摘要
在烟叶外观品质特征定量化的基础上，构造了烟叶质量自适应检验数学模型，目的是减少系统对矛盾数据的敏感性，以实现在标准烟叶样本较少的情况下进行学习。数学模型包括分组分级模型和权重优化模型。本文对构造的模型进行了实际应用，应用结果表明该数学模型对不同年份、不同产地烟叶分组分级效果稳定。
The mathematical models for adaptively examining the quality of tobacco leaves were built with the theory of fuzzy mathematics and artificial neural network, on the basis of the quantitative analysis of external quality of tobacco leaves,in order to reduce the system's sensitivity to contradictory dates and train the system under the condition of having only a few standard tobacco leaves. The models include grouping and grading model and weight optimization model. The models built were put into practice and the result indicated that they have stable effectiveness in grading and grouping for tobacco leaves of different years and different fields.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
1998年第3期99-103,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
机械部教育司基金
关键词
烟草
图像处理
分级
质量
自适应分级
Tobacco, Image processing, Model construction, Grading
