基于灰度共生矩阵的图像纹理特征提取研究

Research on extraction of Image texture Feature based on Gray co-occurrence matrix

计算机视觉系统存在图像数据维度高、图像数据类型冗杂的情况,因此提取图像的纹理特征对图像的后期处理非常重要。采用灰度共生矩阵的图像纹理特征提取方法,可以有效地获取图像在像素灰度间的方向、距离、变化幅度等综合信息,其纹理特征值差异和变化规律可准确描述图像纹理的特征。仿真实验证明了该方法可行,通过对比不同图像的纹理特征值,得出其差异和变化规律基本与实际的纹理特征情况相符合,验证了此方法的有效性。

In computer vision system,the image data dimension is high and the image data type is miscellaneous.Therefore,it is very important to extract the texture feature of the image in post-processing.The gray co-occurrence matrix is used to extract the image texture features,which can effectively obtain the comprehensive information such as the direction,distance and change amplitude of the image between the pixel gray levels,and the difference and change rule of the statistical texture feature values can accurately describe the image texture features.Simulation experiments show that this method is feasible.By comparing the texture feature values of different images,it is concluded that the difference and variation rules are basically consistent with the actual texture features,and the effectiveness of this method is verified.