层次支持向量机在色木孔洞缺陷检测中的应用

针对现有木材无损检测中存在的问题,提出根据木材的声脉冲响应特点,通过自制的声波信号采集装置提取含有孔洞缺陷木材的声脉冲响应信号,再分别从时域和频域对信号进行处理,提取相关的统计信息作为识别特征,再输入到层次支持向量机（SVM）中进行识别的方法.结果表明,该方法对色木孔洞位置的识别准确率在95％以上,具有需构造的SVM分类器数量少、不存在不可识别域、训练和识别速度快的优点.对基于支持向量机的木材孔洞缺陷识别进行探讨,并对其有效性进行验证.

Application of Hierarchical Support Vector Machine in Diagnosing Hole Defects in Acer Mone

For the problems of the existing timber non-destructive testing, this paper proposes that according to the acoustic im- pulse response characteristic of the timber, the acoustic impulse response signal of timber hole defects is extracted using the homemade acoustic signal acquisition device, and then the signals are processed from the time and frequency domain, and the relevant statistical information is extracted as the identifying feature, and input to the level support vector machine （SVM）for training and recognition. The results show that the wood defect recognition accuracy reaches more than 95%. The method has the characteristics such as less SVM classifiers, no reject region, fast training and recognition. The effectiveness of the method is verified by the application to iden- tify wood hole-defects.