In order to non-destructively identify the internal defect condition of the wood

disc specimen with holes

decays and knots is studied as a research object

and a 16-electrode sensor is used on the PXI platform with electrical signal acquisition and measurement functions. The energetic mode Electrical Resistance Tomography（ ERT） technique measures the internal data. For the 208 voltage data of the obtained wood faults

the principal component analysis（ PCA） is used to process the voltage data as input variables

and the genetic algorithm（ GA） is used to optimize the weights and thresholds

and the data for holes

decay

and knots are each 45 group

support vector machine（ SVM） algorithm training

to complete the internal defect image identification of wood. The results show that the GA-SVM algorithm has 94. 55 %

96. 36 % and 92. 73 % recognition rates for holes

decays and knots

respectively

which solves the problem of identifying internal defects in wood and achieves the best prediction results.