Also known as point cloud reconstruction/point cloud meshing, use a series of meshes to approximate you and the point cloud, in graphics, usually using triangular or quadrangle meshes

  1. Read the point cloud

Read the original point cloud. As can be seen from the figure, the resolution of the point cloud is very high, which will cause a large amount of calculation; And there’s a lot of outliers and noise

  1. Sampling under point cloud

The original point cloud is sparse to reduce the computation

  1. We get rid of outliers

The triangulation algorithm is sensitive to outliers

  1. Point cloud is smooth

Due to the measurement noise of the sensor itself, the point cloud obtained will fluctuate somewhat. The point cloud is smoothed by calculating the normal line (similar to the image smoothing operation in image processing).

  1. Point cloud triangulation

Resources

  • Point cloud reconstruction/point cloud triangulation/meshing _Stones1025 blog -CSDN blog _ Point cloud triangulation
  • code
    • PCL Learning: Fast Triangulation of Disordered Point Clouds _zfjBIT column -CSDN blog
    • Point cloud triangulation reconstruction _weixin_42734533 blog -CSDN blog
    • Regrid Remesh