基于感兴趣区域图像分割的红茶茶毫品质数字化评价研究

仝晨; 梁秀华; 王周立; 范冬梅; 林招水; 吴小妹; 阙杨战; 金敏丽; 林杰

doi:10.13305/j.cnki.jts.2025.04.008

PDF(2070 KB)

茶叶科学 ›› 2025, Vol. 45 ›› Issue (4) : 655-670. DOI: 10.13305/j.cnki.jts.2025.04.008

研究报告

基于感兴趣区域图像分割的红茶茶毫品质数字化评价研究

仝晨¹, 梁秀华², 王周立¹, 范冬梅¹, 林招水³, 吴小妹⁴, 阙杨战⁴, 金敏丽^1,5,*, 林杰^1,*

作者信息 +

Digital Evaluation of Black Tea Hairs Quality Based on Region of Interest Image Segmentation

TONG Chen¹, LIANG Xiuhua², WANG Zhouli¹, FAN Dongmei¹, LIN Zhaoshui³, WU Xiaomei⁴, QUE Yangzhan⁴, JIN Minli^1,5,*, LIN Jie^1,*

Author information +

文章历史 +

摘要

茶毫是红茶外形品质的重要评价指标,当前主要依赖于专业人员的感官评价,主观性强且评语抽象,缺乏客观化、数字化的品质评价手段。为构建茶毫品质数字化评价方法,采集3个不同茶毫品质等级的祁门红茶样品图像,采用HSV彩色图像分割技术对感兴趣区域（Region of interest,ROI）提取HSV颜色空间分量特征,构建分割指数（Segmentation index,SI）检索得到茶毫、茶身和阴影的最佳分割阈值,采用掩膜法和像素点判别对图像分割效果进行定性和定量评价,并构建茶毫比例量化方法。结果表明,茶毫、茶身和阴影区域的平均分割准确率达到了98.70%,进一步通过茶毫比例量化结果获得祁门红茶3个茶毫品质等级（“显毫”“多毫”和“少毫”）的推荐毫量比例阈值。不同毫量梯度拼配茶样的线性回归分析（R²=0.958,P<0.01）及滇红、金骏眉的泛化应用效果表明,构建的茶毫品质数字化评价方法在不同毫量区间和不同红茶类别上具有较好的适应性。

Abstract

Tea hairs are an important indicator for evaluating the appearance quality of black tea. Currently, the evaluation mainly relies on the sensory assessment of professionals, which is highly subjective and abstract, lacking objective and digital quality evaluation methods. To construct a digital evaluation method for the quality of tea hairs, this study collected images of Qimen black tea samples with three different tea hairs quality grades. The HSV color image segmentation technique was used to extract the HSV color space component features of the region of interest (ROI), and a segmentation index (SI) was constructed to retrieve the optimal segmentation thresholds for tea hairs, tea body, and shadows. The image segmentation effect was qualitatively and quantitatively evaluated using the masking method and pixel point discrimination, and a quantification method for the proportion of tea hairs was established. The results show that the average segmentation accuracy of tea hairs, tea body, and shadow areas reached 98.70%. Furthermore, through the quantification of the proportion of tea hairs, the recommended hairs proportion thresholds for the three tea hairs quality grades of Qimen black tea (“Prominent golden hairs”, “Observable hairs”, and “Less hairs”) were obtained. The linear regression analysis of different hairs gradient blended tea samples (R²=0.958, P<0.01) and the generalization application effects on ‘Dianhong’ and ‘Jinjunmei’ indicate that the digital evaluation method of tea hairs quality constructed in this study has good adaptability in different hairs quantity intervals and different black tea categories.

导出引用

仝晨, 梁秀华, 王周立, 范冬梅, 林招水, 吴小妹, 阙杨战, 金敏丽, 林杰. 基于感兴趣区域图像分割的红茶茶毫品质数字化评价研究[J]. 茶叶科学. 2025, 45(4): 655-670 https://doi.org/10.13305/j.cnki.jts.2025.04.008

TONG Chen, LIANG Xiuhua, WANG Zhouli, FAN Dongmei, LIN Zhaoshui, WU Xiaomei, QUE Yangzhan, JIN Minli, LIN Jie. Digital Evaluation of Black Tea Hairs Quality Based on Region of Interest Image Segmentation[J]. Journal of Tea Science. 2025, 45(4): 655-670 https://doi.org/10.13305/j.cnki.jts.2025.04.008

中图分类号： S571.1 S126

参考文献

[1] 郭元超. 茶树茸毛的形态特征与演化特点[J]. 茶叶科学简报, 1993(3): 1-4.
Guo Y C.The morphological characteristics and evolution characteristics of tea tree trichomes[J]. Chaye Kexue Jianbao, 1993(3): 1-4.
[2] Cao H L, Li J M, Ye Y J, et al.Integrative transcriptomic and metabolic analyses provide insights into the role of trichomes in tea plant (Camellia Sinensis)[J]. Biomolecules, 2020, 10(2): 311. doi:10.3390/biom10020311.
[3] Li P H, Xu Y J, Zhang Y R, et al.Metabolite profiling and transcriptome analysis revealed the chemical contributions of tea trichomes to tea flavors and tea plant defenses[J]. Journal of Agricultural and Food Chemistry, 2020, 68(41): 11389-11401.
[4] 张明露, 雷睿勇, 邓春梅. 茶毫对茶叶品质的影响分析[J]. 贵茶, 2023(1): 28-31.
Zhang M L, Lei R Y, Deng C M.Analysis of the influence of Tea Hairs on the quality of tea[J]. Journal of Guizhou Tea, 2023(1): 28-31.
[5] 王舒婷, 曲凤凤, 张新富, 等. 基于电子鼻技术的白茶毫香研究[J]. 青岛农业大学学报(自然科学版), 2020, 37(4): 258-263.
Wang S T, Qu F F, Zhang X F, et al.Study on the fragrance of white tea hairs based on electronic nose technology[J]. Journal of Qingdao Agricultural University (Natural Science), 2020, 37(4): 258-263.
[6] 中华人民共和国国家质量监督检验检疫总局、中国国家标准化管理委员会. 茶叶感官审评术语: GB/T 14487—2017[S]. 北京: 中国标准出版社, 2017.
General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China (AQSIQ), Standardization Administration of China (SAC). Tea vocabulary for sensory evaluation: GB/T 14487—2017 [S]. Beijing: China Standard Press, 2015.
[7] Zhu H K, Ye Y, He H F, et al.Evaluation of green tea sensory quality via process characteristics and image information[J]. Food and Bioproducts Processing, 2017, 102: 116-222.
[8] Galmarini M V.The role of sensory science in the evaluation of food pairing[J]. Current Opinion in Food Science, 2020, 33: 149-155.
[9] 胡梦芹. 茶毫发育调控的分子基础研究[D]. 长沙: 湖南农业大学, 2021.
Hu M Q.Studies on molecular basis of the regulation of trichome formation in tea plant (Camellia sinensis L.) [D]. Changsha: Hunan Agricultural University, 2021.
[10] 宋亚康, 张群峰, 张洁, 等. 茶毫氨基酸组成及矿质元素分析[J]. 茶叶科学, 2017, 37(4): 339-346.
Song Y K, Zhang Q F, Zhang J, et al.Study on amino acids composition and mineral elements of tea hairs[J]. Journal of Tea Science, 2017, 37(4): 339-346.
[11] Morris T.Computer vision and image processing[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1992, 48(2): 24-25.
[12] Borianne P, Sarron J, Borne F, et al.Deep mango cultivars: cultivar detection by classification method with maximum misidentification rate estimation[J]. Precision Agriculture, 2023, 24(4): 1619-1637.
[13] Hou L X, Liu Z Y, You J X, et al.Tomato sorting system based on machine vision[J]. Electronics, 2024, 13(11): 2114. doi:10.3390/electronics13112114.
[14] Gill G S, Kumar A, Agarwal R.Monitoring and grading of tea by computer vision: a review[J]. Journal of Food Engineering, 2011, 106(1): 13-19.
[15] 邱晓莹, 曹思宇, 陈昊东, 等. 黄金芽与群体种制成条形、螺形红茶的颜色空间数字化分析[J]. 浙江农业科学, 2024, 65(6): 1445-1450.
Qiu X Y, Cao S Y, Chen H D, et al.Digital analysis of the color space of strip and spiral-shaped black tea made from Huangjinya and population resources[J]. Journal of Zhejiang Agricultural Sciences, 2024, 65(6): 1445-1450.
[16] Pal N R, Pal S K.A review on image segmentation techniques[J]. Pattern Recognit, 1993, 26(9): 1277-1294.
[17] 章展熠, 张宝荃, 王周立, 等. 多茶类CNN图像识别的数据增强优化及类激活映射量化评价[J]. 茶叶科学, 2023, 43(3): 411-423.
Zhang Z Y, Zhang B Q, Wang Z L, et al.Data enhancement optimization and class activation mapping quantitative evaluation for cnn image recognition of multiple tea categorie[J]. Journal of Tea Science, 2023, 43(3): 411-423.
[18] 甘密, 汪飞, 沈强, 等. 基于图像处理的茶叶识别方法研究[J]. 贵茶, 2022(2): 28-32.
Gan M, Wang F, Shen Q, et al.Research on tea recognition method based on image processing[J]. Journal of Guizhou Tea, 2022(2): 28-32.
[19] Ding L, Zhang Y P, Zhang X Y, et al.A survey of image segmentation techniques and performance evaluation[J]. Computer Engineering and Software, 2010, 31(12): 78-83.
[20] Chen W, Shi Y Q, Xuan G R.Identifying computer graphics using HSV color model and statistical moments of characteristic functions[C]//IEEE. Proceedings of the IEEE International Conference on Multimedia and Expo, 2007: 1123-1126.
[21] Zhao B J, Wei D, Sun W Z, et al.Research on tea bud identification technology based on HSI/HSV color transformation[C]//IEEE. Proceedings of the International Conference on Information Science and Control Engineering, 2019: 511-515.
[22] Shelhamer E, Long J, Darrell T.Fully convolutional networks for semantic segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4): 640-651.
[23] 熊俊涛, 卜榕彬, 郭文韬, 等. 自然光照条件下采摘机器人果实识别的表面阴影去除方法[J]. 农业工程学报, 2018, 34(22): 147-154.
Xiong J T, Bu R B, Guo W T, et al.Shadow removal method of fruits recognized by picking robot under natural environment[J]. Nongye Gongcheng Xuebao/Transactions of the Chinese Society of Agricultural Engineering, 2018, 34(22): 147-154.
[24] 周兵, 谭卓昆, 段红星, 等. 基于GA-BP算法的红茶色泽分类研究[J]. 福建茶叶, 2022, 44(2): 15-17.
Zhou B, Tan Z K, Duan H X, et al.Study on color classification of black tea based on GA-BP algorithm[J]. Tea in Fujian, 2022, 44(2): 15-17.
[25] 金美霞, 陈晓阳, 王霆, 等. 木禾种红茶不同工艺色泽品质的多维度比较[J]. 浙江农业科学, 2021, 62(8): 1603-1607.
Jin M X, Chen X Y, Wang T, et al.Multi-dimensional comparative study on color quality of ‘Muhezhong’ black tea produced by different processes[J]. Journal of Zhejiang Agricultural Sciences, 2021, 62(8): 1603-1607.
[26] 冯云丛. 基于信息融合的医学图像阈值分割及评价方法研究[D]. 长春: 吉林大学, 2017.
Feng Y C.Research on medical image threshold segmentation and evaluation methods based on information fusion [D]. Changchun: Jilin University, 2017.
[27] 杨会芳. 基于计算机视觉的鲜茶叶识别与分类系统研究[D]. 重庆: 重庆三峡学院, 2024.
Yang H F.Research on fresh tea leaf recognition and classification system based on computer vision [D]. Chongqing: Chongqing Three Gorges University, 2024.
[28] 张昆明. 基于YOLOv4的自然环境下猴魁茶树芽叶采摘点的三维定位[D]. 合肥: 安徽农业大学, 2022.
Zhang K M.Three-Dimensional localization of Monkey King Tea shoot picking points in natural environment based on YOLOv4 [D]. Hefei: Anhui Agricultural University, 2022.
[29] 郑明君. 基于机器学习的雪茄烟草种植数字化关键技术研究[D]. 泰安: 山东农业大学, 2024.
Zheng M J.Research on digital key technologies for cigar tobacco cultivation based on machine learning [D]. Tai'an: Shandong Agricultural University, 2024.
[30] Zhang L, Zou L, Wu C Y, et al.Method of famous tea sprout identification and segmentation based on improved watershed algorithm[J]. Computers and Electronics in Agriculture, 2021, 184(1): 106108. doi:10.1016/j.compag. 2021.106108.
[31] 胡程喜, 谭立新, 王文胤, 等. 基于改进DeepLabV3+的轻量化茶叶嫩芽采摘点识别模型[J]. 智慧农业(中英文), 2024, 6(5): 119-127.
Hu C X, Tan L X, Wang W Y, et al.Lightweight tea bud picking point recognition model based on improved DeepLabV3+[J]. Smart Agriculture, 2024, 6(5): 119-127.
[32] Zou K, Ge L, Zhou H, et al.An apple image segmentation method based on a color index obtained by a genetic algorithm[J]. Multimedia Tools and Applications, 2022, 81(6): 8139-8153.
[33] 周任虎, 席家新, 丁以纾, 等. 基于K-means聚类算法的烘烤烟叶图像分割研究[J]. 安徽农业科学, 2024, 52(19): 232-237.
Zhou R H, Xi J X, Ding Y S, et al.Research on image segmentation of flue-cured tobacco leaf based on K-means clustering algorithm[J]. Journal of Anhui Agricultural Sciences, 2024, 52(19): 232-237.
[34] 中华人民共和国国家质量监督检验检疫总局、中国国家标准化管理委员会. 红茶第2部分: 工夫红茶: GB/T 13738.2—2017[S]. 北京: 中国标准出版社, 2017.
General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China (AQSIQ), Standardization Administration of China (SAC). Black tea—Part 2: Congou black tea: GB/T 13738.2—2017 [S]. Beijing: China Standard Press, 2017.
[35] 中华全国供销合作总社. 祁门工夫红茶: GH/T 1178—2019[S]. 北京: 中国标准出版社, 2019.
All China Federation of Supply and Marketing Cooperatives. KEEMUN Congou Black tea: GH/T 1178—2019 [S]. Beijing: China Standard Press, 2019.
[36] Zhu H, Ye Y, He H, et al.Evaluation of green tea sensory quality via process characteristics and image information[J]. Food and Bioproducts Processing, 2017, 102: 116-222.
[37] 李晶. 福建红茶品质特征研究与综合评价[D]. 福州: 福建农林大学, 2023.
Li J.Research and comprehensive evaluation of quality characteristics of Fujian black tea [D]. Fuzhou : Fujian Agriculture and Forestry University, 2023.
[38] 赵磊. 茶叶拼配过程中的若干机器学习方法研究[D]. 合肥: 安徽农业大学, 2023.
Zhao L.Research on several machine learning methods in the process of tea blending [D]. Hefei: Anhui Agricultural University, 2023.
[39] 孙彬妹, 刘少群, 刘任坚, 等. 茶树茸毛的研究进展[J]. 茶叶通讯, 2018, 45(4): 3-6.
Sun B M, Liu S Q, Liu R J, et al.Research progress of tea tree hairs[J]. Journal of Tea Communication, 2018, 45(4): 3-6.
[40] 圣阳. 基于红外光谱和拉曼光谱技术的茶叶品质研究[D]. 南京: 南京林业大学, 2022.
Sheng Y.Research on tea quality based on infrared and raman spectroscopy techniques [D]. Nanjing: Nanjing Forestry University, 2022.
[41] 尹鹏, 刘威, 王广铭, 等. 茶树芽叶茸毛及茶毫研究进展[J]. 河南农业, 2016(27): 44-46.
Yin P, Liu W, Wang G M, et al.Research progress of tea bud leaf hair and tea hair[J]. Agriculture of Henan, 2016(27): 44-46.