Wavelet Denoising in Industrial Tomography

Abstract

In recent decades computed tomography (CT) has become a well established technique in non-destructive testing (NDT) approaches for accurate views of external and internal component structures. During the acquisition phase, many artifacts may adversely affect the quality of the sample edge estimation and interfere with surface detection. Their accuracies are important for metrology and correct volume visualization. In particular, the artifacts caused by the scatter radiation can be substantially strong, depending on the sample material and the geometry. In this paper we present a study on the use of wavelet denoising techniques to reduce the impact of such artifacts. The acquired images are filtered before the 3D reconstruction process. The reconstructed volumes are analised with the purpose of evaluating the influence of the denoising process, aiming at a better visual quality and more precise surfaces determination.