Rice Phosphorus Nutrition Diagnosis Method Based on Improved ShuffleNet V2

doi:10.3969/j.issn.1006-8082.2025.02.004

Abstract

Abstract:

In order to more accurately diagnose rice phosphorus nutrition and help rice growth, an improved ShuffleNet V2 rice phosphorus nutrition diagnosis method is proposed. This method improves the model by introducing the ECA attention mechanism into the ShuffleNet V2 network model, and selects the pooling method of Attention Pooling to optimize model training. The transfer learning strategy is adopted to transfer the pre-trained weights on the ImageNet large dataset to the improved ShuffleNet V2 network model to train the rice leaf dataset and construct a rice phosphorus nutrition diagnosis model. The experimental results showed that the improved ShuffleNet V2 network model had higher accuracy, precision, recall and F1 value than other comparative network structure models in the rice tillering stage and rice jointing stage, and the training parameters were small, the training was more stable, and the convergence speed was faster. It proved that the improved ShuffleNet V2 rice phosphorus nutrition diagnosis model had good diagnostic recognition ability, which was helpful to adopt scientific and effective fertilization strategies under big data. At the same time, the improved ShuffleNet V2 network model had also achieved remarkable results on the Plant Village public dataset, which further verified the effectiveness and generalization ability of the improved model.

Key words: rice, phosphorus nutrition diagnosis, ECA attention mechanism, attention pooling, transfer learning

摘要：

为了更精确地诊断水稻的磷素营养状况，进而促进水稻的健康生长，我们提出了一种基于改进ShuffleNet V2的水稻磷素营养诊断方法。该方法的核心是在ShuffleNet V2网络模型引入ECA（Efficient Channel Attention）注意力机制，以优化原有模型。同时，选用注意力池化（Attention Pooling）技术来进一步提升模型训练的效果。在模型训练过程中，采用了迁移学习策略，即将在大规模数据集ImageNet上预训练的权值迁移至经过改进的ShuffleNet V2网络模型中，并利用这些权值对水稻叶片数据集进行训练，从而构建出水稻磷素营养诊断模型。结果显示，相比其他对比的网络结构模型，改进后的ShuffleNet V2网络模型在水稻分蘖期和拔节期的准确率、精确率、召回率以及F1值均表现出更高的水平，且该模型训练参数少、训练过程更稳定、收敛速度更快，证明改进后的ShuffleNet V2水稻磷素营养诊断模型具备了出色的诊断识别能力，能够为大数据背景下的科学、有效施肥策略提供有力支持。改进的ShuffleNet V2网络模型在Plant Village公共数据集上也取得显著效果，验证了其有效性和良好的泛化能力。

关键词: 水稻, 磷素营养诊断, ECA注意力机制, 注意力池化, 迁移学习

CLC Number:

HUANG Shumei, YANG Hongyun, KONG Jie, WU Zheng. Rice Phosphorus Nutrition Diagnosis Method Based on Improved ShuffleNet V2[J]. China Rice, 2025, 31(2): 20-28.

黄淑梅, 杨红云, 孔杰, 吴正. 基于改进ShuffleNet V2的水稻磷素营养诊断方法[J]. 中国稻米, 2025, 31(2): 20-28.

Figures/Tables 12

References 27

[1]	刘璐, 许传山, 孙雅菲, 等. 缺磷响应基因OsSQD2.1对水稻生长发育的影响[J]. 中国水稻科学, 2020, 34(3):237-244.
[2]	李继云, 刘秀娣, 周伟, 等. 有效利用土壤营养元素的作物育种新技术研究[J]. 中国科学(化学), 1995, 25(1):41-48
[3]	彭廷, 熊加豹, 贾晓迪, 等. 有机肥类型及其施肥配比对稻田培肥效应和经济效益的影响[J]. 河南农业大学学报, 2018, 52(1):22-29.
[4]	陈健晓, 王小娟, 屠乃美, 等. 磷水平对不同香稻品种苗期形态生理的影响[J]. 华北农学报, 2020, 35(2):117-125.
[5]	孙志伟, 徐月梅, 许荣越, 等. 水稻低磷胁迫响应及其调控机制的研究进展[J]. 核农学报, 2023, 37(8):1 562-1 570.
[6]	付雪蛟, 马畅, 付立东. 施磷量对滨海盐碱稻区水稻生长发育及产量的影响[J]. 东北农业科学, 2021, 45(5):53-59.
[7]	程明芳, 何萍, 金继运. 我国主要作物磷肥利用率的研究进展[J]. 作物杂志, 2010(1):12-14.
[8]	张锦滨, 王晓云, 孟圆, 等. 不同磷肥用量对水稻产量效益、磷肥利用率及土壤养分的影响[J]. 中国农学通报, 2021, 37(32):96-101.
[9]	冯凡, 孙虎威, 赵全志. 缺磷调控水稻根毛生长发育的机制[J]. 河南农业大学学报, 2018, 52(4):519-525.
[10]	孙洪仁, 张吉萍, 冮丽华, 等. 我国水稻土壤有效磷和速效钾丰缺指标与适宜磷钾施用量研究[J]. 中国稻米, 2018, 24(5):1-10.
[11]	陈晓凯, 李粉玲, 王玉娜, 等, 无人机高光谱遥感估算冬小麦叶面积指数[J]. 农业工程学报, 2020, 36(22):40-49.
[12]	李岚涛, 张萌, 任涛, 等. 应用数字图像技术进行水稻氮素营养诊断[J]. 植物营养与肥料学报, 2015, 21(1):259-268.
[13]	周敬明, 蔡万云. 水稻叶色卡的研制及其配套技术的研究[J]. 农业现代化研究, 1991, 12(6):38-42.
[14]	张金恒, 王珂, 王人潮. 叶绿素计SPAD-502在水稻氮素营养诊断中的应用[J]. 西北农林科技大学学报(自然科学版), 2003, 31(2):177-180.
[15]	杨红云, 周琼, 杨珺, 等. 水稻氮素营养诊断方法研究进展[J]. 中国稻米, 2020, 26(2):5-8+13.
[16]	罗建军. 基于计算机视觉技术与高光谱技术的水稻氮素营养诊断研究[D]. 南昌: 江西农业大学, 2022.
[17]	林芬芳, 丁晓东, 付志鹏, 等. 基于互信息理论的水稻磷素营养高光谱诊断[J]. 光谱学与光谱分析杂志, 2009, 29(9):2 467-2 470.
[18]	张林朋, 杨红云, 钱政, 等. 基于改进的VGG16网络和迁移学习的水稻氮素营养诊断[J]. 中国农业大学学报, 2023, 28(6):219-229.
[19]	钱政, 杨孙哲, 张国卿, 等. 基于卷积神经网络的水稻氮素营养诊断[J]. 中国农业科技导报, 2023, 25(9):113-121.
[20]	班松涛, 田明璐, 常庆瑞, 等. 基于无人机高光谱影像的水稻叶片磷素含量估算[J]. 农业机械学报, 2021, 52(8):163-171.
[21]	SHARMA M, NATH K, SHARMA R K, et al. Ensemble averaging of transfer learning models for identification of nutritional deficiency in rice plant[J]. Electronics, 2022, 11(1): 148.
[22]	ESPEJO-GARCIA B, MALOUNAS I, MYLONAS N, et al. Using EfficientNet and transfer learning for image-based diagnosis of nutrient deficiencies[J]. Computers and Electronics in Agriculture, 2022, 196: 106 868.
[23]	WANG C, YE Y, TIAN Y, et al. Classification of nutrient deficiency in rice based on CNN model with Reinforcement Learning augmentation[C]// 2021 International Symposium on Artificial Intelligence and its Application on Media (ISAIAM). IEEE, 2021: 107-111.
[24]	RICE L, WONG E, KOLTER Z. Overfitting in adversarially robust deep learning[C]// International Conference on Machine Learning. PMLR, 2020: 8 093-8 104.
[25]	MA N, ZHANG X, ZHENG H T, et al. Shufflenet v2: Practical guidelines for efficient cnn architecture design[C]// Proceedings of the European Conference on Computer Vision (ECCV), 2018: 116-131.
[26]	刘春花, 谢富, 宋岩, 等. 植物生长及生理生化特性对低磷胁迫的响应研究进展[J]. 江西农业学报, 2020, 32(2):71-75.
[27]	唐煜杰, 苗欢, 冯茜, 等. 合理施磷量提高水稻拔节期对大气增温的抵御能力[J]. 植物营养与肥料学报, 2023, 29(8):1 461-1 470.