Mining outliers from point cloud by data slice

Present a new outlier detection algorithm for point clouds based slice technology and the Local Distance-Based Outlier Factor, the algorithm changes the 3D data to 2D by data slice and projection, and employed KD-Tree to index the projected points at the same time, so a high processing efficiency can be reached and can deal with the huge quantity of point cloud in semi-automatic style. Several experiments are employed to prove its feasibility and practicability.