基于MLS点云模型特征的提取算法

针对以往特征提取算法提取边界特征效果不理想和特征识别时间较长的问题，提出了一种改进的点云模型特征边界线提取算法.该算法以移动最小二乘法为基础，在对已有的边界特征提取方法研究的基础上，结合研究对象棱角分明的特点，提取点云模型边界特征点集，利用双向搜索方法快速生成模型的特征边界线，并对得到的特征边界线进行特征修复，从而得到较为稳定的点云模型的特征边界线.实验结果表明，该算法能快速地提取点云模型的特征边界线，为后续基于特征边界的建模节约了大量时间.

Feature extraction algorithm for point cloud model based on MLS

In order to overcome the disadvantage that the extraction effect of boundary feature by the existing feature extraction algorithms is unsatisfactory and the feature recognition time is too long, an improved extraction algorithm of feature boundary line for the point cloud model was proposed. Based on the moving least squares （MLS） method and study on the existing boundary feature extraction algorithms as well as through combining with the angular characteristics of study object, the boundary feature point set for the point cloud model was extracted. The feature boundary line of the model was quickly generated with the bidirectional search method, and the feature repair for the obtained feature boundary line was performed. Therefore, a more stable feature boundary line for the point cloud model was attained. The experimental results show that the proposed algorithm can quickly extract the feature boundary line for the point cloud model, which can save a lot of time for the following modeling based on the feature boundary.