提出了一种基于模糊熵迭代的点云精简算法，在提高算法运行效率的同时，获得的精简点云模型具有更好的细节特征。对所有点云数据进行快速X-Y边界提取以保留点云边界特征；计算所有数据点的曲率，将除边界外的数据点按照曲率分组并计算每组数据点个数和曲率平均值；利用数据点的曲率构造点云模型的模糊集，计算最小模糊熵，从而得到最佳曲率划分阈值；对曲率小于阈值的数据点按迭代次数不同进行相应比例稀释，对曲率大于阈值的数据点在满足剩余点个数要求的条件下进行迭代计算模糊熵操作，不满足个数要求时数据点全数保留。实验证明该算法既能够保留点云的细节特征以逼近点云原型，又具有良好的运算效率。

A novel algorithm based on fuzzy entropy iteration is proposed to simplify point cloud data. Better detail features of the streamlined point cloud model are retained while the algorithm′s operating efficiency is improved. To retain the boundary characteristics, X-Y boundary of all point cloud data is extracted quickly. Curvature of each data point is calculated. Data points except boundary points are grouped according to the curvatures. Then the number of data points and the average of the curvatures in each group are computed. Fuzzy sets of point cloud model are constructed according to the curvatures. Then the minimum fuzzy entropy is calculated to obtain the curvature threshold, so that the data points can be divided best. The data points are diluted when their curvatures are less than the threshold. The dilution ratio depends on the iteration number. When satisfying the requirements of number, the data points are processed when their curvatures are bigger than the threshold. In this iteration process, new minimum fuzzy entropy and new curvature threshold are obtained. If not satisfying the requirements of number, the data points whose curvatures are bigger than the threshold are retained. The experimental results show that the proposed algorithm approximates the point cloud model and has a satisfactory computing efficiency.