Surface Reconstruction Algorithm For Tree Trunk Based On Point Cloud

The construction of 3D surface models that accurately reflects the convexity and unevenness of tree trunk surfaces is a fundamental part of tree and forest scene visualization.Recent years,3D surface reconstruction of tree trunks has gained the attention of many researchers.The existing trunk surface reconstruction algorithms either have intersections or holes,or can not effectively reflect the real surface characteristics of tree trunks.In order to construct a non-intersecting,closed surface model that may effectively reflect the geometric features of the trunk surface,the subdivision-based trunk surface reconstruction algorithm and stem surface reconstruction algorithm based on maximization the opposite angle of the base edge are proposed.Subdivision-based trunk surface reconstruction algorithm.First,vertical segmentation is performed based on the maximum and minimum values of the trunk point cloud in the Z-axis and the thickness values of the segments.For each segment,the point cloud is divided into different angular partitions according to its center of mass and angular parameter values,and the point set of angular partition centroids is obtained.Secondly,the coarse-grained trunk surface model is obtained by using the trunk 3D surface reconstruction algorithm based on tangent plane projection with the point set of the center points of the angular partition as input.Finally,to obtain the fine-grained surface model,the Loop subdivision algorithm and the Sqrt3 subdivision algorithm are improved and applied to the coarse-grained surface model respectively.The experimental results show that the improved Loop subdivision algorithm and the improved Sqrt3 subdivision algorithm are closer to the real surface of the tree trunk than the Loop subdivision algorithm and the Sqrt3 subdivision algorithm when used in the reconstructed tree trunk surface model.Compared with the improved Sqrt3 subdivision algorithm,the improved Loop subdivision algorithm may also reflect the unevenness of the trunk surface well,in addition to being more realistic to the real surface of the trunk.Stem surface reconstruction algorithm based on maximization the opposite angle of the base edge.The set of points at the center of the angular partition is as input.First,a triangle normal vector angle threshold control method is put forward to optimize the determination method of searching the third point with the base edge;a triangle disjoint control approach is proposed to realize the construction of triangular mesh with flow structure;a closed point determination and removal strategy are come up with timely determine and remove each point that completes the construction of triangular mesh;an unclosed edge processing strategy is presented to solve the problem of unclosed edges generated by the triangle normal vector angle threshold control and triangle disjoint.The problem of unclosed edges generated by the control strategy is solved.Secondly,Through the above four strategies the algorithm of constructing Delaunay triangular mesh by maximizing the diagonal of base edges in two-dimensional space,a coarse-grained trunk surface model may be obtained.Finally,the improved Loop subdivision algorithm is applied to the coarse-grained surface model to obtain the fine-grained surface model.The experimental results show that the trunk surface reconstruction algorithm based on the diagonal maximization of the base edge can build a closed and disjoint trunk surface model,which may reflect the unevenness of the trunk surface.It can be used as the input of Tetgen mesh generator to build a tetrahedral solid model of the trunk.Compared with the improved Loop subdivision algorithm based on subdivision,the Hausdorff distance obtained by the improved Loop subdivision algorithm based on maximizing the diagonal of the base edge is smaller and closer to the real surface of the trunk.