基于灰度共生矩阵的棉花图像纹理特征研究
Research on Texture Feature of Cotton Images based on Gray Level Cooccurrence Matrix
本文采用基于统计的灰色共生矩阵方法对不同含水量的棉花图像进行纹理分析。在对图像预处理的基础上，研究了矩阵的不同灰度级，像素间的不同距离和不同方向上的取值对灰度共生矩阵方法的纹理特征值的影响。通过分析得到灰度级取16，像素间距离取8，方向取0度和90度或45度和135度的平均值对于反映棉田纹理的特点较为适宜。利用定量的灰度共生矩阵提取棉花样本中6种纹理综合特征值。评价各个纹理特征对不同样本类型的区分效果，为有效的区分的不同含水量棉田样本提供了依据。
his paper introduces gray level coocurrence matrix (GLCM) into cotton image of different water content for texture analysis. On the basis of images pretreatment, the impact of grey level, pixel distance and pixel angle over characteristic value of gray level coocurrence matrix is researched. After analyzing diversification rule of texture feature, concludes that gray level equals to 16,distances equals to 8,and angle equals to the average of 0°and 90°，or 45°and 135°in matrix are suitable for cotton texture analysis. Then six texture features are extracted from quantitative matrix, and evaluate the efficient of distinguish cotton types using texture features.
李书颖
农业科学研究生物科学研究方法、生物科学研究技术计算技术、计算机技术
灰度共生矩阵纹理分析特征提取
gray level cooccurrence matrixexture analysisFeature extraction
李书颖.基于灰度共生矩阵的棉花图像纹理特征研究[EB/OL].(2009-01-21)[2025-08-11].http://www.paper.edu.cn/releasepaper/content/200901-972.点此复制
评论