the point-cloud model，and then iterative simplification is used to further simplify the initially simplified model.In order to effectively combine the two schemes，uses quadric error matrix to transform related information between the two steps in the whole process.For the model with boundary，this paper presents a simple and effective method for detecting the boundary.Experimental results show that the quality of the simplified models using our method is superior to the uniform clustering method，and close to the iterative method.But the memory footprint and simplifying time are far below than the iterative method.

Key words: point-cloud simplification, clustering, iterative simplification, quadric error matrix

摘要： 目前点云简化方法很多采用单一的聚类或迭代简化策略。结合两者优点，首先对点云模型进行均匀聚类然后进行迭代简化。为了将两者有效结合，将二次误差矩阵应用在整个简化过程中，简洁地传递了两个过程中的相关信息。对于有边缘的点云模型，给出了一个简便有效的边缘检测方法。实验表明，该方法简化质量高于聚类简化而接近迭代简化，但内存占用和简化时间却远低于迭代简化方法。

关键词: 点云简化, 聚类, 迭代简化, 二次误差矩阵