Analysis of Point Cloud Deformation Monitoring for Sandy Coast Based on PointNet++
Due to its special physical structure,sandy coasts are very susceptible to deformation by seawater erosion under the influence of climate change.In order to obtain the deformation informa-tion of sandy coasts,the M3C2 algorithm is usually used to calculate the deformation of the study ar-ea,but this method only considers the neighbourhood relationship between point clouds and lacks the global characterization of point clouds.Therefore,in this experiment,we used the one-kilome-tre coastal point cloud data monitored in Kijkduin area of the Netherlands for a period of seven months as the study object,and extracted local and global features of point clouds at different scales by PointNet++ deep learning algorithm,and constructed the distance metric of point cloud feature vectors to calculate the local and global features of point clouds within the study area.The local and global features of the point cloud were extracted by the PointNet++ deep learning algorithm,and the distance metric of the point cloud feature vectors was constructed to calculate the morphology of the point cloud in the study area.The experimental results show that the shape variables in the western and central parts of the study area are significant,and the maximum monthly average shape variable is 0.305 1 m.Based on the variation of shape variables,it is analyzed that the shape variables of the sandy coast have a close relationship with the rainfall and temperature.This research methodolo-gy takes into full consideration of the local and global characteristics of point clouds,thus realizing the efficient deformation monitoring of point clouds on sandy coasts,which is of great significance for the protection of sandy coasts.
sandy coastmulti-temporal point cloudPointNet++deformation analysisinter-point cloud distance
国家科技期刊平台
NSTL
万方数据
维普
