基于超声技术的木质材料无损分类研究

Nondestructive classification of wood materials based on ultrasonic technology

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。

The method of wood material classification based on ultrasonic nondestructive testing can effectively overcome the defects of traditional classification method, such as the destruction of the test piece, the long detection period, the difference and the operation complexity. Selecting feature extraction method, combined with the optimization of the classifier model, the classification accuracy can achieve 97%,the single category accuracy of the classification is not less than 80%.