AUTHOR=Niu Qunfeng , Liu Jiangpeng , Jin Yi , Chen Xia , Zhu Wenkui , Yuan Qiang 

TITLE=Tobacco shred varieties classification using Multi-Scale-X-ResNet network and machine vision

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.962664

DOI=10.3389/fpls.2022.962664

ISSN=1664-462X

ABSTRACT=Identifying four tobacco shred types of tobacco, including expanded tobacco silk, cut stem, tobacco silk, and reconstituted tobacco shred, is the primary task in calculating the tobacco shred blending ratio. The classification precision directly affects the subsequent determination of tobacco shred components. Because the types of tobacco shred, especially the expanded tobacco silk and tobacco silk, have no apparent difference in macro-scale characteristics. The tobacco shred has small size and irregular shape characteristics, bringing significant challenges to the recognition and classification based on machine vision. Aiming at this problem, this paper provides a complete set of solutions for screening tobacco shred samples, taking images, image preprocessing, establishing data sets, and identifying types. In this paper, ResNet50 is used as the mainframe of the classification and recognition network. By increasing the multi-scale structure, optimizing the number of blocks and loss function, a new tobacco shred image classification method based on MS-X-ResNet (Multi-Scale-X-ResNet) network is proposed. The experimental results show that the final classification accuracy of the network with tobacco shred data set is 96.56%. The image recognition of single tobacco shred requires 103ms, achieving high classification accuracy and efficiency. The image preprocessing and deep learning algorithms for classification and identification of tobacco shred proposed in this paper provide a new implementation idea for the actual production and quality detection of tobacco and a new way for online real-time type identification of other agricultural products.