The anisotropic properties of timber lead to variations in wave velocity depending on the chord angle, which significantly affects the accuracy of ultrasonic Computed Tomography (CT) imaging. To address this, it is proposed a velocity-compensated Linear Traveltime Interpolation-Simultaneous Iterative Reconstruction Technique (LTI-SIRT) imaging method that accounts for directional wave speed differences. The approach integrates ray-tracing-based forward and backward LTI computations with the SIRT iterative reconstruction algorithm.
A theoretical compensation model considering both chordwise and radial wave propagation was developed, and healthy-layer velocity data were obtained through both experiments and theoretical calculations. These were used to correct the travel times in defective layers. Validation experiments on timbere piles with small (4 cm) and large (13.5 cm) artificial voids demonstrated that both experimental (E-LTI-SIRT) and theoretical (T-LTI-SIRT) compensation methods significantly improved defect detection accuracy compared to the original LTI-SIRT approach.
For small defects, recognition accuracy improved from over 6\% error to below 3\%, while large-defect detection errors dropped by as much as 9\%. Positional deviations were also reduced, with minimum errors as low as 3.90 mm. The results show that the theoretical compensation method performs comparably to the experimental one, offering a practical alternative in scenarios where healthy-layer data cannot be directly measured. This improves the reliability and applicability of ultrasonic CT imaging for nondestructive assessment of historic timber structures.