基于改进Swin-Transformer的柑橘病叶分类模型

doi:10.13733/j.jcam.issn.2095-5553.2024.01.035

摘要/Abstract

摘要： 针对柑橘病害人工检测效率低、成本高、准确度低等问题，结合人工智能技术对柑橘病叶进行分类识别。首先，建立模拟复杂环境下的柑橘病叶数据集。其次，提出一种改进的SwinTrasnformer柑橘病叶分类模型，包含局部感知通道增强注意力模块(LPCE)，以提升模型的感受野和特征表达能力，通过通道之间的相关性进行加权，使模型更容易提取关键特征。试验证明本文模型的分类识别准确率达到98.52%，精确率、召回率和F1score分别达到98.17%、98.24%、98.28%，均超过基线模型。该模型为柑橘病害的检测提供技术支撑。

关键词: 柑橘病叶, 深度学习, 分类识别, Swin-Transformer, 注意力模块

Abstract: To address the problems of low efficiency, high cost, and low accuracy in manual detection of citrus diseases, this article combines artificial intelligence technology to classify and identify diseased citrus leaves. Firstly, a dataset of citrus disease leaves under simulated complex environments is established. Secondly, an improved Swin Transformer model for citrus disease leaf classification is proposed, which includes a Local Perception Channel Enhanced Attention Module (LPCE) to enhance the models receptive field and feature representation capabilities. Through weighted correlation between channels, the model is made to extract key features more easily. Experiments demonstrate that the classification accuracy of the proposed model reaches 98.52%, with Precision, Recall, and F1score reaching 98.17%, 98.24%, and 98.28% respectively, all exceeding the baseline model. It providing technical support for the detection of citrus diseases.

Key words: citrus diseased leaves, deep learning, classification recognition, SwinTransformer, attention module

中图分类号:

S432: TP391.43

方俊泽, 郭正, 李歌, 邢素霞, 王瑜. 基于改进Swin-Transformer的柑橘病叶分类模型[J]. 中国农机化学报, 2024, 45(1): 252-258.

Fang Junze, Guo Zheng, Li Ge, Xing Suxia, Wang Yu. Classification model of citrus disease leaf based on improved SwinTransformer[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(1): 252-258.

参考文献

［1］孙爱华, 张礼杰, 李虎, 等. 不同水分处理的三峡库区柑橘品质综合评价［J］. 节水灌溉, 2022(12): 24-30.
Sun Aihua, Zhang Lijie, Li Hu, et al. A comprehensive evaluation of citrus quality in the Three Gorges Reservoir Area under different water treatments ［J］. Water Saving Irrigation, 2022(12): 24-30.
［2］边柯橙, 杨海军, 路永华. 深度学习在农业病虫害检测识别中的应用综述［J］. 软件导刊, 2021, 20(3): 26-33.
Bian Kecheng, Yang Haijun, Lu Yonghua. Application review of deep learning in detection and identification of agricultural pests and diseases ［J］. Software Guide, 2021, 20(3): 26-33.
［3］ Wu Y L, Chen Y Y, Lian X Q, et al. Study on the identification method of citrus leaves based on hyperspectral imaging technique ［J］. Spectroscopy and Spectral Analysis, 2021, 41(12): 3837-43.
［4］ Wang K J, Guo D M, Zhang Y, et al. Detection of Huanglongbing (citrus greening) based on hyperspectral image analysis and PCR ［J］. Frontiers of Agricultural Science and Engineering, 2019, 6(2): 172-180.
［5］ Pydipati R, Burks T F, Lee W S. Identification of citrus disease using color texture features and discriminant analysis ［J］. Computers and Electronics in Agriculture, 2006, 52(1-2): 49-59.
［6］杜英杰, 宗哲英, 王祯, 等. 农作物病害诊断方法现状和展望［J］. 江苏农业科学, 2023, 51(6): 16-23.
Du Yingjie, Zong Zheying, Wang Zhen, et al. Current situation and prospect of diagnostic methods for crop diseases ［J］. Jiangsu Agricultural Sciences, 2023, 51(6): 16-23.
［7］李亚文, 张军, 陈月星. 基于Kmeans和特征提取的植物叶部病害检测与实现［J］. 陕西农业科学, 2021, 67(6): 33-37, 41.
Li Yawen, Zhang Jun, Chen Yuexing. Detection of plant leaf diseases based on Kmeans and feature extraction ［J］. Shaanxi Journal of Agricultural Sciences, 2021, 67(6): 33-37, 41.
［8］孙瑜, 张永梅, 武玉军. 基于粒子群算法和支持向量机的黄花菜叶部病害识别［J］. 中国农学通报, 2022, 38(8): 135-140.
Sun Yu, Zhang Yongmei, Wu Yujun. Recognition of hemerocallis citrina leaf disease based on PSO and SVM ［J］. Chinese Agricultural Science Bulletin, 2022, 38(8): 135-140.
［9］孙俊, 朱伟栋, 罗元秋, 等. 基于改进MobileNetV2的田间农作物叶片病害识别［J］. 农业工程学报, 2021, 37(22): 161-169.
Sun Jun, Zhu Weidong, Luo Yuanqiu, et al. Recognizing the diseases of crop leaves in fields using improved MobilenetV2 ［J］. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(22): 161-169.
［10］刘拥民, 胡魁, 聂佳伟, 等. 基于MSDBResNet的水稻病虫害识别［J/OL］. 华南农业大学学报: 1-11［2023-08-25］. http://kns.cnki.net/kcms/detail/44.1110.S.20230707.1325.004.html.
Liu Yongmin, Hu Kui, Nie Jiawei, et al. Rice pest identification based on MSDBResNet ［J/OL］. Journal of South China Agricultural University: 1-11［2023-08-25］. http://kns.cnki.net/kcms/detail/44.1110.S.20230707.1325.004.html.
［11］杜海顺, 张春海, 安文昊, 等. 基于多层信息融合和显著性特征增强的农作物病害识别［J］. 农业机械学报, 2023, 54(7): 214-222.
Du Haishun, Zhang Chunhai, An Wenhao, et al. Crop disease recognition based on multilayer information fusion and saliency feature enhancement ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(7): 214-222.
［12］ Wang B. Identification of crop diseases and insect pests based on deep learning ［J］. Scientific Programming, 2022.
［13］ Mallick M T, Biswas S, Das A K, et al. Deep learning based automated disease detection and pest classification in Indian mung bean ［J］. Multimedia Tools and Applications, 2023, 82(8): 12017-41.
［14］ Xiao Z T, Yin K, Geng L, et al. Pest identification via hyperspectral image and deep learning ［J］. Signal Image and Video Processing, 2022, 16(4): 873-80.
［15］ Liu Y W, Zhang X, Gao Y X, et al. Improved CNN method for crop pest identification based on transfer learning ［J］. Computational Intelligence and Neuroscience, 2022.
［16］ Liu Z, Hu H, Lin Y T, et al. Swin Transformer V2: Scaling up capacity and resolution ［C］. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022: 11999-12009.
［17］ Liu Z, Lin Y, Cao Y, et al. Swin Transformer: Hierarchical vision transformer using shifted windows ［C］. Proceedings of the IEEE International Conference on Computer Vision. IEEE Press, 2021: 10012-10022
［18］ Chen L C, Papandreou G, Kokkinosi I, et al. DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4): 834-848.
［19］ Chen L C, Papandreou G, Schroff F, et al. Rethinking atrous convolution for semantic image segmentation ［EB/OL］. https://arxiv.org/abs/1706.05587v3, 2022-05-09.
［20］钱坤, 李晨瑄, 陈美杉, 等. 基于改进Swin Transformer的舰船目标实例分割算法［J/OL］. 系统工程与电子技术: 1-12［2023-09-04］. http://kns.cnki.net/kcms/detail/11.2422.TN.20220625.1335.012.html.
Qian Kun, Li Chenxuan, Chen Meishan, et al. A ship target instance segmentation algorithm based on improved Swin Transformer ［J/OL］. Systems Engineering and Electronics: 1-12［2023-09-04］. http://kns.cnki.net/kcms/detail/11.2422.TN.20220625.1335.012.html.
［21］ Wang Q L, Wu B G, Zhu P F, et al. ECANet: Efficient channel attention for deep convolutional neural networks ［C］. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2020: 11531-11539.
［22］ He T, Zhang Z, Zhang H, et al. Bag of tricks for image classification with convolutional neural networks ［C］. IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2019: 558-567.
［23］ Kim Y. Convolutional neural networks for sentence classification ［J］. Eprint Arxiv, 2014.
［24］ He K, Zhang X, Ren S, et al. Deep residual learning for image recognition ［C］. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
［25］ Simonyan K, Zisserman A. Very deep convolutional networks for largescale image recognition ［J］. arXiv: 1409.1556.2014(9).
［26］ Dosovitskiy A, Beyer L, Kolesnikov A, et al. An image is worth 16×16 words: Transformers for image recognition at scale ［C］. 9th International Conference on Learning Representations, 2021: 1-22.

[1]	叶荣, 马自飞, 高泉, 李彤, 邵郭奇, 王白娟. 基于改进YOLOv5sECAASFF算法的茶叶病害目标检测[J]. 中国农机化学报, 2024, 45(1): 244-251.
[2]	林森, , 许童羽, 葛禹豪, 马璟, 孙添龙, 赵春江. 基于改进YOLOv5l的设施番茄3D信息检测方法[J]. 中国农机化学报, 2024, 45(1): 274-284.
[3]	侯炳法, 李小敏, 牟向伟, 姚华平. 圣女果温室巡检机器人系统设计与试验[J]. 中国农机化学报, 2024, 45(1): 285-294.
[4]	吴潇, 杨颖, 刘刚, 张倩, 宁远霖. 基于迁移学习和改进ResNet34的猪个体识别方法[J]. 中国农机化学报, 2023, 44(9): 214-221.
[5]	孙松丽, 温宏愿, 刘宾龄, 钟锦扬, 毛政兴. 基于ROS和深度学习的杏鲍菇智能分选系统设计[J]. 中国农机化学报, 2023, 44(8): 95-102.
[6]	李宗南, 蒋怡, 王思, 李源洪, 黄平, 魏鹏. 基于YOLOv5模型的飞蓬属入侵植物目标检测[J]. 中国农机化学报, 2023, 44(7): 200-206.
[7]	王春桃, , , , 梁炜健, 郭庆文, 钟浩, 甘雨, 肖德琴, , . 农业害虫智能视觉检测研究综述[J]. 中国农机化学报, 2023, 44(7): 207-213.
[8]	陈维美, 刘馨蔚, 王铁伟, 徐文凯, 李娟. 基于两级融合深度学习的松材线虫病识别[J]. 中国农机化学报, 2023, 44(7): 214-219.
[9]	李明, 丁智欢, 赵靖暄, 陈思铭, 李文勇, 杨信廷. 基于改进YOLOv5s的日光温室黄瓜霜霉病孢子囊检测计数方法[J]. 中国农机化学报, 2023, 44(5): 63-70.
[10]	杨承磊, 兰玉彬, 王庆雨, 别晓婷, 单常峰, 王国宾, . 神经网络在温室小气候预测中的应用[J]. 中国农机化学报, 2023, 44(5): 89-99.
[11]	付豪, , 赵学观, 翟长远, , 郑康, 郑申玉, 王秀. 基于深度学习的杂草识别方法研究进展[J]. 中国农机化学报, 2023, 44(5): 198-207.
[12]	段宇飞, 董庚, 孙记委, 王焱清, . 基于SE-ResNet网络的油茶果果壳与茶籽分选模型[J]. 中国农机化学报, 2023, 44(4): 89-95.
[13]	轩梦辉, 赵思夏, 徐立友, 陈小亮, 李团飞. 改进VMD和LSTM的联合收割机装配质量检测方法[J]. 中国农机化学报, 2023, 44(3): 132-140.
[14]	王新彦, 易政洋. 基于改进YOLOv5的割草机器人工作环境障碍物检测方法研究[J]. 中国农机化学报, 2023, 44(3): 171-176.
[15]	何雨霜, 王琢, 王湘平, 肖进, 罗友谊, 张俊峰. 深度学习在农作物病害图像识别中的研究进展[J]. 中国农机化学报, 2023, 44(2): 148-155.