Automatic extraction method for complex river cross-section terrain in multi-beam point clouds
A method for automatically extracting cross-section terrain from complex underwater terrain,and redundant multi-beam point cloud data,without a topological structure and uneven density distribution,has been proposed based on an improved Douglas-Peucker algorithm.The method eliminated noise points based on the spatial distribution of the point clouds in the river scene.Linear data structures were created by segmenting the point cloud into sections using pile points,which were then transformed into an independent coordinate system to facilitate the extraction of cross-sections and the organization of point clouds.The affinity between the cross-section curve and the fitting curve was analyzed to preserve the characteristic points and smooth points while removing redundant points,finally resulting in a refined river cross-section.The qualitative results showed that the method has a high degree of automation,retains the complete characteristic points,and effectively reduces redundancy.Quantitative evaluations showed that the average percentage of area difference was 0.02%,and the average percentage reduction in redundancy was 80.85%.The accuracy of the shape and the river cross-section area were not lost while reducing data redundancy.
multi-beam point cloudriver cross-sectionDP algorithmpercentage of difference areapercentage of degraded redundancy
国家科技期刊平台
NSTL
万方数据
