论文部分内容阅读
机载激光扫描点云的三维数字图像表达模型将二维形态学运算推广至三维,给出基于三维数字图像的膨胀和腐蚀运算方法。针对点云三维数字图像,提出基于三维数学形态学和聚类分析的分割方法。将点云三维数字图像进行膨胀和聚类分析,依据聚类结果得到点云的分割结果。讨论了本方法两个参数与点云分辨率、地物间隔之间的关系。选用两套实例数据进行实验,并将第一套数据计算结果与MeanShift算法、渐进三角网加密算法进行比较,从分割评价因子、精度、计算效率等方面分析本文方法与其他两种方法的优劣,最后分析了本文方法的稳定性。
The three-dimensional digital image representation model of airborne laser scanning point cloud extends the two-dimensional morphological operation to three-dimensional and gives the expansion and erosion operation method based on the three-dimensional digital image. For point cloud three-dimensional digital images, a segmentation method based on three-dimensional mathematical morphology and cluster analysis is proposed. The point cloud 3D digital image expansion and clustering analysis, based on the clustering results point cloud segmentation results. The relationship between two parameters of this method and the resolution of point cloud and the distance between objects are discussed. Two sets of example data are selected for experiment. The first set of data is compared with MeanShift algorithm and incremental triangle network encryption algorithm. The advantages and disadvantages of this method and the other two methods are analyzed from the aspects of segmentation evaluation factor, accuracy and computational efficiency Finally, the stability of this method is analyzed.