基于颜色空间模型的烟叶图像分割研究

Study on segmentation of tobacco images based on color model

对于图像分割,在实际运用中尚无通用的算法实现对所有图像分割。针对烟叶图像的分割,传统的基于区域生长、阈值分割等方法分割效果并不理想。为此,文中提出基于颜色空间模型的K-means聚类分割算法。先提取烟叶图像的颜色分量,以RGB颜色分量和HSV分量为聚类样本点进行K-means聚类分割,比较两种算法的结果。结果表明:基于HSV颜色空间模型对烟叶分割效果比较好,适用于对烟叶的图像分割,为准确提取烟叶特征奠定基础。

For image segmentation,there is no general algorithm to segment all the images in the practical applications. For the segmentation of tobacco images,the effect of the traditional segmentation methods based on region growing and threshold segmentation are not satisfactory. Therefore,this paper proposes a K-means clustering segmentation algorithm based on color space model. Firstly,the color components of the tobacco image are extracted,the RGB color components and the HSV components are used as the clustering sample points to do the K-means clustering segmentation,and the results of the two algorithms are compared. The results show that the segmentation of tobacco leaves based on HSV color space model is better. It is suitable for image segmentation of tobacco,and lays the foundation for accurate extraction of tobacco characteristics.