基于人工智能的海南黄花梨木与越南黄花梨木识别研究

Study on Identification Eigenvalue of Hainan Huanghua Pear and Vietnam Huanghua Pear Based on Image Processing

为实现对海南黄花梨木及越南黄花梨木成品家具在无损伤的情况下进行分类识别,本文通过采集成品家具纹理图像并分别对图像在RGB和HSV颜色空间下的颜色直方图各分量的峰值进行分析,计算图像灰度共生矩阵(GLCM)在步长d取值为2和4的情况下4个方向上图像纹理的对比度、相关性、熵、平稳度、能量,5个特征值用SPSS软件做显著性分析。通过分析最终确定R、G、B、V、相关性、对比度、能量,7个特征值作为海南黄花梨与越南黄花梨识别的特征值向量,通过对比BP神经网络与SVM向量机2种识别方法,发现SVM向量机准确率更高,准确率可达到94%。结果表明本文方法可作为一种在无损伤的情况下识别海南黄花梨木与越南黄花梨木成品家具的可靠方法。

In order to realize the hainan yellow rosewood and Vietnam rose wood furniture products in the case of no damage identification, this paper respectively on image in RGB and HSV color space of color histogram peak analysis of each component of image gray level cooccurrence matrix in step d values of 2 and 4 cases respectively from four directions contrast of image texture, smooth, entropy, correlation degree, five energy eigenvalue analysis with SPSS software.Seven eigenvalues of R, G, B, V, correlation, contrast and energy were determined as the eigenvalue vectors of sea yellow and sea yellow, and the sample images were trained and recognized by BP neural network.The results show that the method presented in this paper can be used as a very effective method for the identification of sea-yellow and sea-yellow.