-
摘要: 正确识别木材对木材科学和木材产业具有重要意义。提出一种提取木材显微图像特征并进行识别的新方法。首先进行傅里叶变换得到木材显微图像的傅里叶变换功率谱图,然后进行独立成分分解得到功率谱图的独立基,所提取的特征就是木材显微图像的功率谱图在独立基上的投影系数,最后利用支持向量机对待识别图像在独立基上的投影系数进行分类,实现木材识别。在200幅木材显微图像库上进行小样本实验,取得了较高的识别率。实验结果表明,该方法具有较大应用潜力。图4参12Abstract: Wood recognition plays an important role in wood science and industry. A new method for wood micrograph feature extraction and wood recognition is proposed. The fast Fourier transform(FFT) is firstly used to get power spectrum of wood micrograph,then the independent component analysis algorithm is performed to extract feature which is the coefficient of wood micrograph projecting to the independent components. Finally the wood can be recognized by classifying the coefficients using support vector machines. A small sample experiment using 200 wood micrographs indicates a high recognition rate and effectiveness of the method.[Ch,4 fig. 12 ref.]
-
链接本文:
https://zlxb.zafu.edu.cn/article/doi/10.11833/j.issn.2095-0756.2010.06.004
计量
- 文章访问数: 4767
- HTML全文浏览量: 397
- PDF下载量: 1400
- 被引次数: 0