Optimisation of acquisition parameters for radiography using numerical simulation

Abstract

X-ray imaging is a powerful tool for the inspection of industrial castings. The radiograph quality is highly dependent of a large number of inspection parameters: generator settings, object geometrical parameters and detector characteristics. When trying to optimise these parameters, the large dimensionality of the problem makes manual investigation of the solution difficult. Within European project VERDICT, we propose to use simulation to predict the optimal imaging conditions, and thus to help the operator to tune the system prior to the inspection. We use a simulation software tool called 'Sindbad'. For every particular set of acquisition parameters, a corresponding radiograph is simulated. A main loop scans the whole parameters space to find the optimal parameters set, using a defect detectability criteria as cost function. Detector response is precisely modelled, both for film and flat panel. Different versions of the detectability criteria are discussed, depending on whether detection is performed by numerical processing or human inspection. Experimental validation of our approach is performed using components from aeronautics.