Automated image stitching for enhanced visual inspections of nuclear power stations
Paul Murray, Graeme M. West, Stephen Marshall, S.D.J. McArthur · Strathprints: The University of Strathclyde institutional repository (University of Strathclyde) · 2013
In the UK, visual inspection of the fuel channels of the Advanced Gas-cooled Reactor (AGR) nuclear power stations forms an integral part of understanding the health of the reactor cores. During a statutory outage, video footage of the inside of selected fuel channels is recorded. Features of interest and anomalies are manually identified by an expert who extracts frames from the video to create a composite image for the feature of interest. This is a laborious and time consuming process which can be costly to station operators who must produce these images before returning the station to service. This paper describes an automatic technique capable of generating a 2D image of the entire internal bore of the channel. The technique uses the position of the camera coupled with advanced image processing techniques to generate a high-resolution image of the whole channel. This allows surface details to be viewed in relation to each other, and the rest of the channel, while facilitating a direct comparison of any anomalies over time. In addition, the time taken by this automated technique to produce a full core image is a fraction of that taken to manually stitch an image for a much smaller area.