一种面向海底陡坡的ALB点云滤波与恢复方法OA

Point clouds are a fundamental data source for reconstructing three-dimensional marine scenes.To address the uncertainties in processing Airborne LiDAR Bathymetry(ALB)point clouds over steep seabed slopes,this paper proposes a point cloud filtering algorithm that integrates adaptive morphological filtering with hyperbolic surface fitting.Firstiy,a Gaussian function is applied to fit the primary peak of the ALB point cloud elevation histogram,thereby determining the sea surface elevation.By optimizing the depth threshold,sea-surface points are separated while shallow seabed points are retained.Subsequently the point cloud is then divided into grids to construct local reference surfaces,and elevation residuals are analyzed to identify candidate noise points.Meanwhile,a K-D tree is constructed to calculate the number of neighboring points;if the number is less than the minimum threshold,the candidate noise label is preserved and water-body filtering is performed accordingly.Finally,a three-dimensional seabed topography model is accomplished using hyperbolic surface fitting,and residual analysis is used to recover valid steep-slope points that may have been erroneously removed.Experimental results demonstrate that the proposed algorithm achieves a total error rate of 1.64%and a Kappa coefficient of 0.946,outperforming existing methods.It not only accomplishes the effective extraction of ALB seabed topography but also effectively preserves the characteristic details of steep-slope areas.