An optimised resistance tomography approach for detecting knots in logs based on the shortest current path model

Existing electrical resistance tomography (ERT) methods for wood knot detection show significant limitations, such as lacking an effective model of the local impedance discontinuities in knot regions, and struggle to meet the requirements of high-precision and quantitative non-destructive testing and evaluation. In order to address the key limitations of traditional electrical resistivity imaging algorithms, this study was to establish a model based on the shortest current path, and then to optimise the knot image reconstruction method. Twelve log specimens totally were used to study the optimal current path and validate the optimised ECIDW algorithm. Results of tomographic images from three path models showed the shortest path model had higher accuracy and the lowest relative error of knot localisation and knot area estimation. Except for two specimens, the relative error of knot area from other specimens was within 25%. The optimised ECIDW algorithm proposed in this paper was capable of identifying the location and area of knots. Therefore, this work presented a more reliable image reconstruction method for non-destructive testing of wood, enabling high-precision detection and quantitative evaluation of internal knot defects.