Point cloud analysis of 3D measurement data using a priori information
Start: October 2017
Student: Sofia Catalucci
Supervisors: Richard Leach, Petros Stavroulakis
Funding: Faculty
Point clouds extracted from the 3D data from surface measuring instruments are used for many purposes, including metrology, reverse engineering and quality inspection. A multitude of visualisation, animation, rendering and mass customisation applications can also be realised though point cloud processing. This project will concentrate on point cloud processing which would be useful for non-contact metrology applications, such as fringe projection, laser scanning and photogrammetry.
The aim of the research project is to investigate methods of pre-processing CAD files to improve the measurement process and assist the faster and more efficient acquisition of full 3D point clouds. More specifically, point clouds will be tagged with a priori data in regions of interest that have different measurement requirements (resolution, accuracy). The project will also focus on enabling rapid alignment and comparison between point clouds and CAD files of measured objects, in order to minimise 3D point cloud data acquired in in-line measurement applications, whilst also performing automatic rejection of extraneous points, thus also streamlining the 3D point cloud post processing required. Appropriate metrics for qualitative and quantitative post measurement comparisons of 3D point clouds and CAD data will be investigated. This will enable measurement systems to identify regions where the scan needs to be repeated or has not been successful.
Investigations and comparisons will be carried out using an industrial 3D metrology software solution, such as Polyworks, in combination with customised algorithms and control procedures for specific measurement applications developed in the lab during the course of the project.
ICP registered point clouds showing the data portion acquired from a fringe projection system in a single measurement cycle