数字农科院2.0

A novel recognition method for the rolling degree of roasted green tea based on optimized segmentation algorithm and multi-source physical features fusion deep learning framework

文献类型： 外文期刊

作者： Jin, Huaqiang;Huang, Aowen;Tan, Junfeng;Sun, Zhe;Xu, Yingjie;Gu, Jiangping;Li, Kang;Shi, Lin;Yao, Qiwei;Shen, Xi

作者机构：

关键词： Tea leaf;Rolling quality;Machine vison;Background removal;Deep neural network

期刊名称： COMPUTERS AND ELECTRONICS IN AGRICULTURE

ISSN： 0168-1699

年卷期： 2025 年 242 卷

页码：

收录情况： SCIE（2025版） ; ; EI（2025版）

摘要： Rolling is the core process in roasted green tea (RGT) production, critically shaping its physical form and flavor constituents' release. The rolling quality assessment of RGT currently relies on subjective manual evaluation, which presents a significant barrier to automation and intelligent processing within the tea industry. To overcome this limitation, this study proposes an innovative deep learning-based framework combined with an optimized segmentation algorithm for automated RGT rolling degree assessment. This approach employs Otsumaximin Latin hypercube sampling differential evolution (Otsu-MLHSDE) algorithm to precisely segment complex processing background and tea regions. The Squeeze-and-Excitation attention mechanism and multi-source physical features fusion (SE-MPFF) model is developed to recognize the rolling degree, enhancing both the predictive accuracy and interpretability of the deep learning framework. Experimental results demonstrated that the proposed SE-MPFF model achieved a recognition accuracy of 97.32 %. It also delivered rapid diagnosis, with an average processing time of 0.210 s per recognition task. These results showed a promising potential for using optimized segmentation and deep learning methods to recognize rolling degree, thereby providing a basis for continuous processing and intelligent quality control in RGT production.

分类号：

  • 相关文献

[1]Combining novel feature selection strategy and hyperspectral vegetation indices to predict crop yield. Fei S.,Li L.,Han Z.,Chen Z.,Xiao Y.. 2022

[2]Instance Segmentation and Number Counting of Grape Berry Images Based on Deep Learning. Yanmin Chen,Xiu Li,Mei Jia,Jiuliang Li,Tianyang Hu,Jun Luo. 2023

作者其他论文 更多>>