基于掩码自编码的农作物病虫害分类方法

doi:10.16180/j.cnki.issn1007-7820.2024.10.004

摘要/Abstract

摘要：

作物病虫害使农业生产遭受损失,但仅依靠人工调查难以满足田间需求。基于机器视觉可实现病虫害自动分类,为农业精准高效生产提供保障。然而现有利用深度学习的方法易受刚性卷积感受野影响,数据增强手段低效且样本量匮乏。针对这些问题,文中提出一种基于掩码自编码学习范式的农业经济作物病虫害分类方法,来弥补现有技术在识别准确率方面的不足。通过对作物图像随机掩蔽、特征提取和依高维映射的全局重建,所提算法能充分挖掘输入的高阶语义隐式表征,建模同一图像内远距离上下文关系,从而训练鲁棒性更强的模型。通过相对总变分变换消除了高频噪声对预训练特征提取过程的干扰。所提方法与当前基于主流卷积网络的方法的对比结果表明,所提方法可显著提升现有方法的性能,准确率由基于ResNet50基准网络的90.48%提升至95.24%。

关键词: 机器视觉, 深度学习, 农业经济作物, 病虫害检测, 掩码自编码, 相对总变分, 神经网络, 卷积感受野

Abstract:

Crop diseases and insect pests cause a large amount of economic losses in agricultural production activities, and it is difficult to meet field production requirements of the current society if only relying on manual surveys by agronomist staffs. Applications of the machine vision technology can realize the automatic classification and detection of crop diseases and insect pests, and provide the guarantee for accurate and efficient agricultural productions. However, existing detection methods based on the deep learning framework and convolutional neural networks are constrained by factors such as rigid convolutional receptive field, inefficient data enhancement operator, and small sample size. In order to make up for the above shortcomings of existing detection technologies in term of recognition accuracy, a method for the classification of agricultural economic crop diseases and insect pests based on the masked autoencoding learning paradigm is proposed in this study. Through local random content masking, semantic feature extraction, and global context reconstruction of high-dimensional mapping of input crop images, the proposed algorithm can fully mine implicit representations of high-level semantics of images and model the long-distance contextual relationship in the same map, so as to train a more robust model with less data samples. Moreover, the model eliminates the interference of the high-frequency noise on the pre-training feature extraction processing by the relative total variational transformation. The results of comparison between the proposed method and current methods based on mainstream convolutional networks show that the proposed method can significantly improve the performance of existing methods, and the accuracy rate is improved from 90.48% to 95.24% based on ResNet50 benchmark network.

Key words: machine vision, deep learning, agricultural cash crop, diseases and pests detection, masked autoencoding, relative total variation, neural network, convolutional receptive field

中图分类号:

TP391

鞠萍, 宋岩, 张英杰, 徐一夫, 邵杭. 基于掩码自编码的农作物病虫害分类方法[J]. 电子科技, 2024, 37(10): 23-29.

JU Ping, SONG Yan, ZHANG Yingjie, XU Yifu, SHAO Hang. Research on Agricultural Crop Diseases and Pests Classification Based on Masked Autoencoding[J]. Electronic Science and Technology, 2024, 37(10): 23-29.

图/表 9

图1

图2

图3

表1

图4

表2

图5

表3

表4

参考文献 35

[1]	温艳兰, 陈友鹏, 王克强, 等. 基于机器视觉的病虫害检测综述[J]. 中国粮油学报, 2022, 37(10):271-279.
	Wen Yanlan, Chen Youpeng, Wang Keqiang, et al. An overview of plant diseases and insect pests detection based on machine vision[J]. Journal of the Chinese Cereals and Oils Association, 2022, 37(10):271-279.
[2]	徐天成, 吴敏, 贺冬仙, 等. 机器视觉在农业工程中的应用[J]. 农业工程, 2021, 11(8):40-48.
	Xu Tiancheng, Wu Min, He Dongxian, et al. Application of machine vision in agricultural engineering[J]. Agricultural Engineering, 2021, 11(8):40-48.
[3]	Loti N N A, Noor M R M, Chang S W. Integrated analysis of machine learning and deep learning in chilipest and disease identification[J]. Journal of the Science of Food and Agriculture, 2021, 101(9):3582-3594.
[4]	邵杭, 王永雄. 基于并行对抗与多条件融合的生成式高分辨率图像修复[J]. 模式识别与人工智能, 2020, 33(4):363-374. doi: 10.16451/j.cnki.issn1003-6059.202004009
	Shao Hang, Wang Yongxiong. Generative high-resolution image inpainting with parallel adversarial network and multicondition fusion[J]. Pattern Recognition and Artificial Intelligence, 2020, 33(4):363-374.
[5]	邵杭, 王永雄, 秦宇龙. 基于深度学习与显著性的数字图像构图优化[J]. 电子科技, 2021, 34(3):36-42.
	Shao Hang, Wang Yongxiong, Qin Yulong. Digital image composition optimization based on salient feature algorithm[J]. Electronic Science and Technology, 2021, 34(3):36-42.
[6]	赵轩, 周凡, 余汉成. 基于改进特征提取及融合模块的YOLO-v3模型[J]. 电子科技, 2022, 35(7):40-45.
	Zhao Xuan, Zhou Fan, Yu Hancheng. Improved YOLO-v3 model based on new feature extraction and fusion module[J]. Electronic Science and Technology, 2022, 35(7):363-374.
[7]	Sinha A, Shekhawat R S. Review of image processing approaches for detecting plant diseases[J]. IET Image Processing, 2020, 14(8):1427-1439.
[8]	Liu J, Wang X. Plant diseases and pests detection based on deep learning:A review[J]. Plant Methods, 2021, 17(22):1-18.
[9]	于润润, 姜晓燕, 朱凯赢, 等. 基于上下文注意力机制的实时语义分割[J]. 电子科技, 2022, 35(12):57-63.
	Yu Runrun, Jiang Xiaoyan, Zhu Kaiying, et al. Real-time image semantic segmentation based on contextual attention mechanism[J]. Electronic Science and Technology, 2022, 35(12):57-63.
[10]	Lee S H, Chan C S, Mayo S J, et al. How deep learning extracts and learns leaf features for plant classification[J]. Pattern Recognition, 2017, 71(11):1-13.
[11]	Zheng Y Y, Kong J L, Jin X B, et al. Cropdeep:The crop vision dataset for deep-learning-based classification and detection in precision agriculture[J]. Sensors, 2019, 19(5):1058-1078.
[12]	Fuentes A, Yoon S, Park D S. Deep learning-based techniques for plant diseases recognition in real-field scenarios[C]. Auckland: International Conference on Advanced Concepts for Intelligent Vision Systems,2020:3-14.
[13]	He K, Chen X, Xie S, et al. Masked autoencoders are scalable vision learners[C]. New Orleans: IEEE/CVF Conference on Computer Vision and Pattern Recognition,2022:16000-16009.
[14]	Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]. Long Beach: Neural Information Processing Systems,2017:1-11.
[15]	Dong X, Bao J, Zhang T, et al. Bootstrapped masked autoencoders for vision bert pretraining[C]. Tel Aviv: European Conference on Computer Vision,2022:247-264.
[16]	Xiao J, Bai Y, Yuille A, et al. Delving into masked autoencoders for multilabel thorax disease classification[C]. Waikoloa: IEEE/CVF Winter Conference on Applications of Computer Vision,2023:3588-3600.
[17]	Wang W, Jia Y, Wang Q, et al. An image enhancement algorithm based on fractional-order phase stretch transform and relative total variation[J]. Computational Intelligence and Neuroscience, 2021, 13(1):1-10.
[18]	Thakur P S, Khanna P, Sheorey T, et al. Trends in vision-based machine learning techniques for plant disease identification:A systematic review[J]. Expert Systems with Applications, 2022, 11(1):1-11.
[19]	Fuentes A, Lee J, Lee Y, et al. Anomaly detection of plant diseases and insects using convolutional neural networks[C]. Jeju Island: International Society for Ecological Modelling Global Conference,2017:1-2.
[20]	Hasan M J, Mahbub S, Alom M S, et al. Rice disease identification and classification by integrating support vector machine with deep convolutional neural network[C]. Dhaka: International Conference on Advances in Science,Engineering and Robotics Technology,2019:1-6.
[21]	Thenmozhi K, Reddy U S. Crop pest classification based on deep convolutional neural network and transfer learning[J]. Computers and Electronics in Agriculture, 2019, 164(1):1-11.
[22]	Nagasubramanian K, Jones S, Singh A K, et al. Plant disease identification using explainable 3D deep learning on hyperspectral images[J]. Plant Methods, 2019, 15(1):1-10.
[23]	Fang T, Chen P, Zhang J, et al. Crop leaf disease grade identification based on an improved convolutional neural network[J]. Journal of Electronic Imaging, 2020, 29(1):13-17.
[24]	Wen J, Shi Y, Zhou X, et al. Crop disease classification on inadequate low-resolution target images[J]. Sensors, 2020, 20(16):4601-4017.
[25]	Priyadharshini R A, Arivazhagan S, Arun M, et al. Maize leaf disease classification using deep convolutional neural networks[J]. Neural Computing and Applications, 2019, 31(12):8887-8895.
[26]	Atila U, Ucar M, Akyol K, et al. Plant leaf disease classification using efficientnet deep learning model[J]. Ecological Informatics, 2021, 61(1):1-13.
[27]	Zhou G, Zhang W, Chen A, et al. Rapid detection of rice disease based on FCM-KM and faster R-CNN fusion[J]. IEEE Access, 2019, 7(1):143190-143206.
[28]	Tian Y, Yang G, Wang Z, et al. Detection of apple lesions in orchards based on deep learning methods of cyclegan and YOLOv3-dense[J]. Journal of Sensors, 2019, 1(1):1-13.
[29]	Akshai K P, Anitha J. Plant disease classification using deep learning[C]. Coimbatore: International Conference on Signal Processing and Communication,2021:407-411.
[30]	Amin M F, Othman Z, Ahmad S S S, et al. Analysis onthe impact of imagenet preprocessing image mode using VGG19 pretrained model in plant disease classification[J]. Mechanical Engineering Research Day, 2022, 1(1):154-155.
[31]	He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]. Las Vegas: IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[32]	Kumar R, Singh D, Chug A, et al. Evaluation of deep learning based ResNet-50 for plant disease classification with stability analysis[C]. Madurai: International Conference on Intelligent Computing and Control Systems,2022:1280-1287.
[33]	Shao H, Wang Y. Generative image inpainting with salient prior and relative total variation[J]. Journal of Visual Communication and Image Representation, 2021, 79(1):1-9.
[34]	余小东, 杨孟辑, 张海清, 等. 基于迁移学习的农作物病虫害检测方法研究与应用[J]. 农业机械学报, 2020, 51(10):252-258.
	Yu Xiaodong, Yang Mengji, Zhang Haiqing, et al. Research and application of crop diseases detection method based on transfer learning[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(10):252-258.
[35]	于明, 李若曦, 阎刚, 等. 基于颜色掩膜网络和自注意力机制的叶片病害识别方法[J]. 农业机械学报, 2022, 53(8):337-344.
	Yu Ming, Li Ruoxi, Yan Gang, et al. Crop diseases recognition method via fusion color mask and self-attention mechanism[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(8):337-344.

作物	病虫害类别	训练样本	验证样本
草莓	健康	242	35
草莓	叶枯病	775	110
番茄	健康	1 208	173
番茄	白粉病、疮痂病、早疫病、晚疫病、叶霉病、斑点病、斑枯病、红蜘蛛损伤、黄化曲叶病毒病、花叶病毒病	10 268	1 466
柑桔	健康	367	52
柑桔	黄龙病	3 627	531
辣椒	健康	1 025	147
辣椒	疮痂病	664	94
马铃薯	健康	1 430	204
马铃薯	早疫病、晚疫病	1 410	202
苹果	健康	1 185	169
苹果	黑星病、灰斑病、雪松锈病	972	139
葡萄	健康	294	42
葡萄	黑腐病、轮斑病、褐斑病	2 456	352
桃子	健康	251	36
桃子	疮痂病	1 627	232
樱桃	健康	598	85
樱桃	白粉病	226	30
玉米	健康	376	54
玉米	灰斑病、锈病、叶斑病、花叶病毒病	2 717	387

网络架构	分类数	准确率/%
基于VGG-19网络方法	61	85.07
基于ResNet50网络方法	61	90.48
CDCNN-V2方法	61	91.51
FCMSAN方法	61	93.92
本文方法	61	95.24

网络架构	分类数	准确率/%
基于VGG-19网络的方法	2	95.03
基于ResNet50网络的方法	2	97.05
CDCNN-V2方法	2	96.77
FCMSAN方法	2	97.30
本文方法	2	97.54

网络架构	分类数	准确率/%
基于VGG-19网络的方法	20	92.06
基于ResNet50网络的方法	20	92.97
CDCNN-V2方法	20	93.02
FCMSAN方法	20	94.28
本文方法	20	95.27