基于遗传算法BP神经网络芒果成熟度判别模型的构建与优化

doi:10.12190/j.issn.2096-1197.2023.05.11

摘要/Abstract

摘要：

【目的】基于遗传算法，构建及优化BP神经网络芒果成熟度判别模型。【方法】以小台芒为试验材料，测定不同成熟度芒果的硬度、糖度以及拍摄RGB图像，分析芒果硬度、糖度及图像特征变量（R、G、B、H、S、V分量）与芒果成熟度之间的相关性；利用遗传算法得到最优网络权值及阈值，构建BP神经网络芒果成熟度判别模型，并对其进行优化。【结果】随着芒果成熟度的增加，糖度表现出较明显的增长趋势、硬度表现出较明显的下降趋势；图像的特征分量（R、G、B、H、S分量）在不同成熟度阶段呈显著性差异（P<0.05）；基于遗传算法优化BP神经网络的芒果成熟度判别模型经过预测后，分类准确率为96.3%。【结论】基于遗传算法优化BP神经网络的芒果成熟度判别模型能较为准确地判断不同成熟度的芒果。

关键词: 芒果, 成熟度, 遗传算法, BP神经网络, 判别模型

Abstract:

【Objective】To construct and optimize BP neural network mango maturity discrimination model based on genetic algorithm.【Methods】Using Xiaotai Mango as the experimental material，the firmness and sugar content of mangoes with different maturity levels were measured，and RGB images captured. The correlation between mango maturity and mango firmness，sugar content，and image feature variables（R，G，B，H，S，V components） were analyzed. BP neural network mango maturity discrimination model was constructed and optimized using optimal network weights and thresholds obtained by genetic algorithm.【Results】With the increase of mango maturity，sugar content showed clear increasing trend，while firmness showed decreasing trend. Image feature variables （R，G，B，H，S components） had significant differences （P<0.05） at different maturity stages. The BP neural network mango maturity discrimination model optimized by genetic algorithm had an classification accuracy of 96.3%.【Conclusion】The BP neural network mango maturity discrimination model optimized by genetic algorithm could accurately identify mangoes with different maturity levels.

Key words: Mango, Maturity, Genetic algorithm, BP neural network, Discrimination model

中图分类号:

TP391.41

房芯如, 邢靖萱, 索郎措, 江亿平. 基于遗传算法BP神经网络芒果成熟度判别模型的构建与优化[J]. 北方农业学报, 2023, 51(5): 103-113.

FANG Xinru, XING Jingxuan, Suolangcuo , JIANG Yiping. Construction and optimization of BP neural network mango maturity discrimination model based on genetic algorithm[J]. Journal of Northern Agriculture, 2023, 51(5): 103-113.

图/表 17

图1

图2

图3

图4

表1

图5

图6

图7

图8

表2

表3

图9

表4

图10

表5

图11

表6

参考文献 20

[1]	张业歆, 刘家粮, 潘永贵, 等. 氯吡苯脲浸泡处理延缓采后芒果成熟和软化的生理机制[J]. 食品科学, 2021, 42(19):234-241.
[2]	张明洁, 张京红, 张亚杰, 等. 我国芒果种植气候适宜性区划[J]. 江苏农业科学, 2022, 50(2):124-130.
[3]	童悦, 徐祥彬, 史学群, 等. 草珊瑚乙醇提取物的抗氧化性及其对芒果采后主要病原菌的抑菌活性研究[J]. 保鲜与加工, 2022, 22(4):19-26.
[4]	普红梅, 王海丹, 杨芳, 等. 不同预处理的芒果保鲜效果比较[J]. 现代食品科技, 2022, 38(1):197-205.
[5]	卫赛超, 谢晶. 不同包装方式对芒果低温模拟运输贮藏中品质及代谢的影响[J]. 食品科学, 2022, 43(5):227-234.
[6]	LI L, LI C B, SUN J A, et al. The effects of 1-methylcyclopropene in the regulation of antioxidative system and softening of mango fruit during storage[J]. Journal of Food Quality, 2020, 2020(2):1-11.
[7]	MISHRA P, WOLTERING E, HARCHIOUI N E. Improved prediction of “Kent” mango firmness during ripening by near-infrared spectroscopy supported by interval partial least square regression[J]. Infrared Physics & Technology, 2020, 110:103459.
[8]	SHEU S C, TSOU P C, LIEN Y Y, et al. Rapid and specific detection of mango mangifera indica in processed food using an isothermal nucleic acid amplification assay[J]. European Food Research and Technology, 2020, 246(4):759-766. doi: 10.1007/s00217-020-03440-z
[9]	LIANG M H, SU X G, YANG Z F, et al. Carotenoid composition and expression of carotenogenic genes in the peel and pulp of commercial mango fruit cultivars[J]. Scientia Horticulturae, 2020, 263:109072. doi: 10.1016/j.scienta.2019.109072
[10]	BAI J W, ZHANG L, CAI J R, et al. Laser light backscattering image to predict moisture content of mango slices with different ripeness during drying process[J]. Journal of Food Process Engineering, 2021, 44(12):13900.
[11]	THONG N D, THINH N T, CONG H T. Mango classification system uses image processing technology and artificial intelligence[J]. International Conference on System Science and Engineering, 2019, 11(5):1169.
[12]	SHAH S S A, ZEB A, QURESHI W S, et al. Towards fruit maturity estimation using NIR spectroscopy[J]. Infrared Physics & Technology, 2020, 111:103479.
[13]	XU D H, WANG H W, JI H W, et al. Hyperspectral imaging for evaluating impact damage to mango according to changes in quality attributes[J]. Sensors, 2018, 18(11):3920. doi: 10.3390/s18113920
[14]	李水梅, 李洪艳, 余静, 等. 嗅觉可视化技术在食品品质分析中的应用[J]. 分析化学, 2022, 50(12):1796-1803.
[15]	PAMUNGKAS W A, BINTORO N. Evaluation of watermelon ripeness using self-developed ripening detector[J]. IOP Conference Series:Earth and Environmental Science, 2021, 653(1):12020. doi: 10.1088/1755-1315/653/1/012020
[16]	黄毅英. 基于视觉技术的芒果运输过程监控系统设计[J]. 农机化研究, 2020, 42(4):81-85.
[17]	向阳, 林洁雯, 李亚军, 等. 芒果双面成熟度在线检测分级系统[J]. 农业工程学报, 2019, 35(10):259-266.
[18]	辛华健, 韦胜超. 基于计算机视觉的芒果分级系统研究[J]. 计算机产品与流通, 2019(6):94.
[19]	贺少鹏, 邱媛媛. 水果品质的无损检测技术应用研究[J]. 中国食品, 2022(15):122-124.
[20]	陈文悦, 何军, 朱立学, 等. 基于迁移学习的芒果成熟度分类算法研究[J]. 仲恺农业工程学院学报, 2022, 35(4):56-61.