Hyperspectral measurement and identification of chinese herbal slices based on bidirectional game parallel deconvolution network

CHEN Yan; CAO Jianguo; XU Yan; YAN Huishen

您当前的位置：

首页 >

文章列表页 >

Hyperspectral measurement and identification of chinese herbal slices based on bidirectional game parallel deconvolution network

Optical Image Processing | 更新时间：2026-05-20

- Hyperspectral measurement and identification of chinese herbal slices based on bidirectional game parallel deconvolution network
- OPTICAL TECHNIQUE Vol. 52, Issue 3, Pages: 368-377(2026)
- 作者机构：
  
  1.扬州职业技术大学，江苏扬州225000
  2.南京中医药大学药学院，江苏南京210000
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP394.1;TH691.9
- Received：26 November 2025，
  
  Revised：2026-01-03，
  
  Published：20 May 2026
- 稿件说明：
移动端阅览
陈雁,巢建国,徐艳等.双向博弈平行反卷积网络下的饮片高光谱测量鉴别[J].光学技术,2026,52(03):368-377.

CHEN Yan,CAO Jianguo,XU Yan,et al.Hyperspectral measurement and identification of chinese herbal slices based on bidirectional game parallel deconvolution network[J].OPTICAL TECHNIQUE,2026,52(03):368-377.
陈雁,巢建国,徐艳等.双向博弈平行反卷积网络下的饮片高光谱测量鉴别[J].光学技术,2026,52(03):368-377. DOI：

CHEN Yan,CAO Jianguo,XU Yan,et al.Hyperspectral measurement and identification of chinese herbal slices based on bidirectional game parallel deconvolution network[J].OPTICAL TECHNIQUE,2026,52(03):368-377. DOI：

摘要

实现饮片智能鉴别代替传统感官判断，可被推广于日常生活或医药生产等领域。结合高、低卷积层特征提取网络提取丰富的饮片特征，并在生成对抗网络模型中引入全特征连接网络，可进一步提取高光谱曲线、均值、方差、最大和最小局部隐层特征。然后在生成和判别网络中增加条件参数，通过博弈思想提升网络的稳定性和有效性，最后将所得的特征传送到分类装置中完成分类鉴别。所提方法在简单和复杂背景下测试与训练样本占比小于4:7后，饮片鉴别准确率分别约为90.2%和79.6%，并可植入手机完成饮片鉴别。所提方法的各指标均优于IC-Net、Res-TCMNet、ResNet18、GhostNet。饮片图像背景复杂程度与清晰度均会影响饮片鉴别的准确率。

Abstract

Intelligent identification of prepared slices of Chinese materia medica replacing traditional sensory judgment can be popularized in daily life

pharmaceutical production and other fields. A feature extraction network with high and low convolutional layers is combined to extract abundant features of prepared slices

and a full-feature connection network is introduced into the Generative Adversarial Network （GAN） model to further extract the hyperspectral curves

mean values

variances

and the maximum and minimum local hidden layer features. Conditional parameters are then added to the generator and discriminator networks

and the stability and effectiveness of the network are improved based on the game theory. Finally

the extracted features are fed into the classification module to complete the classification and identification. The proposed method is tested under simple and complex background conditions with the ratio of test samples to training samples less than 4:7

achieving an identification accuracy of approximately 90.2% and 79.6% for prepared slices respectively

and it can be embedded in mobile phones for on-site identification. All evaluation indicators of the proposed method outperform those of IC-Net

Res-TCMNet

ResNet18 and GhostNet. Moreover

the complexity and definition of the background of prepared slice images both have a significant impact on the accuracy of prepared slice identification.

关键词

Keywords

references

周维 , 周永强 , 倪喆鑫 , 等 . 高原地区中医药的融合传承与创新发展思考 [J ] . 世界科学技术-中医药现代化 , 2025 , 27 （ 04 ）: 918 — 923 .

Zhou Wei , Zhou Yongqiang , Ni Zhexin , et al . Thoughts on the inheritance, and innovarive development of traditional chinese medicine in the plateau region [J ] . Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology , 2025 , 27 （ 04 ）: 918 — 923 .

李若彤 , 胡会强 , 曹诗宇 , 等 . 基于高光谱成像结合机器学习的半夏饮片鉴别 [J ] . 光谱学与光谱分析 , 2025 , 45 （ 05 ）: 1236 — 1242 .

Li Ruotong , Hu Huiqiang , Cao Shiyu , et al . Identification of pinelliae rhizoma decoction pieces by hyperspectral imaging combined with machine learning [J ] . Spectroscopy and Spectral Analysis , 2025 , 45 （ 05 ）: 1236 — 1242 .

姜帆 , 徐爱军 , 林鹏 , 等 . “十四五”时期我国中医药强省建设预期成效综合评价 [J ] . 卫生经济研究 , 2023 , 40 （ 11 ）: 16 — 21 .

Jiang Fan , Xu Aijun , Lin Peng , et al . Study on comprehensive evaluation of the expected results of building a strong province of traditional chinese medicine in china during the 14th five year plan period [J ] . Health Economics Research , 2023 , 40 （ 11 ）: 16 — 21 .

王志杰 , 樊薛津 , 王豫骞 , 等 . 机器学习方法在中医药研究中的应用进展 [J ] . 药物评价研究 , 2024 , 47 （ 08 ）: 1906 — 1913 .

Wang Zhijie , Fan Xuejin , Wang Yuqian , et al . Application of machine learning methods in research of traditional Chinese medicine [J ] . Drug Evaluation Research , 2024 , 47 （ 08 ）: 1906 — 1913 .

陶欧 , 林兆洲 , 张宪宝 , 等 . 基于饮片切面图像纹理特征参数的中药辨识模型研究 [J ] . 世界科学技术-中医药现代化 , 2014 , 16 （ 12 ）: 2558 — 2562 .

Tao Ou , Lin Zhaozhou , Zhang Xianbao , et al . Research on identification model of chinese herbal medicine by texture feature parameter of transverse section image [J ] . Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology , 2014 , 16 （ 12 ）: 2558 — 2562 .

谷瑞 , 宋翠玲 , 李元昊 . 基于BIM-Res2Net-ICA的中药饮片图像识别模型 [J ] . 时珍国医国药 , 2025 , 36 （ 06 ）: 1192 — 1200 .

Gu Rui , Song Cuiling , Li Yuanhao . To Construct an image recognition model for Chinese herbal medicine decoction pieces based on BIM-Res2Net-ICA [J ] . Journal of Li-shizhen Traditional Chinese Medicine , 2025 , 36 （ 06 ）: 1192 — 1200 .

罗晋 , 张俊华 , 罗旭东 , 等 . 一种融合注意力机制的轻量级重楼饮片分类方法 [J ] . 计算机应用与软件 , 2025 , 42 （ 06 ）: 186 — 192+201 .

Luo Jin , Zhang Junhua , Luo Xudong , et al . A lightweight paris l. decoction pieces recognition method based on attention mechanism [J ] . Computer Applications and Software , 2025 , 42 （ 06 ）: 186 — 192+201 .

李玥辰 , 赵晓 , 王若男 , 等 . 基于改进卷积神经网络的中药饮片图像识别 [J ] . 科学技术与工程 , 2024 , 24 （ 09 ）: 3596 — 3604 .

Li Yuechen , Zhao Xiao , Wang Ruonan , et al . Image recognition of chinese herbal slices based on improved convolution neural network [J ] . Science Technology and Engineering , 2024 , 24 （ 09 ）: 3596 — 3604 .

张琦 , 区锦锋 , 周华英 . 基于Xception与迁移学习的中药饮片图像识别研究 [J ] . 现代电子技术 , 2024 , 47 （ 03 ）: 29 — 33 .

Zhang Qi , Ou Jinfeng , Zhou Huaying . Research on traditional Chinese medicine piece image recognition based on Xception and transfer learning [J ] . Modern Electronic Technique , 2024 , 47 （ 03 ）: 29 — 33 .

韩勇 , 兰杰 , 郭瑞瑶 , 等 . 基于深度学习的中药材饮片识别 [J ] . 西北农业学报 , 2023 , 32 （ 11 ）: 1859 — 1867 .

Han Yong , Lan Jie , Guo Ruiyao , et al . Identification of chinese herbal medicine slices based on deep learning [J ] . Acta Agriculturae Boreali-occidentalis Sinica , 2023 , 32 （ 11 ）: 1859 — 1867 .

张谊 , 万华 , 涂淑琴 . 基于计算机视觉的中药饮片分类技术综述与案例研究 [J ] . 计算机应用 , 2022 , 42 （ 10 ）: 3224 — 3234 .

Zhang Yi , Wan Huam Hu Shuqin . Technical review and case study on classification of Chinese herbal slices based on computer vision [J ] . Journal of Computer Applications , 2022 , 42 （ 10 ）: 3224 — 3234 .

吴冲 , 谭超群 , 黄永亮 , 等 . 基于深度学习算法的川贝母、山楂及半夏饮片的智能鉴别 [J ] . 中国实验方剂学杂志 , 2020 , 26 （ 21 ）: 195 — 201 .

Wu Chong , Tan Chaoqun , Huang Yongliang , et al . Intelligent identification of fritillariae cirrhosae bulbus,crataegi fructus and pinelliae rhizoma based on deep learning algorithms [J ] . Chinese Journal of Experimental Traditional Medical Formulae , 2020 , 26 （ 21 ）: 195 — 201 .

张一凡 , 周苏娟 , 孟江 , 等 . 基于机器视觉系统的姜炭炮制程度判别及颜色-成分相关性分析 [J ] . 中国药房 , 2022 , 33 （ 22 ）: 2712 — 2718 .

Zhang Yifan , Zhou Sujuan , Meng Jiang , et al . Discrimination of processing degree of Zingiber officinale charcoal and analysis of the correlation between color and component based on machine vision system [J ] . China Pharmacy , 2022 , 33 （ 22 ）: 2712 — 2718 .

Miao J F , Huang Y A , Wang Z S , et al . Image recognition of traditional Chinese medicine based on deep learning [J ] . Frontiers in Bioengineering and Biotechnology , 2023 , 11 : 1199803 .

孙兴 , 李杨 , 毛天驰 , 等 . 改进YOLOv8在中药饮片自动识别标注中的研究与应用 [J ] . 中草药 , 2025 , 56 （ 01 ）: 216 — 226 .

Sun Xing , Li Yang , Mao Tianchi , et al . Research and application of improved YOLOv8 in automatic recognition andannotation of traditional Chinese medicine decoction pieces [J ] . Chinese Traditional and Herbal Drugs , 2025 , 56 （ 01 ）: 216 — 226 .

Kang Y F , Wu M X , Gao Y L , et al . Deep learning-based classification of traditional Chinese medicine: a novel approach [J ] . Quantitative Imaging in Medicine and Surgery , 2025 , 15 （ 8 ）: 7483 — 7496 .

丁羽 , 戴慧莲 , 姜德利 , 等 . 基于智能感官和近红外光谱技术结合化学计量学的酒当归炮制程度快速辨识研究 [J ] . 中草药 , 2025 , 56 （ 01 ）: 68 — 78 .

Ding Yu , Dai Huilian , Jiang Deli , et al . Research on rapid discrimination of different processing degrees of Angelicae Sinensis Radix with yellow rice wine based on intelligent sense and near-infrared technology combined with chemometrics [J ] . Chinese Traditional and Herbal Drugs , 2025 , 56 （ 01 ）: 68 — 78 .

Zhang C , Han M . Multi-feature hyperspectral image classification with L-2,L-1 norm constrained joint sparse representation [J ] . Int J Remote Sens , 2021 , 42 （ 12 ）： 4789 — 4808 .

Liu B , Yu A Z , Tan X . Slow feature extraction for hyperspectral image classification [J ] . Remote Sens Lett , 2021 , 12 （ 5 ）： 429 — 438 .

Gu G H , Li Z Y , Feng L J , et al . Multi-semantic region weighting and multi-scale flatness weighting based image retrieval [J ] . Soft Comput , 2021 , 25 （ 7 ）: 5699 — 5708 .

罗春梅 , 张风雷 . 基于均值特征和改进深度神经网络的说话人识别算法 [J ] . 声学技术 , 2021 , 40 （ 04 ）: 503 — 507 .

Luo Chunmei , Zhang Fenglei . Speaker recognition based on mean feature andimproved deep neural network [J ] . Technical Acoustics , 2021 , 40 （ 04 ）: 503 — 507 .

杨磊 , 陈换过 , 高祥冲 , 等 . 基于慢特征方差聚类的风电机组主传动系统过渡状态划分 [J ] . 机电工程 , 2025 , 42 （ 10 ）: 1877 — 1887 .

Yang Lei , Chen Huanguo , Gao Xiangchong , et al . Transition state division of wind turbine main transmission system based on slow feature variance clustering [J ] . Journal of Mechanical & Electrical Engineering , 2025 , 42 （ 10 ）: 1877 — 1887 .

Li B , Du J , Zhang X P . Feature extraction using maximum nonparametric margin projection [J ] . Neurocomputing , 2016 , 188 : 225 — 232 .

Wang L , Zhang J S , Zang F . An efficient feature extraction method, global between maximum and local within minimum, and its applications [J ] . Mathematical Problems in Engineering , 2011 , 2011 : 176058 .

Du X L , Li K , Lv Y , et al . Motor imaging EEG signal recognition of resnet18 network based on deformable convolution [J ] . Electronics , 2022 , 11 （ 22 ）: 3674 .

Xiao J , Huang H S , Chen X R , et al . Identification of tea bud with improved DCGAN algorithm and GhostNet-RCLAM network [J ] . Journal of Food Measurement and Characterization , 2023 , 17 （ 4 ）: 4191 — 4207 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Design of simulated eye for calibration of vision screening instrument based on eccentric photography principle

Infrared and visible light images fusion based on frequency transformation enhancement

The deployment of high-energy laser weapons C-RAM

A robust zero-phase joint extraction method for quasi-static deformation signals in optomechanical structures

Design of microscope imaging system for double-telecentric and two-position zoom microscope

Related Author

No data

Related Institution

No data

⁰