Quality and speed improvements in industrial CT by the use of an additional optical sensor

A method for quality and speed improvements is introduced for a CT with exactly known model. The intention is the exploitation of the knowledge of the model for the correction process. We show that at least with an additional optical sensor we are able to register the model by a few projections only and during the data collection. Until now non-linearities in the CT as beam hardening were corrected by a process that needs a binary volume gained by a time-consuming post processing. With the known model pose we are able to perform the correction for the non-linearities already during the data collection. This means a great gain of time. Also the length derived the models will be more accurate than the lengths gained by first binary volume that still might possess some artefacts.