CAJ | 학술논문

针对形态学运算在机载LiDAR滤波中最大窗口尺寸的选择问题，提出了一种基于形态学开重建的迭代多尺度点云滤波算法。通过循环迭代多尺度开重建运算，克服开重建对矮小地物的误判问题，自动获取开重建运算的最大窗口尺寸，解决了对城市区域滤波的窗口适宜尺寸选择问题，提高了地物与地面点分类的正确性。使用ISPRS提供的城区样本测试数据开展实验，结果表明：其Ⅰ类、Ⅱ类及总误差均值分别达到3.10%、6.05%和4.11%，在Ⅱ类误差不显著增加的情况下，Ⅰ类误差和总误差均值同比均为最小，整体分类与自动识别性能优于常规滤波算法。
침대형태학운산재궤재LiDAR려파중최대창구척촌적선택문제，제출료일충기우형태학개중건적질대다척도점운려파산법。통과순배질대다척도개중건운산，극복개중건대왜소지물적오판문제，자동획취개중건운산적최대창구척촌，해결료대성시구역려파적창구괄의척촌선택문제，제고료지물여지면점분류적정학성。사용ISPRS제공적성구양본측시수거개전실험，결과표명：기Ⅰ류、Ⅱ류급총오차균치분별체도3.10%、6.05%화4.11%，재Ⅱ류오차불현저증가적정황하，Ⅰ류오차화총오차균치동비균위최소，정체분류여자동식별성능우우상규려파산법。
Aimed at the maximum window size problem of LiDAR morphological method on unknown region, a morphological filter of iterative multi- scale opening by reconstruction (IMORF) was proposed on the basis of traditional morphological filtering algorithms. Multi-scale opening by reconstruction (MORF) was utilized to get maximum window size automatically, which can help user settle the suitable window size problem of unknown region. MORF was used iteratively to settle the classification error of the low objects that were nearby high and large objects. The experimental results for ISPRS urban data show that IMORF can classify terrain and off-terrain points effectively, and the mean of TypeⅠ, Type Ⅱ and total error are 3.10%, 6.05% and 4.11% respectively. Compared with other traditional filtering methods, the mean of Type Ⅰ Error and Total Error of IMORF are minimum with Type Ⅱ Error increased not obviously.