Improving and Extending the Information on Principal Component Analysis for Local Neighborhoods in 3d Point Clouds

Principal Component Analysis (PCA) is often utilised in point cloud processing as provides an efficient method to approximate local point properties through the examination of the local neighbourhoods. This process does sometimes suffer from the assumption that the neighbourhood contains only a single surface, when it may contain multiple discrete surface entities, as well as relating the properties from PCA to real world attributes. This paper will present two methods. The first is a correction method to filter out the presence of multiple surfaces through an iterative process. The second is to combine the PCA preformed on the neighbourhood of point coordinates and normal approximations in order to estimate the radius of curvature in the maximum and minimum curvature directions.