CAJ | 학술논문

平均曲率是分析三维表面的重要几何特征之一。根据平均曲率进行海量散乱点云数据的精简，首先通过空间包围盒法建立K邻域，然后对K邻域内的点拟合二次曲面计算平均曲率，最后以邻近区域内点的平均曲率中误差为阈值，结合点的精简概率判定点是否保留。通过与传统方法对比，证实了文中方法在保留特征点和压缩上具有较好的优势。
평균곡솔시분석삼유표면적중요궤하특정지일。근거평균곡솔진행해량산란점운수거적정간，수선통과공간포위합법건립K린역，연후대K린역내적점의합이차곡면계산평균곡솔，최후이린근구역내점적평균곡솔중오차위역치，결합점적정간개솔판정점시부보류。통과여전통방법대비，증실료문중방법재보류특정점화압축상구유교호적우세。
The mean curvature is one of the important geometric features of 3D surface analysis .This paper reduces scattered point cloud data according to mean curvature .Firstly ,K‐neighborhood of points is established based on bounding box method .Secondly ,mean curvature is calculated based on quadric surface w hich is fitted by the points of K‐neighborhood .Finally ,mean square error of mean curvature of points in K‐neighborhood combined with reduction probability is used to decide whether to reserve these points . Compared with traditional method ,this method has some advantage in feature keeping and data reduction .