基于多模态融合技术的番茄灰霉病智能协同诊断模型研究

doi:10.13733/j.jcam.issn.2095-5553.2023.11.018

摘要/Abstract

摘要： 为快速、准确地监测设施番茄灰霉病的发生情况，选取灰霉病发生的环境因子特征和图像特征两类数据，分别构建基于单因子的灰霉病的识别模型，并研究两个模型间的关联识别模型。首先，连续采集番茄灰霉病发生与不发生设施温室的最高空气温度和平均空气湿度，构建基于Logistic回归分析的设施番茄灰霉病温湿度预测模型；然后，开展番茄叶部灰霉病RGB图像的采集和预处理，建立图像数据集，构建基于ResNet50-CBAM卷积神经网络的番茄灰霉病RGB图像识别模型；最后，运用多模态融合技术，以温湿度预测模型为文本模态，图像识别模型为图像模态，构建番茄灰霉病智能协同诊断模型。试验结果表明：在VGG16，MobileNet V2，ResNet50和ResNet50-CBAM四个网络模型中，ResNet50-CBAM网络结构准确率最高，达到95.48%，使用基于多模态融合技术的番茄灰霉病智能协同诊断技术的准确率达98.3%，比温湿度预测模型提高14.7%，比RGB图像识别模型提高2.82%。

关键词: 番茄灰霉病, 多模态融合, 回归分析, 深度神经网络

Abstract: In order to monitor the occurrence of tomato Botrytis quickly and accurately, we selected two kinds of data of environmental factor feature and image feature of the occurrence of gray mold to construct the recognition model of Botrytis based on single factor respectively, and studied the correlation recognition model between the two models. Firstly, we collected the maximum air temperature and average air humidity of the occurrence and nonoccurrence of Botrytis, and established the prediction model of Botrytis based on Logistic regression analysis. Secondly, RGB images of Botrytis on tomato leaves were collected and prepossessed to establish an image data set. Then we established a ResNet50-CBAM CNN RGB image recognition model of Botrytis based on it. Finally, the temperature and humidity prediction model was used as the text mode, and the image recognition model was used as the image mode, we constructed the intelligent collaborative diagnosis model of Botrytis by the multimode fusion technology. Experimental results showed that among the four network models VGG16, MobileNet V2, ResNet50 and RESNET50-CBAM, the ResNet50-CBAM network structure had the highest accuracy, reaching 95.48%. The accuracy of intelligent collaborative diagnosis of Botrytis based on multimode fusion technology was 98.3%, which was 14.7% higher than that of temperature and humidity prediction model, and 2.82% higher than that of RGB image recognition model.

Key words: tomato Botrytis, multimode fusion, logistic analysis, deep neural network

中图分类号:

S274.1

杜亚茹, 黄媛, 杜鹏飞, 高欣娜, 武猛, 杨英茹, . 基于多模态融合技术的番茄灰霉病智能协同诊断模型研究[J]. 中国农机化学报, 2023, 44(11): 115-122.

Du Yaru, , Huang Yuan, , Du Pengfei, , Gao Xinna, , Wu Meng, , Yang Yingru, . Study on intelligent collaborative diagnosis model of tomato Botrytis based on multimode fusion technology[J]. Journal of Chinese Agricultural Mechanization, 2023, 44(11): 115-122.

参考文献

［1］　林建吾, 张欣, 陈孝玉龙, 等. 基于轻量化卷积神经网络的番茄病害图像识别［J］. 无线电工程, 2022(8): 1-11.
Lin Jianwu, Zhang Xin, Chen Xiaoyulong, et al. Tomato disease image recognition based on lightweight Convolutional Neural Network ［J］. Radio Engineering, 2022(8): 1-11.
［2］　谢传奇, 何勇. 利用光谱反射特性对番茄叶片早疫病害程度的识别［J］. 中国科技论文, 2017, 12(6): 671-675.
Xie Chuanqi, He Yong. Identification of early blight disease degree of tomato leaf using spectral reflectance ［J］.China Sciencepaper, 2017, 12(6): 671-675.
［3］　孔汶汶, 虞佳佳, 刘飞, 等. 基于高光谱成像技术的番茄茎秆灰霉病早期诊断研究［J］. 光谱学与光谱分析, 2013, 33(3): 733-736.
Kong Wenen, Yu Jiajia, Liu Fei, et al. Early diagnosis of gray mold on tomato stalks based on Hyperspectral data ［J］. Spectroscopy and Spectral Analysis, 2013, 33(3): 733-736.
［4］　瞿华香, 赵萍, 陈桂鹏, 等. 基于无线传感器网络的精准农业研究进展［J］. 中国农学通报, 2014, 30(33): 268-272.
Zhai Huaxiang, Zhao Ping, Chen Guipeng, et al. Advances in precision agriculturebased wireless sensor networks ［J］. Chinese Agricultural Science Bulletin, 2014, 30(33): 268-272.
［5］　杨英茹, 黄媛, 高欣娜, 等. 基于Logistic回归模型的设施番茄病毒病预警模型构建［J］. 河北农业科学, 2019, 23(5): 91-94.
Yang Yingru, Huang Yuan, Gao Xinna, et al. Early warning model construction of greenhouse tomato virus disease based on Logitic regression model ［J］. Journal of Hebei Agricultural Sciences, 2019, 23(5): 91-94.
［6］　陈杰, 杨银娟, 严宝华, 等. 番茄黄化曲叶病毒病预警模型的建立［J］. 中国植保导刊, 2017, 37(3): 49-52.
Chen Jie, Yang Yinjuan, Yan Baohua, et al. Establishment of early warning model for tomato yellow leaf curl virus disease ［J］. China Plant Protection, 2017, 37(3): 49-52.
［7］　王晓蓉, 吕雄杰, 贾宝红. 基于物联网技术的日光温室黄瓜白粉病预警系统研究［J］. 农学学报, 2016, 6(8): 50-53.
Wang Xiaorong, Lü Xiongjie, Jia Baohong. Construction of cucumber powdery mildew early warning system in solar greenhouse based on internet of things ［J］. Journal of Agriculture, 2016, 6(8): 50-53.
［8］　曹跃腾, 朱学岩, 赵燕东, 等. 基于改进ResNet的植物叶片病虫害识别［J］. 中国农机化学报, 2021, 42(12): 175-181.
Cao Yueteng, Zhu Xueyan, Zhao Yandong, et al. Recognition of plant leaf diseases, insect, and pests based on improved ResNet ［J］. Journal of Chinese Agricultural Mechanization, 2021, 42(12): 175-181.
［9］　〖ＺＫ（〗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.〖ＺＫ）〗［9］　Xu Y, Zhao B, Zhai Y, et al. Maize diseases identification method based on multiscale convolutional global pooling neural network ［J］. IEEE Access, 2021, 9: 27959-27970.
［11］　〖ＺＫ（〗Pham T N, Van Tran L, Dao S V T. Early disease classification of mango leaves using feedforward neural network and hybrid metaheuristic feature selection ［J］. IEEE Access, 2020, 8: 189960-189973.〖ＺＫ）〗［12］　〖ＺＫ（〗〖JP3〗鲍文霞, 黄雪峰, 胡根生, 等. 基于改进卷积神经网络模型的玉米叶部病害识别［J］. 农业工程学报, 2021, 37(6): 160-167.〖JP〗〖JP2〗Bao Wenxia, Huang Xuefeng, Hu Gensheng, et al. Identification of maize leaf diseases using improved convolutional neural network ［J］.Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(6): 160-167.〖JP〗〖ＺＫ）〗［10］　李书琴, 陈聪, 朱彤, 等. 基于轻量级残差网络的植物叶片病害识别［J］. 农业机械学报, 2022, 53(3): 243-250.
Li Shuqin, Chen Cong, Zhu Tong, et al. Plant leaf disease identification based on lightweight residual network ［J］. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(3): 243-250.
［11］　郝菁, 贾宗维. 基于图像识别的苹果叶片病害识别模型对比研究［J］. 中国农学通报, 2022, 38(12): 153-158.
Hao Jing, Jia Zongwei. Comparative study on apple leaf disease recognition models based on image recognition ［J］. Chinese Agricultural Science Bulletin, 2022, 38(12): 153-158.
［12］　孟亮, 郭小燕, 杜佳举, 等. 一种轻量级CNN农作物病害图像识别模型［J］. 江苏农业学报, 2021, 37(5): 1143-1150.
Meng Liang, Guo Xiaoyan, Du Jiaju, et al. A lightweight CNN model for image recognition of crop disease ［J］. Jiangsu Journal of Agricultural Sciences, 2021, 37(5): 1143-1150.
［13］　周宏威, 沈恒宇, 袁新佩, 等. 基于迁移学习的苹果树叶片病虫害识别方法研究［J］. 中国农机化学报, 2021, 42(11): 151-158.
Zhou Hongwei, Chen Hengyu, Yuan Xinpei, et al. Research on identification method of apple leaf diseases based on transfer learning ［J］. Journal of Chinese Agricultural Mechanization, 2021, 42(11): 151-158.
［14］　李庆盛, 缪楠, 张鑫, 等. 基于注意力机制非对称残差网络和迁移学习的玉米病害图像识别［J］. 科学技术与工程, 2021, 21(15): 6249-6256.
Li Qingsheng, Miao Nan, Zhang Xin, et al. Image recognition of maize disease based on asymmetric convolutional attention residual network and transfer learning ［J］. Science Technology and Engineering, 2021, 21(15): 6249-6256.
［15］　陆雅诺, 陈炳才. 基于注意力机制的小样本啤酒花病虫害识别［J］. 中国农机化学报, 2021, 42(3): 189-196.

[1]	高芳征, 汤文俊, 陈光明, 黄家才. 基于改进YOLOv3的复杂环境下西红柿成熟果实快速识别[J]. 中国农机化学报, 2023, 44(8): 174-183.
[2]	袁馨, 李薇, 燕振刚. 晋城市农业经济增长影响因素实证分析——以阳城县为例[J]. 中国农机化学报, 2023, 44(1): 234-240.
[3]	吴小伟, 张璐, 唐莉莉, 韦家武, 陆海莉, 周忠诚. 基于BoxBehnken模型大蒜机播成本影响因素分析[J]. 中国农机化学报, 2022, 43(5): 211-217.
[4]	彭虹. 中国与非洲农产品贸易网络结构与合作态势分析——基于“一带一路”倡议[J]. 中国农机化学报, 2022, 43(4): 216-222.
[5]	李小敏, 张日红, 陈天赐, 侯炳法, 张权, 李雄, . 基于深度学习的林下落果识别方法与试验[J]. 中国农机化学报, 2021, 42(9): 202-208.
[6]	李向军;张成亮;赵大勇;吴家安;李庆达;周福君;. 气吹式垄台秸秆根茬处理机设计与试验[J]. 中国农机化学报, 2021, 42(5): 13-18.
[7]	安小宇;赵复兴;柳海涛;. 土壤墒情预测相关性数据耦合性试验分析[J]. 中国农机化学报, 2021, 42(4): 134-141.
[8]	丁龙朋, 李枫楠, 王成霄, 李景彬, 葛建兵, . 盘刷式矮密栽培落地红枣集条装置研制[J]. 中国农机化学报, 2021, 42(12): 107-113.
[9]	田立权;张正中;吕美巧;金仁雕;熊永森;. 半喂入联合收割机双速回转脱分装置设计与性能试验[J]. 中国农机化学报, 2020, 41(9): 8-15.
[10]	李向军;李连豪;. 组合式灭茬、苗带旋耕整地机设计与试验[J]. 中国农机化学报, 2020, 41(8): 13-19.
[11]	卜小东;张馨月;黄可京;郭辉;. 基于机器视觉的玉米苗期土壤湿度检测试验研究[J]. 中国农机化学报, 2020, 41(7): 150-156.
[12]	高国华;谢海峰;李博文;田博兴;张赛;. 甘薯跌落碰撞损伤试验研究与分析[J]. 中国农机化学报, 2019, 40(9): 85-90.
[13]	谢万里;李宏志;周辉;尹绍武;. 基于迁移学习与卷积神经网络的鱼濒死预警系统研究[J]. 中国农机化学报, 2019, 40(2): 186-192.