中文摘要：

为提高烟叶等级分类效率和烟叶产品品质，减轻人工劳动强度，基于极限学习机提出了一种烟叶成熟度快速分类方法：首先将烟叶图像归一化处理，将烟叶图像平均分成4块，然后提取烟叶图像的分块颜色直方图特征，利用主成分分析法对提取的特征进行降维处理，最后利用极限学习机进行识别判断。仿真实验结果表明，将极限学习机应用于烟叶成熟度分类，测试精度可达96．43％，其训练速度和泛化性均优于BP神经网络和支持向量机，能够快速、准确地判断烟叶成熟度，具有潜在的实用价值。

英文摘要：

For improving the efficiency of grading and the quality of tobacco leaf and reducing the labor intensity of operators, a leaf maturity classifying method based on extreme learning machine was proposed. Firstly, the image of tobacco leaf was equally divided into 4 blocks after being normalized; secondly, principal component analysis （PCA） was conducted to reduce the dimension of the extracted characteristics; finally, extreme learning machine was adopted to identify leaf maturity. The results of simulation experiment showed that the precision of test reached 96.43% when extreme learning machine was applied. Extreme learning machine could identify the maturity of tobacco leaf rapidly and accurately, it was of potential practical uses and was better than support vector machine and BP neural network in terms of training speed and generalization.