General Approaches for 3D Point-Cloud Evaluation, Classification and Material Thickness

Abstract

Analysis of geometric point cloud measurements pose very difficult challenges if removed from prior knowledge regarding the measurement process or measured object. The goal is to investigate how collections of 3D coordinates produced from 3D scan measurement systems can be structured, evaluated and analyzed solely from the information available in the data. The key difficulty is to create frameworks and methods able to be generally applicable to any unstructured point cloud.

The problems and dependencies regarding handling of unstructured geometric point clouds are presented alongside methods to resolve them. A method for utilizing the commercially available software ATOS Professional used by the aerospace manufacturer GKN Aerospace to find material thickness of welding junctions is discussed. A general method for determining material thickness for 3D scanned data with low curvature is proposed and implemented through Python programming.

The methods are continuously tested throughout the implementation of the presented material thickness method using fabricated 2D and 3D geometrical point cloud shapes with random noise to mimic real measurement variance. Large grounds for further work has been uncovered where parts of the presented methods have room for improvement as well as the discovery of other solutions and analytical methods.