5.3.4 Unsicherheit bei der Datenfusion von Koordinatenmessungen
- 16. GMA/ITG-Fachtagung Sensoren und Messsysteme 2012
2012-05-22 - 2012-05-23
- 5.3 Messunsicherheit
- M. Galovska, M. Petz, R. Tutsch - Technische Universität Braunschweig
- 577 - 588
The complexity of the tasks in coordinate metrology is increasing and the requirements concerning the measurement time and uncertainty are becoming more stringent. This makes a measurement strategy with a single setup and a single measurement not always possible. Measurement object characterization often requires a multi-sensor and a multi-orientation measurement strategy that involves multiple view measurements as well as multiple probing. The process required to combine data from different sources defined as data fusion.
The goal of data fusion is data collecting or data improving to obtain new or more precise knowledge about physical quantities. Being used in a wide range of measurement processes, data fusion allows increasing effectiveness, enhancing scope and applicability, increasing accuracy and precision of the measurement. The task of data fusion also defines the configuration of the sources: competitive, complementary or cooperative.
There are three levels of data fusion (classification by output): raw-data level, feature level and decision level. Typical for coordinate metrology is the fusion of point clouds implemented on the low level (raw-data level) of data fusion. In fusion of coordinate measurements data the registration process is crucial. It is needed to transfer all measured points to a common coordinate system. The registration is realized either by the capabilities of the precision positioning or by matching of overlapping areas of the point clouds. The registration process is studied in this paper.