The main basis for the identification of tobacco leaf quality and ripening degree was summarized.The convolution neural network reconstructed by the automatic encoder pre-training method was used to construct the tobacco leaf quality identification model.The experimental data were used to verify the experimental results and the results showed that the reconstructed depth training self-encoder achieved 99.92% accuracy in classification performance.%概述了烟叶质量和熟成度分类的主要依据,采用自动编码器预训练方法重构的卷积神经网络构建了烟叶质量识别模型,并采用实地采集的数据集进行了实验验证,结果表明重构的深度训练自编码器在分类性能上达到99.92%的准确度.
展开▼