Effects of temperature and relative humidity on tomographic imaging of log knots based on the shortest current path model

Abstract Internal moisture and temperature of wood significantly affect its electrical resistance, destabilizing electrical resistance tomography (ERT) accuracy in log knot detection. To address this, this paper investigates the eccentric circle search-based inverse distance weighting (ECIDW) algorithm, based on the shortest current path model, under varying temperature and moisture gradients. ERT experiments on four Chinese fir ( Cunninghamia lanceolata (Lamb.) Hook.) logs (No. 1–No. 4) revealed a significant nonlinear coupling effect between environmental factors and imaging quality. Compared to extreme conditions, a specific window drastically improves accuracy. An optimal relative humidity of 60 % effectively suppresses cavitation and uneven moisture artifacts, maintaining area prediction errors for specimens No. 1, No. 2, and No. 4 within 25 %. Concurrently, an optimal temperature of 20 °C balances thermal activation responses; it sharply reduced the error of specimen No. 4 from 53.7 % at 10 °C to 6.3 %, and prevented feature collapse in No. 2 at 25 °C. Ultimately, operating within this optimized 20 °C and 60 % relative humidity window allows the ECIDW algorithm to significantly reduce background noise, improve the signal-to-noise ratio, and accurately predict knot position, area, and shape in Chinese fir logs.