Automated analysis of AGR fuel channel inspection videos

Remote Visual Inspection of the fuel channels which form the reactor cores of the UK’s fleet of Advanced Gas-Cooled Reactors (AGRs) occurs during planned periodic outages and provides station operators with a detailed understanding of core condition. A typical single fuel channel inspection generates a large amount of footage which must be analysed before the station is returned to power (provided it is safe to do so). While manual approaches are currently used, inspection videos can be analysed efficiently using techniques from image processing and computer vision. For example, the ASIST (Automated Software Image Stitching Tool) software processes inspection videos to construct a single image known as a chanorama (channel panorama) which allows the full inside surface of a single fuel channel to be viewed in a snapshot. To accurately characterise defects such as cracks in chanoramas, their dimensions need to be measured. This requires channel features of known size to serve as references to calculate scaling factors. These features vary from station to station and include overall brick dimensions, trepanned holes and keyways. In this paper, we propose a series of algorithms to automatically detect and measure the dimensions of each of these known features. In turn, this information can be used to generate a scaling factor which can be applied when sizing any cracks detected in a given chanorama.