Leaf color is an important indicator of flue-cured tobacco quality; however, there is a big difference among similar color tobacco leaves from different areas. Color features of tobacco leaves were obtained by using Region Growing Method in the pre-processing, and then made region classification by using Support Vector Machine (SVM). The results showed that in the case of small samples, radial basis function was the kernel function of the SVM, and the appropriate model parameters were determined. The classification accuracies for training set and test set of the SVM model were 100% and 86.67%, showing that SVM has a perfect performance in color region classification of tobacco leaves.%颜色是烤烟烟叶品质的重要外在指标之一,在生产中,同类颜色烟叶在不同产地却往往存在着较大的差异.采用区域生长方法对烟叶图像进行分割预处理,然后提取烟叶的颜色特征,再运用一种新的机器学习算法—支持向量机分类方法对我国烟叶颜色特征进行区域分类.结果发现在小样本情况下,采用径向基函数作为支持向量模型的核函数,并确定了适当的模型参数,所建立模型对烟叶颜色区域特征的回判识别率达100%,预测识别率达86.67%.支持向量机对典型产地烟叶颜色的分类识别具有良好的应用性能.
展开▼