Papadaki, H.; (2002) Automated measurement of complex engineering surfaces using multi station photogrammetry. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Automated measurement of complex engineering surfaces using multi station photogrammetry..pdf Download (29MB) | Preview

Abstract

Accurate three-dimensional models of objects are increasingly required in engineering applications. Accurate surface point measurement to targeted objects is already established with close range photogrammetry using retro-reflective targets, dense image networks and high-resolution digital cameras within the self-calibrating bundle adjustment. However, in cases of complex objects, targets provide inadequate information about the object surface and more dense information is required. The use of alternative measurement methods, such as coordinate measuring machines, laser scanning or structured light projection, is not always appropriate or within the cost specifications of the application. This thesis describes a new methodology that robustly and reliably produces dense surface measurements of complex objects within the established multi-photo close range photogrammetric framework. The method represents a new fully automated process that densifies a triangulation mesh originating from a sparse set of retro- reflective targets measured in an image network, to produce a dense 3D point cloud that is an accurate representation of the object surface. In order to achieve this the technique utilises image texture information, a variety of feature detection algorithms and a set of geometric constraints, which are derived from image network geometry and dynamic Delaunay triangulation. Since the presence of appropriate image texture information is fundamental to the technique, a rapid Fourier-based image analysis tool that can detect the presence of suitable image content is also integrated within the method. Finally, data quality of the 3D point cloud is investigated. The precision and reliability of the derived 3D data are ensured through the inclusion of a least squares bundle adjustment within the method. An accuracy evaluation is achieved based on the acquisition of surface measurement data from a mechanical probe-based coordinate measuring machine (Leitz CMM, 101012). Results verify the capabilities of the methodology to produce data of high precision and accuracy.

Download activity - last month

Export as CSVXMLJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item