One of the regular tasks, during a nuclear power plant operation, is to check the condition of the fuel assemblies and their internals. In some cases, the inspection has to be included in the periodic shutdown of a nuclear power plant and, due to the time limitation, is on the critical path. This several-days-long inspection is done by a special camera, which acquires video of the outside of the fuel assemblies. The experts trace these records online and, based on the overview, identify the places where a more detailed shot must be taken. The aim of this work is automation of the identification process of those problematic locations and, in this way, save many hours of human work (which is, due to a limited time frame, sensitive to a human error).
The work aims to design algorithms and implement a prototype of video processing and defect detection software. All preprocessing, classification and post-processing steps will be tested separately and for each step quality control will be done.
Seznam odborné literatury
Gonzalez, R. C. and Woods, R. E. [2018]. Digital Image Processing, 4th ed., Pearson/Prentice Hall, NY.