欢迎访问林业科学,今天是

林业科学 ›› 2014, Vol. 50 ›› Issue (8): 126-130.doi: 10.11707/j.1001-7488.20140818

• 论文与研究报告 • 上一篇    下一篇

原木内部孔洞缺陷二维超声图像构建

徐华东, 王立海, 宋世全   

  1. 东北林业大学工程技术学院 哈尔滨 150040
  • 收稿日期:2013-07-20 修回日期:2014-05-15 出版日期:2014-08-25 发布日期:2014-07-31
  • 基金资助:

    国家林业“948”项目(2014-4-78);国家自然科学基金青年项目(31300474);中国博士后科学基金资助项目(2014M551203);中央高校基本科研专项资金(DL12BB18)。

Two Dimensional Image Construction of Ultrasonic Wave for

Xu Huadong, Wang Lihai, Song Shiquan   

  1. College of Engineering & Technology, Northeast Forestry University Harbin 150040
  • Received:2013-07-20 Revised:2014-05-15 Online:2014-08-25 Published:2014-07-31
  • Contact: 王立海

摘要:

采用超声波手段,以椴木圆盘为研究对象,在试样完好和含不同大小孔洞时,测试并提取超声波信号特征值。依据超声波特征值构建训练集和测试集,利用支持向量机对原木孔洞缺陷的大小进行分类辨识,进而提出一种定量判别原木横截面内缺陷点位置的方法,分析并改进该方法存在的不足;在此基础上,实现原木横截面孔洞缺陷二维超声图像构建。结果表明:1) 支持向量机用于原木横截面孔洞缺陷直径大小的分类识别是可行的,准确率达到84.78%;2) 原木横截面孔洞缺陷二维图像模拟图与实物图重合度高,模拟效果较理想。

关键词: 超声波, 原木, 支持向量机, 无损检测, 图像构建

Abstract:

The ultrasonic propagation parameters in amur linden(Tilia amurensis)log specimen which was in intact and defective status respectively were measured and obtained. These parameters were then used as training set and test set to classify the hole size in log based on support vector machine(SVM). Furthermore, a kind of method to quantitatively determine the location of defect point on the cross section of log was proposed and improved. Based on this, the two dimensional simulation image of internal hole defect in log was constructed. The results showed that: 1) It was feasible to classify the hole size in log using SVM and the identification accuracy was 84.78%.2) The two dimensional simulation image of hole defect in the cross section was in good agreement with the actual image of log specimen.

Key words: ultrasonic wave, log, support vector machine(SVM), nondestructive test, image construction

中图分类号: