基于LSTM-TE模型的冬枣可溶性固形物含量高光谱估测

doi:10.3964/j.issn.1000-0593(2025)08-2326-09

摘要
参考文献
相关文章 (15)

全文: PDF (21323 KB)
输出: BibTeX | EndNote (RIS)

摘要：冬枣因其丰富的营养成分、香甜的口感和优良的风味，深受消费者的喜爱。随着消费者对水果品质要求的提高，可溶性固形物含量（SSC）作为衡量水果成熟度和口感质量的重要指标，已成为水果品质评估中的关键因素。因此，实现冬枣SSC的高效、无损预测具有重要的应用价值和实际意义。提出一种融合长短时记忆网络（LSTM）与Transformer编码器的LSTM-TE模型，旨在实现冬枣SSC的快速无损预测。采集900个冬枣样本的高光谱数据并测定其SSC值，结合多种光谱数据预处理方法[包括多元散射校正（MSC）、矢量归一化（VN）、Savitzky-Golay（SG）滤波、一阶导数（D1）和二阶导数（D2）等]对数据进行处理，通过PLSR、SVR、VGG16、ResNet18、LSTM五种模型系统比较了10种预处理组合的效果，确定最优预处理方案为MSC-SG-D1。在该预处理方法的基础上，进一步构建了PLSR、SVR、VGG16、ResNet18、LSTM和LSTM-TE的多模型对比体系，并对其在测试集上的性能进行了对比分析。实验结果表明，LSTM-TE 模型在测试集上的决定系数为0.959 8，均方根误差为1.269 0，较传统机器学习模型PLSR (R²_p=0.817 3)提升17.4%，较单一 LSTM模型（R²_p=0.865 2）提升10.9%。该模型通过LSTM的时序特征捕捉能力与Transformer编码器的全局依赖建模优势，有效挖掘了高光谱数据中的非线性特征关系。本研究为冬枣品质的在线检测与分级提供了新的技术方案，对高光谱技术在精准农业中的应用具有重要参考价值。

关键词：可溶性固形物含量(SSC)；高光谱；深度学习；LSTM-TE模型

Abstract：Winter jujube is favored by consumers for its rich nutritional content, sweet taste, and excellent flavor. With the increasing demands of consumers for fruit quality, Soluble Solids Content (SSC) has become a key factor in fruit quality evaluation, serving as an important indicator for measuring fruit ripeness and taste quality. Therefore, efficient and non-destructive prediction of winter jujube SSC has significant practical value and importance. This paper proposes an LSTM-TE model, which integrates Long Short-Term Memory (LSTM) networks with a Transformer Encoder, aiming to achieve rapid and non-destructive prediction of winter jujube SSC. By collecting hyperspectral data from 900 winter jujube samples and determining their SSC values, multiple spectral data preprocessing methods (including Multivariate Scatter Correction (MSC), Vector Normalization (VN), Savitzky-Golay (SG) filtering, first derivative (D1), and second derivative (D2)) were applied to process the data. The effects of 10 preprocessing combinations were compared through five models: PLSR, SVR, VGG16, ResNet18, and LSTM, to determine the optimal preprocessing scheme, MSC-SG-D1. Based on this preprocessing method, a multi-model comparison system was further constructed, including PLSR, SVR, VGG16, ResNet18, LSTM, and LSTM-TE, and their performance was analyzed on the test set. Experimental results showed that the LSTM-TE model achieved a coefficient of determination of 0.959 8 and a root mean square error (RMSE) of 1.269 0 on the test set, an improvement of 17.4% compared to the traditional machine learning modelPLSR (R²_p=0.817 3) and 10.9% compared to the single LSTM model (R²_p=0.865 2). This model effectively explored the nonlinear feature relationships in hyperspectral data by capturing temporal features through LSTM and leveraging the global dependency modeling advantages of the Transformer encoder. This study provides a new technical solution for the online detection and grading of winter jujube quality, offering important reference value for applying hyperspectral technology in precision agriculture.

Key words：Soluble Solids Content (SSC);Hyperspectral;Deep learning;LSTM-TE model

收稿日期: 2025-01-13 修订日期: 2025-04-08

中图分类号:

S237

基金资助: 中央引导地方科技发展资金项目(246Z0309G)，河北省高等教育科研项目(BJ2025097)，河北省农业科技成果转化资金项目(202460104030028)，河北省博士在读研究生创新能力培养项目(CXZZBS2024073)资助

通讯作者: 宋宇斐 E-mail: songyufei0311@163.com

作者简介: 刘傲然, 2001年生, 河北师范大学计算机与网络空间安全学院硕士研究生 e-mail: 347354547@qq.com

引用本文:

刘傲然，孟惜，刘智国，宋宇斐，赵雪曼，智丹宁. 基于LSTM-TE模型的冬枣可溶性固形物含量高光谱估测[J]. 光谱学与光谱分析, 2025, 45(08): 2326-2334.
LIU Ao-ran, MENG Xi, LIU Zhi-guo, SONG Yu-fei, ZHAO Xue-man, ZHI Dan-ning. Hyperspectral Estimation of Soluble Solids Content in Winter Jujube Based on LSTM-TE Model. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2025, 45(08): 2326-2334.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2025)08-2326-09 或 https://www.gpxygpfx.com/CN/Y2025/V45/I08/2326

[1] LIU Tian-zhen, YUAN Ying-chun, TENG Gui-fa, et al(刘天真，苑迎春，滕桂法, 等). Jiangsu Agricultural Sciences(江苏农业科学), 2024, 52(1): 163.
[2] LIU Zi-han, LI Ming, ZHAO Zhi-yao, et al(刘子涵, 李明, 赵峙尧, 等). Journal of Fruit Science(果树学报), 2024, 41(12): 2606.
[3] Zhang D Y, Xu Y F, Huang W Q, et al. Infrared Physics &Technology, 2019, 98: 297.
[4] XIANG Zhi-yong, MIAO Yu-bin(相志勇, 苗玉彬). Journal of Chinese Agricultural Mechanization(中国农机化学报), 2024, 45(11): 139.
[5] Guo Z M, Wang M M, Agyekum A A, et al. Journal of Food Engineering, 2020, 279: 109955.
[6] Kim M J, Yu W H, Song D J, et al. Sensors, 2024, 24: 1512.
[7] Yang B, Gao Y, Yan Q, et al. Sensors, 2020, 20: 5021.
[8] Yang J, Xu J, Zhang X, et al. Analytica Chimica Acta, 2019, 1081: 6.
[9] Furbank R T, Silva-Perez V, Evans J R, et al. Plant Methods, 2021, 17(1): 108.
[10] LeCun Y, Bengio Y，Hinton G. Nature, 2015, 521: 436.
[11] Sun Z Z, Tian H, Hu D, et al. Food Chemistry, 2025, 464: 141488.
[12] Yu X J, Lu H D, Wu D. Postharvest Biology and Technology, 2018, 141: 39.
[13] Qi H, Li H, Chen L, et al. Foods, 2024, 13: 251.
[14] PENG Run-dong, LI Yao-xiang, CHEN Ya, et al(彭润东, 李耀翔, 陈雅, 等). Journal of Central South University of Forestry & Technology(中南林业科技大学学报), 2024, 44(3): 179.
[15] MENG Heng, ZHANG Tao, WANG Jin, et al(孟衡, 张涛, 王金, 等). Power System Technology(电网技术), 2024, 48(10): 4297.
[16] WANG Hao-yu, WEI Zi-yuan, YANG Yong-xia, et al(王浩宇, 魏子渊, 杨永霞, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2024, 44(8): 2208.
[17] Yang Z F, Xiao H, Zhang L, et al. Analytical Methods, 2019, 11(31): 3936.
[18] LI Shan-you, WANG Bo-rui, LU Jian-qi, et al(李山有, 王博睿, 卢建旗, 等). Chinese Journal of Geophysics(地球物理学报), 2024, 67(2): 587.