IPSO-VMD based signal feature extraction and internal defect detection of hardwood logs through acoustic impact test

Abstract

Accurate quality detection of hardwood logs can realize efficient utilization of wood and provide significant benefits to wood enterprises. However, due to the differences in the extraction principle of acoustic parameters and the interaction mechanism between the parameters and wood properties, the quality assessment results of hardwood logs are different to some extent. For this, a method for feature extraction of the acoustic signal and defect detection was proposed based on the improved particle swarm optimization-variational modal decomposition (IPSO-VMD). By analyzing the sparse features of defect signals, the minimum mean envelope entropy was set as the fitness function of IPSO optimized VMD to search for the optimal number pair (L, α). And the search for IPSO was accelerated to acquire the global optimal solution by improving the inertia weight and learning factor of PSO. Then, the effective sub-modes decomposed from the IPSO-VMD were selected based on Hilbert marginal spectrum and energy ratio of sub-band components, and the frequency sub-band distribution and energy rates of the effective modes sifted were used as the characteristic parameters characterized the defect signal to realize the accurate quality detection of hardwood logs. The sawing results of the sample logs showed that the major defect types and priorities in logs were detected with an accuracy of 88.4% and 74.4% respectively based on the IPSO-VMD method, even for which not identifiable by global parameters could be also examined effectively. The effectiveness of the new feature parameters can provide a reliable basis for the accurate quality detection of hardwood logs through fusing the multi-parameter features and constructing the artificial intelligence recognition system in the future.