Noise reduction in ultrasonic computerized tomography by preprocessing for projection data

In this study, an ultrasonic computerized tomography (CT) using time-of-flights (TOFs) has been used for the nondestructive inspection of steel billets with high acoustic attenuation. One of the remaining problems of this method is noise in CT images, which makes it difficult to distinguish defects from noise. Conventionally, noise is suppressed by a low-pass filter (LPF) in the process of filtered back projection (FBP). However, it has been found that there is residual noise even after filtering. To cope with this problem, in this study, the noise observed in ultrasonic testing was examined. As a result, it was found that the TOF data used for CT processing contains impulse noise, which remains in the CT image even after filtering, owing to the existence of transducer directivity. To remove impulse noise selectively, we propose a noise reduction technique for ultrasonic CT for steel billet inspection, that is, preprocessing (outlier detection and removal) of TOF data. The performance of the proposed technique was evaluated experimentally. The obtained results suggest that the proposed technique can remove impulse noise selectively and markedly improve the quality of the CT image. Hence, the proposed technique can improve the performance of ultrasonic CT for steel billet inspection.