Recognition of Tobacco Leaf Curing Stage Based on Deep Learning

The aim of this study is to improve the intelligent curing technology in Jiangxi and promote its high-quality development. Four different deep learning models, namely AlexNet, VGG16, GoogLeNet, and ResNet18, were used to recognize the typical curing stages of tobacco leaves in Jiangxi. Firstly, a dataset of typical curing stages of tobacco leaves in the Jiangxi region was established. Secondly, the four deep learning models were trained and optimized. Finally, the performance of the four models was evaluated and analyzed. The experimental results showed that all four deep learning models were capable of recognizing the typical curing stages of tobacco leaves in Jiangxi. However, the ResNet18 deep learning model performed the best, achieving an accuracy of 95.80% on the test dataset. In terms of performance, it exhibited higher classification accuracy and stability compared to the other three deep learning models. This study has important practical value and significance for the automated management and quality control of tobacco leaf curing processes in the Jiangxi region.