Flaw Shape Characterization in 3D Volume Visualization in Ultrasonic Testing

In general the ultrasonic testing is used for flaw detection and localization in tested materials. The flaws in materials are commonly characterized in measured ultrasonic signals by faults echoes. In the case of materials with coarse-grained structure the ultrasonic signal contains undesirable echoes caused by the scattering from these grains considered as noise. Another source of noise is caused by the electronic circuitry which is called electronic noise. These sources of noise make the flaw detection more difficult. The efficient methods for noise reduction have to be used. This paper evaluates different filtering methods used for noise reduction. Methods based on Wiener filtering, stationary wavelet transform, neural networks in 2D space are compared with conventional mathematical filtering methods based on median filter and mean value. The goal of this paper is to determine efficient method for noise reduction in B-scan for the exact flaw shape characterization. Filtered B-scans are used for the construction of 3D-volume visualization of flaws in the tested materials. Final 3D-volume visualization represents the flaw shape and is useful in the case of characterization of the flaw visualization