农产品质量的计算机辅助检验与分级(第Ⅱ报)——烟叶自动分级模型的建立与训练

Computer Aided Quality Inspecting and Grading for Agricultural Products(2nd half)—— Construction and Training of Grading Model for Tobacco Leaf

摘要: 在外观品质特征定量检测的基础上，研究了烟叶的自动分级技术。其基本原理是，建立标准样本烟叶的特征数据库，将待测烟叶与样本烟叶比较而进行分级。研究了数据库记录及烟叶差异的计算方法；应用神经网络技术构造了识别模型，进行了智能化训练。

Abstract: A new automatic grading technology of tobacco leaf was developed. The principle of grading is comparing the inspected tobacco leaf with the standard tobacco leaves. Recording theory of external characters of standard tobacco leaves in database, and the calculation approach of difference between the inspected and the standard tobacco leaves were studied. The technique of neural network was used to build up a trainable recognizing model.