枸杞果柄结合力影响因素分析及模型预测

doi:10.16420/j.issn.0513-353x.2024-0466

摘要/Abstract

摘要：

分析枸杞果柄结合力及其相关物理特性参数之间的内在联系与影响规律，发掘重要果实参数，建立果柄结合力预测模型，以期为提高枸杞机械采摘的效率和精确度提供依据。以青海省主栽枸杞品种‘宁杞1号’和‘宁杞7号’成熟期果实为研究对象，通过图像处理方法获得不同成熟度枸杞果实的RGB颜色模型分量，分析了不同品种、成熟度、采摘期温度、湿度及果实在枝条上的位置等对果柄结合力的影响，以及果柄参数、果实参数等与果柄结合力的相关性，构建果柄结合力预测模型。结果显示，可通过RGB颜色模型区分成熟度，但不同成熟度果柄结合力重叠部分较多，不能实现精确采摘。双因素方差分析结果显示品种和成熟度均对结合力有显著影响；果柄参数、果实参数和果实硬度与果柄结合力显著相关；正交试验表明果实在枝条上的位置、采摘时的温度和湿度对结合力没有显著影响。对偏最小二乘回归（PLSR）、随机森林（RF）和反向传播神经网络（BPNN）等建模方法的参数进行优化，建立果柄结合力的预测模型，其中BPNN表现出了优秀的预测性能（R_p²= 0.888，RMSEP = 0.440，RPD = 2.756），能够实现对果柄结合力的有效预测。

关键词: 枸杞, 机械采收, 果柄, 结合力, 形态学, 影响因素, 预测模型

Abstract:

To analyse the intrinsic connection and influence law between the stalk binding force of wolfberry and its related physical property parameters，to discover the important fruit parameters，and to establish a prediction model of stalk binding force，with a view to providing a basis for improving the efficiency and accuracy of mechanical picking of wolfberry. Taking the ripening fruit of‘Ningqi 1’and ‘Ningqi 7’，the main cultivars of wolfberry in Qinghai Province，as the research object，the RGB colour model components of wolfberry fruits with different maturity were obtained by image processing methods，and different cultivars，maturity were analysed，temperature，humidity and the position of the fruit on the branch on the impact of the stalk binding force，as well as the correlation between the stalk parameters，fruit parameters and the stalk binding force，to construct a prediction model of the stalk binding force. The results showed that the maturity could be distinguished by the RGB colour model，but the overlapping part of the stalk binding force of different maturity levels was large，which could not be achieved for precise picking. The results of two-way ANOVA showed that both cultivar and maturity had a significant effect on the binding force；stalk parameters，fruit parameters and fruit hardness were significantly correlated with the stalk binding force；orthogonal tests showed that the position of the fruit on the branch，temperature and humidity at the time of picking did not have a significant effect on the binding force. The parameters of partial least squares regression（PLSR），random forest（RF）and back-propagation neural network（BPNN）modelling methods were optimised to establish a prediction model for fruit stalk binding，in which the BPNN showed excellent prediction performance（R_p² = 0.888，RMSEP = 0.440，RPD = 2.756），which was able to achieve effective prediction of fruit stalk binding.

Key words: wolfberry, mechanical harvesting, fruit stalk, binding force, morphology, influencing factors, predictive modelling

陈运, 刘全程, 姜鑫娜, 魏雨青, 王凡, 闫磊, 赵健, 曹兴达, 邢红. 枸杞果柄结合力影响因素分析及模型预测[J]. 园艺学报, 2024, 51(11): 2645-2656.

CHEN Yun, LIU Quancheng, JIANG Xinna, WEI Yuqing, WANG Fan, YAN Lei, ZHAO Jian, CAO Xingda, XING Hong. Analysis and Model Prediction of Factors Affecting the Binding Force of Fruit Stalks of Wolfberry[J]. Acta Horticulturae Sinica, 2024, 51(11): 2645-2656.

图/表 8

参考文献 20

[1]	Chen Jiawei, Zhou Deqiang, Cui Chenhao, Ren Zhijun, Zuo Wenjuan. 2023. Prediction model of farinograph characteristics of wheat flour based on near infrared spectroscopy. Spectroscopy and Spectral Analysis, 43 (10):3089-3097. (in Chinese)
	陈嘉伟, 周德强, 崔晨昊, 任志俊, 左文娟. 2023. 近红外光谱的小麦粉粉质特性预测模型研究. 光谱学与光谱分析, 43 (10):3089-3097.
[2]	Ding Haiyang, Dong Mingli, Liu Chenhua, Lu Xitian, Guo Chentong. 2024. Infrared and visible image fusion based on saliency adaptive weight map. Laser & Optoelectronics Progress, 61 (10):396-403. (in Chinese)
	丁海洋, 董明利, 刘陈华, 陆熙田, 郭晨彤. 2024. 基于显著性自适应权重图的红外可见光图像融合. 激光与光电子学进展, 61 (10):396-403.
[3]	Huang Guowei, Bi Xinsheng, Xiao Binbin, Qiu Bo. 2013. Analysis of influence factors about separation force between tomato fruit and stalk. Food & Machinery, 29 (3):187-190. (in Chinese)
	黄国伟, 毕新胜, 肖彬彬, 邱博. 2013. 番茄果实与茎秆分离力的影响因素分析. 食品与机械, 29 (3):187-190.
[4]	Jiang X, Hu X, Huang H, Tian J, Bu Y, Huang D, Luo H. 2021. Detecting total acid content quickly and accurately by combining hyperspectral imaging and an optimized algorithm method. Journal of Food Process Engineering, 44 (11):13844.
[5]	Lei X, Zhang X, Yang Q, Ouyang Q, Zhang C. 2016. Research progress on the tensile mechanical properties of plant roots. Journal of Zhejiang A & F University, 33 (4):703-711.
[6]	Li S, Song Q, Liu Y, Zeng T, Liu S, Jie D, Wei X. 2023. Hyperspectral imaging-based detection of soluble solids content of loquat from a small sample. Postharvest Biology and Technology, 204:112454.
[7]	Lin Xinzhang, Lei Jin, Ren Zichao, Zhang Jie, Liang Bingqin, Wang Bingpeng, Wang Weibing. 2023. Study on physical parameters and biological characteristics of Xinjiang seabuckthorn fruit after frozen storage. Xinjiang Agricultural Sciences, 60 (10):2479-2485. (in Chinese) doi: 10.6048/j.issn.1001-4330.2023.10.017
	林鑫章, 雷金, 任子超, 张杰, 梁炳钦, 王炳鹏, 王卫兵. 2023. 冻藏后沙棘果实的分离力与相关物理特性参数分析. 新疆农业科学, 60 (10):2479-2485. doi: 10.6048/j.issn.1001-4330.2023.10.017
[8]	Liu Ying, Fan Kaixin, Ma Chenwei. 2023. Theoretical analysis of vibration of self-propelled wolfberry picker. Forestry Machinery & Woodworking Equipment, 51 (10):39-43. (in Chinese)
	刘英, 范开欣, 马晨伟. 2023. 枸杞采收机振动采收机理分析. 林业机械与木工设备, 51 (10):39-43.
[9]	Ma F, Li L, Wang Y, Li P, Zhu C. 2020. Biomechanical properties of wolfberry plant organs. Engenharia Agrícola, 40 (2):162-176.
[10]	Ma Jiawei. 2017. Research status and prospect of the mechanized technology of picking wolfberry in China. Mechanical Research & Application, 30 (4):151-153,155. (in Chinese)
	马嘉伟. 2017. 我国枸杞机械化采摘技术的现状及发展趋势. 机械研究与应用, 30 (4):151-153,155.
[11]	Manjunatha K, Savadi S, Adiga J D, Balasubramanian D, Naik R, Muralidhara B M, Chethan C R. 2024. Evaluation of fruit detachment forces and related characteristics reveals differential fruit detachments at developmental stages and cultivar differences in cashew. Scientia Horticulturae, 331:113128.
[12]	Pang Chunmei, Li Yanping. 2023. Research progress on functional components in wolfberries. Food Industry,(4):116-118. (in Chinese)
	逄春梅, 李艳萍. 2023. 枸杞中功能性成分的研究进展. 食品界,(4):116-118.
[13]	Tie Jianshe. 2022. Exploration of the transformation of Qinghai wolfberry industry under the background of green economy. Food Research and Development, 43 (6):229-230. (in Chinese)
	铁建设. 2022. 绿色经济背景下青海枸杞产业转型探索. 食品研究与开发, 43 (6):229-230.
[14]	Wang Cungui. 2023. Breeding and cultivation management techniques of Lycium barbarum in Qinghai. Science and Technology of Qinghai Agriculture and Forestry,(4):91-94,98. (in Chinese)
	王存桂. 2023. 青海枸杞育苗与栽培管理技术. 青海农林科技,(4):91-94,98.
[15]	Wang S, Song W, Wang M, Wang J, Ding T, Zhou D, Ma S. 2022. Tests and analyses on physical and mechanical properties of fresh black fungus in picking season. PLoS ONE, 17 (10):e0275565.
[16]	Wu Di, Lin Xiaoyuan, Liu Kai,Ning Hang,Luo Wenhui,Zhao Guangfa. 2024. Recognition of antitussive components in Farfarae flos based on grey relational analysis and partial least squares regression. Journal of Central South University(Medical Science), 49 (3):435-446. (in Chinese) doi: 10.11817/j.issn.1672-7347.2024.230344 pmid: 38970518
	吴笛, 蔺晓源, 刘凯, 宁杭, 罗文汇, 赵光发. 2024. 基于灰色关联度分析和偏最小二乘回归辨识款冬花止咳活性成分. 中南大学学报(医学版), 49 (3):435-446. pmid: 38970518
[17]	Xu Shengrong, Zhang Enhe, Ma Ruili, Wang Qi, Liu Qinglin, Wu Rui. 2017. Hydraulic characteristics of Lycium barbarum L. seedlings under drought stress and re-watering conditions. Chinese Journal of Eco-Agriculture, 25 (8):1190-1197. (in Chinese)
	胥生荣, 张恩和, 马瑞丽, 王琦, 刘青林, 武睿. 2017. 干旱胁迫及复水对耐旱枸杞水力学特性的影响. 中国生态农业学报, 25 (8):1190-1197.
[18]	Zhang Ning, He Maoqi, Fang Wen. 2024. A prediction model for the activated roasting leaching composition of titanium-containing blast furnace slag based on PSO-GRNN. China Mining Magazine, 33 (S1):453-459,468. (in Chinese)
	张宁, 何茂琪, 方文. 2024. 基于PSO-GRNN的含钛高炉渣活化焙烧浸出成分预测模型. 中国矿业, 33 (S1):453-459,468.
[19]	Zhao J, Ma T, Inagaki T, Chen Q, Gao Z, Sun L, Cai H, Chen C, Li C, Zhang S, Satoru T, Chen J. 2021. Finite element method simulations and experiments of detachments of Lycium barbarum L. Forests, 12 (6):699.
[20]	Zhu C, Ding J, Zhang Z, Wang Z. 2022. Exploring the potential of UAV hyperspectral image for estimating soil salinity:effects of optimal band combination algorithm and random forest. Spectrochimica Acta Part A:Molecular and Biomolecular Spectroscopy, 279:121416.

成熟度 Maturity	品种 Cultivar	果柄长度/mm Fruit stalk length	果柄直径/mm Fruit stalk diameter		果实 Fruit
成熟度 Maturity	品种 Cultivar	果柄长度/mm Fruit stalk length	细端 Slender end	粗端 Thick end	横径/mm Diameter	纵径/mm Vertical diameter	质量/g Weight	硬度/N Hardness
青果 Green fruit	宁杞1号 Ningqi 1	15.34 ± 3.32	0.67 ± 0.11^*	1.59 ± 0.34^*	5.16 ± 0.53	12.36 ± 1.62^*	0.168 ± 0.049	11.57 ± 2.92^*
青果 Green fruit	宁杞7号 Ningqi 7	15.30 ± 2.60	0.62 ± 0.12	1.49 ± 0.30	5.24 ± 0.48	12.78 ± 1.54	0.163 ± 0.054	10.68 ± 2.83
黄果 Yellow fruit	宁杞1号 Ningqi 1	16.97 ± 3.81^*	0.69 ± 0.13^*	1.66 ± 0.41^*	7.15 ± 0.90^*	14.83 ± 2.33^*	0.333 ± 0.104^*	4.60 ± 1.22^*
黄果 Yellow fruit	宁杞7号 Ningqi 7	19.95 ± 3.58	0.64 ± 0.08	1.75 ± 0.35	7.49 ± 0.92	15.58 ± 2.16	0.386 ± 0.106	5.60 ± 1.46
红果 Red fruit	宁杞1号 Ningqi 1	19.42 ± 3.64^*	0.74 ± 0.12^*	1.86 ± 0.34	10.88 ± 1.17	20.71 ± 2.67^*	1.021 ± 0.283^*	2.66 ± 0.54
红果 Red fruit	宁杞7号 Ningqi 7	20.51 ± 3.37	0.67 ± 0.09	1.77 ± 0.33	10.32 ± 1.29	18.90 ± 2.26	0.813 ± 0.262	2.62 ± 0.64

成熟度 Maturity	品种 Cultivar	果柄长度/mm Fruit stalk length	果柄直径/mm Fruit stalk diameter		果实 Fruit
成熟度 Maturity	品种 Cultivar	果柄长度/mm Fruit stalk length	细端 Slender end	粗端 Thick end	横径/mm Diameter	纵径/mm Vertical diameter	质量/g Weight	硬度/N Hardness
青果 Green fruit	宁杞1号 Ningqi 1	15.34 ± 3.32	0.67 ± 0.11^*	1.59 ± 0.34^*	5.16 ± 0.53	12.36 ± 1.62^*	0.168 ± 0.049	11.57 ± 2.92^*
青果 Green fruit	宁杞7号 Ningqi 7	15.30 ± 2.60	0.62 ± 0.12	1.49 ± 0.30	5.24 ± 0.48	12.78 ± 1.54	0.163 ± 0.054	10.68 ± 2.83
黄果 Yellow fruit	宁杞1号 Ningqi 1	16.97 ± 3.81^*	0.69 ± 0.13^*	1.66 ± 0.41^*	7.15 ± 0.90^*	14.83 ± 2.33^*	0.333 ± 0.104^*	4.60 ± 1.22^*
黄果 Yellow fruit	宁杞7号 Ningqi 7	19.95 ± 3.58	0.64 ± 0.08	1.75 ± 0.35	7.49 ± 0.92	15.58 ± 2.16	0.386 ± 0.106	5.60 ± 1.46
红果 Red fruit	宁杞1号 Ningqi 1	19.42 ± 3.64^*	0.74 ± 0.12^*	1.86 ± 0.34	10.88 ± 1.17	20.71 ± 2.67^*	1.021 ± 0.283^*	2.66 ± 0.54
红果 Red fruit	宁杞7号 Ningqi 7	20.51 ± 3.37	0.67 ± 0.09	1.77 ± 0.33	10.32 ± 1.29	18.90 ± 2.26	0.813 ± 0.262	2.62 ± 0.64

代号 Designator	A 采摘温度/℃ Picking temperature	B 采摘湿度/% Picking humidity	C 枸杞在枝条上的位置 The site of the wolfberry on the branch	结合力/N Binding force
1	15 ~ 20	25 ~ 45	1（基部 Head section）	1.448
2	15 ~ 20	45 ~ 65	2（中部 Central section）	2.152
3	15 ~ 20	65 ~ 85	3（梢部 End section）	1.688
4	20 ~ 25	25 ~ 45	2	1.977
5	20 ~ 25	45 ~ 65	3	1.973
6	20 ~ 25	65 ~ 85	1	2.232
7	25 ~ 30	25 ~ 45	3	1.501
8	25 ~ 30	45 ~ 65	1	1.251
9	25 ~ 30	65 ~ 85	2	1.974

代号 Designator	A 采摘温度/℃ Picking temperature	B 采摘湿度/% Picking humidity	C 枸杞在枝条上的位置 The site of the wolfberry on the branch	结合力/N Binding force
1	15 ~ 20	25 ~ 45	1（基部 Head section）	1.448
2	15 ~ 20	45 ~ 65	2（中部 Central section）	2.152
3	15 ~ 20	65 ~ 85	3（梢部 End section）	1.688
4	20 ~ 25	25 ~ 45	2	1.977
5	20 ~ 25	45 ~ 65	3	1.973
6	20 ~ 25	65 ~ 85	1	2.232
7	25 ~ 30	25 ~ 45	3	1.501
8	25 ~ 30	45 ~ 65	1	1.251
9	25 ~ 30	65 ~ 85	2	1.974

来源 Source	类型 Ⅲ 的平方和 Sum of squares of Type Ⅲ	自由度 df	均方 Mean square	统计量 F	显著性P Significance
校正模型Correction model	0.864	6	0.144	3.637	0.231
节点增量Nodal increment	30.250	1	30.250	763.728	0.001
A	0.364	2	0.182	4.590	0.179
B	0.236	2	0.118	2.980	0.251
C	0.265	2	0.132	3.343	0.230
误差Error	0.079	2	0.040
总计Total	31.194	9
总修正Total corrected	0.944	8