基于超声无损检测的木质材料分类方法,并通过选择合适的特征选择或特征提取方法,结合优化的分类器模型,可以实现高达97%的分类准确率,且单个类别的分类准确率不低于80%,实现木质材料无损分类。有效克服了传统木材分类方法破坏试件、检测周期长、适用性差和操作复杂等缺陷。 Based on the classification method of wood materials by ultrasonic non-destructive testing, the classification accuracy of up to 97% can be achieved by selecting the appropriate feature selection or feature extraction method combined with the optimized classifier model, and the classification accuracy of single category is not less than 80 %, To achieve non-destructive classification of wood materials. Effectively overcome the traditional timber classification method to destroy the test piece, long testing cycle, poor applicability and complicated operation and other defects.