一种基于深度学习的水稻种子分类方法

doi:10.3969/j.issn.1006-8082.2024.05.006

摘要/Abstract

摘要：

水稻是一种重要的粮食作物,其种子质量特别是纯度直接影响着水稻的产量和品质。传统的种子分类方法主要依靠人工视觉,效率低、误差率高。为了提高水稻种子分类的准确性和效率,建立了一种基于深度学习的水稻种子分类方法。本研究自行拍摄并构建了5个不同水稻品种含有80 000张水稻种子图片数据集,首先利用卷积神经网络(CNN)从水稻种子图像中提取特征,然后利用自建RiceFastNet模型对提取的特征分类。结果表明,该方法在识别外观相似度较高的不同品种水稻种子时,分类准确率超过96%,优于传统的种子分类方法。本研究提出的新方法在提高水稻种子检测准确率方面具有潜力和优势。

关键词: 水稻, 种子分类, 深度学习, 卷积神经网络

Abstract:

Rice is an important cereal crops, and its seed quality, especially its purity, directly affects the yield and quality of rice. Traditional seed classification methods mainly rely on artificial vision, which is inefficient and has a high error rate. In order to improve the accuracy and efficiency of rice seed classification, a method of rice seed classification based on deep learning was proposed. In this study, a dataset containing 80,000 rice seed images of five different rice varieties was self-captured and constructed. The method first used convolutional neural network(CNN) to extract features from rice seed images, and then used self-built RiceFastNet to classify the extracted features. The results showed that this method still achieved a classification accuracy of over 96% when identifying rice seeds of different varieties with high appearance similarity, which is superior to traditional seed classification methods. The new method established in this study has the potential and advantages in improving the accuracy of rice seed detection.

Key words: rice, seed classification, deep learning, convolutional neural network

中图分类号:

王晓飞, 刘维, 巫浩翔, 陈浩, 张丽婷, 潘朝阳, 何秀英. 一种基于深度学习的水稻种子分类方法[J]. 中国稻米, 2024, 30(5): 49-56.

WANG Xiaofei, LIU Wei, WU Haoxiang, CHEN Hao, Zhang Liting, PAN Zhaoyang, HE Xiuying. A Method of Rice Seed Classification Based on Deep Learning[J]. China Rice, 2024, 30(5): 49-56.

图/表 10

图1 种子矩阵

图2 5种水稻种子的随机图片实例

表1 研究中使用的硬件规格及分类器网络参数

图3 分类模型研究路线

表2 RiceFastNet网络结构及参数

图4 训练1个批次加载的种子图片实例

表3 不同学习率下RiceFastNet 分类准确率

图5 不同学习率下CNN网络训练准确率与验证准确率变化情况

表4 相同条件下不同神经网络对自建80 000张图片数据集分类表现

表5 相同条件下不同神经网络对公开数据集(75 000张图片)分类表现

参考文献 38

[1]	RUBAIYATH BIN RAHMAN A N M, ZHANG J H. Trends in rice research: 2030 and beyond[J]. Food and Energy Security, 2022, 12(2): 1-17.
[2]	LU Y, ZHANG X, ZENG N, et al. Image classification and identification for rice leaf diseases based on improved WOACW_SimpleNet[J]. Frontiers in Plant Science, 2022, 13: 1 008 819.
[3]	SHARMA G, ANAND V, KUMAR V. A suitable approach for classifying skin disease using deep convolutional neural network[C]// 2023 World Conference on Communication & Computing. New York: Institute of Electrical and Electronics Engineers, 2023.
[4]	SHARMA G, ANAND V, CHAUHAN R, et al. From paddy to pixel: An in-depth exploration into classifying diverse rice varieties leveraging advanced convolutional neural network architectures[C]// 2023 Global Conference on Information Technologies and Communications. New York: Institute of Electrical and Electronics Engineers, 2023.
[5]	NAIK N K, SETHY P K, DEVI A G, et al. Hybrid enhanced featured AlexNet for milled rice grain identification[J]. Ingénierie des Systèmes d'Information, 2023, 28(3): 663-668.
[6]	CHEN L Y, LI S B, BAI Q, et al. Review of image classification algorithms based on convolutional neural networks[J]. Remote Sensing, 2021, 13(22): 4 712.
[7]	FELIZARDO K B, PAREDES A M C, ARBOLEDA E R. Advancements in Artificial Intelligence (AI) for enhanced insights and automation in rice agriculture: A systematic review[J]. International Journal of Science and Research Archive, 2024, 11(1): 444-463.
[8]	PEKER N. Classification of rice varieties using a deep neural network model[C]// International Symposium on Intelligent Manufacturing and Service Systems. Singapore: Springer Nature Singapore, 2023: 510-521.
[9]	季长清, 高志勇, 秦静, 等. 基于卷积神经网络的图像分类算法综述[J]. 计算机应用, 2022, 42(4):1044-1 049.
[10]	阙昊天, 赵鑫, 黄敏, 等. 基于特征波段卷积神经网络的小麦种子品种分类研究[J]. 中国粮油学报, 2023, 38(5):141-148.
[11]	黄琼. 基于图像的水稻种子品种分类识别方法研究[D]. 南昌: 江西农业大学, 2022.
[12]	SINGH R, SHARMA N, GUPTA R. Rice type classification using proposed CNN model[C]// 2023 2nd International Conference on Vision Towards Emerging Trends in Communication and Networking Technologies (ViTECoN). New York: Institute of Electrical and Electronics Engineers, 2023.
[13]	MOHAMED ALI Z M G H. A Comparative study between traditional machine learning and deep learning models to classify rice types[D]. Dublin:National College of Ireland, 2023.
[14]	SINGH H, MADHOK H, JULKA N. Rice kernels classification with deep learning using a modified dataset mimicking real-world conditions[J]. Educational Research (IJMCER), 2023, 5(3): 15-23.
[15]	SETIAWAN R, OUMAROH H. Classification of rice grain varieties using ensemble learning and image analysis techniques[J]. Indonesian Journal of Data and Science, 2024, 5(1): 54-63.
[16]	叶文超, 罗水洋, 李金豪, 等. 近红外光谱与图像融合的杂交水稻种子分类方法研究[J]. 光谱学与光谱分析, 2023, 43(9):2935-2 941.
[17]	SOKUDLOR N, LALOON K, JUNSIRI C, et al. Enhancing milled rice qualitative classification with machine learning techniques using morphological features of binary images[J]. International Journal of Food Properties, 2023, 26(2), 2 978-2 992.
[18]	DIN N M U, ASSAD A, DAR R A, et al. RiceNet: A deep convolutional neural network approach for classification of rice varieties[J]. Expert Systems with Applications, 2024, 235: 121 214.
[19]	DOGAN M, TASPINAR Y S, CINAR I, et al. Dry bean cultivars classification using deep cnn features and salp swarm algorithm based extreme learning machine[J]. Computers and Electronics in Agriculture, 2023, 204: 107 575.
[20]	TASCI M, ISTANBULLU A, KOSUNALP S, et al. An efficient classification of rice variety with quantized neural networks[J]. Electronics, 2023, 12(10): 2 285.
[21]	RAJALAKSHMI R, FAIZAL S, SIVASANKARAN S, et al. RiceSeedNet: Rice seed variety identification using deep neural network[J]. Journal of Agriculture and Food Research, 2024, 16: 101 062.
[22]	OROZCO J, MANIAN V, ALFARO E, et al. Graph convolutional network using adaptive neighborhood Laplacian matrix for hyperspectral images with application to rice seed image classification[J]. Sensors, 2023, 23(7): 3 515.
[23]	阙昊天. 基于深度神经网络的小麦品种高光谱图像识别[D]. 无锡: 江南大学, 2023
[24]	WANG Y, SONG S Y. Variety identification of sweet maize seeds based on hyperspectral imaging combined with deep learning[J]. Infrared Physics & Technology, 2023, 130: 104 611.
[25]	LIU W H, ZENG S, WU G J, et al. Rice seed purity identification technology using hyperspectral image with LASSO logistic regression model[J]. Sensors, 2021, 21(13): 4 384.
[26]	MENG Y, YUAN W S, AKTILEK E U, et al. Fine hyperspectral classification of rice varieties based on self-attention mechanism[J]. Ecological Informatics, 2023, 75: 102 035.
[27]	KANG Z, FAN R, ZHAN C, et al. The rapid non-destructive differentiation of different varieties of rice by fluorescence hyperspectral technology combined with machine learning[J]. Molecules, 2024, 29(3): 682.
[28]	LI C, TAN Y, LIU C, et al. Spectral fusion based on hyperspectral imaging technology for discrimination of rice varieties[J]. Preprints 2024.
[29]	MAKMUANG S, TERDWONGWORAKUL A, VILAIVAN T, et al. Mapping hyperspectral NIR images using supervised self-organizing maps: Discrimination of weedy rice seeds[J]. Microchemical Journal, 2023, 190: 108 599.
[30]	孟颖. 基于双尺度高光谱信息的寒地水稻品种精细分类研究[D]. 哈尔滨: 东北农业大学, 2023.
[31]	徐元统. 基于高光谱的水稻品种鉴别系统设计与实现[D]. 哈尔滨: 黑龙江大学, 2019.
[32]	SINGH V, GOURISARIA M K, GM H, et al. Weed detection in soybean crop using deep neural network[J]. Pertanika Journal of Science & Technology, 2023, 31(1): 401-423.
[33]	PAN W T, SUN M L, YUN Y, et al. Identification method of wheat grain phenotype based on deep learning of ImCascade R-CNN[J]. Smart Agriculture, 2023, 5(3): 110-120.
[34]	YASAR A, GOLCUK A, SARI O F. Classification of bread wheat varieties with a combination of deep learning approach[J]. European Food Research and Technology, 2024, 250(1): 181-189.
[35]	ZHANG N, ZHANG E, LI F. A soybean classification method based on data balance and deep learning[J]. Applied Sciences, 2023, 13(11): 6 425.
[36]	SABLE A, SINGH P, KAUR A, et al. Quantifying soybean defects: A computational approach to seed classification using deep learning techniques[J]. Agronomy, 2024, 14(6): 1 098.
[37]	ÜNAL Z, AKTAS H. Classification of hazelnut kernels with deep learning[J]. Postharvest Biology and Technology, 2023, 197: 112 225.
[38]	BAYRAM F, YILDIZ M. Classification of some barley cultivars with deep convolutional neural networks[J]. Journal of Agricultural Sciences, 2023, 29(1): 262-271.