改进残差网络甜瓜叶片病害的识别研究

doi:10.3778/j.issn.1002-8331.2305-0134

摘要/Abstract

摘要： 针对甜瓜叶片不同程度的病害识别研究较少，人工检测实时性差且存在识别准确率较低等问题，提出了一种基于改进残差网络模型的甜瓜叶片病害识别方法。将传统的ResNet50模型作为骨干网络，将ReLU激活函数替换为ELU激活函数；将ResNet50的模型的第一层卷积中的7×7卷积核替换成Incption结构，在全连接层之后加入Dropout层，增强模型的表达能力并缓解过拟合问题；引入多头自注意力机制（MHSA），提高模型的泛化能力。进行数据预处理，将训练集与测试集的比例划分为7∶3，采用数据增强的方式对小样本数据集进行扩充。实验结果表明：改进的残差网络模型准确率与原模型相比提高了1.03个百分点，识别准确率达到98.72%且模型参数量为19.3?MB。与其他网络模型相比准确率大幅提升，可以为甜瓜叶片病害的高效识别和及时预防治理提供参考。

关键词: 甜瓜叶片病害, 图像识别, 残差网络, 多头自注意力机制, 深度学习

Abstract: In view of the fact that there are few studies on the identification of diseases of muskmelon leaves in different degrees, the manual detection has poor real-time performance and the identification accuracy is low, a method for identification of muskmelon leaf diseases based on the improved residual network model is proposed. The traditional ResNet50 model is used as the backbone network, and the ReLU activation function is replaced by the ELU activation function. The 7×7 convolution kernel in the first layer convolution of the ResNet50 model is replaced with the Incption structure, in the fully connected layer, then the Dropout layer is added to enhance the expressive ability of the model and alleviate the overfitting problem. The multi-head self-attention (MHSA) mechanism is introduced to improve the generalization ability of the model. This paper carries out data preprocessing, divides the ratio of training set and test set into 7∶3, and uses data enhancement to expand the small sample data set. The experimental results show that the accuracy rate of the improved residual network model is 1.03 percentage points higher than the original model, the recognition accuracy rate reaches 98.72%, and the model parameter size is 19.3 MB. Compared with other network models, the accuracy rate has been greatly improved, which can provide a reference for efficient identification and timely prevention and treatment of melon leaf diseases.

Key words: melon leaf diseases, image recognition, residual network, multi-head self-attention (MHSA) mechanism, deep learning

黄英来, 姜忠良. 改进残差网络甜瓜叶片病害的识别研究[J]. 计算机工程与应用, 2024, 60(15): 189-197.

HUANG Yinglai, JIANG Zhongliang. Research on Identification of Melon Leaf Diseases with Improved Residual Network[J]. Computer Engineering and Applications, 2024, 60(15): 189-197.

参考文献

[1] 张俊姝. 大棚甜瓜病害综合防治技术[J]. 吉林蔬菜, 2018(3): 37-38.
ZHANG J S. Integrated control technology of muskmelon diseases in greenhouses[J]. Jilin Vegetables, 2018(3): 37-38.
[2] 赵春梅, 陈柏杰, 金荣荣. 不同光照强度对甜瓜叶色黄化突变体幼苗生理指标的影响[J]. 蔬菜, 2019(5): 18-23.
ZHAO C M, CHEN B J, JIN R R. Effects of different illumination intensity on physiological indexes of yellow leaves melon mutant[J]. Vegetables, 2019(5): 18-23.
[3] 王亚红, 范迎宾, 李默莎, 等. 连续性床旁血液滤过联合解磷定 VitB6对急性有机磷农药中毒患者的效果及对心肌损伤炎症症状的影响[J]. 河北医学, 2023, 29(3): 461-466.
WANG Y H, FAN Y B, LI M S, et al. Effect of continuous bedside hemofiltration combined with desmophosphate and VitB6 on patients with acute organophosphorus pesticide poisoning and the effect on myocardial injuryand inflammatory symptoms[J]. Hebei Medical Journal, 2023, 29(3): 461-466.
[4] BHANGE M, HINGOLIWALA H A. Smart farming: pomegranate disease detection usingimage processing[J]. Procedia Computer Science, 2015, 58: 280-288.
[5] 鲍文霞, 吴德钊, 胡根生, 等. 基于轻量型残差网络的自然场景水稻害虫识别[J]. 农业工程学报, 2021, 37(16): 145-152.
BAO W X, WU D Z, HU G S, et al. Rice pest identification in natural scene based on lightweight residual network[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(16): 145-152.
[6] 朱帅, 王金聪, 任洪娥, 等. 基于多特征融合的残差网络果树叶片病害识别[J]. 森林工程, 2022, 38(1): 108-114.
ZHU S, WANG J C, REN H E, et al. Fruit tree leaf disease recognition based on residual network and multi feature fusion[J]. Forest Engineering, 2022, 38(1): 108-114.
[7] 周孟然, 姚旭. 改进多尺度残差网络病害叶片图像识别[J]. 计算机工程与应用, 2023, 59(11): 195-202.
ZHOU M R, YAO X. Improved leaf image recognition of disease in multi-scale residual networks[J]. Computer Engineering and Applications, 2023, 59(11): 195-202.
[8] 秦丰, 刘东霞, 孙炳达, 等. 基于深度学习和支持向量机的4种苜蓿叶部病害图像识别[J]. 中国农业大学学报, 2017, 22(7): 123-133.
QIN F, LIU D X, SUN B D, et al. Image recognition of four different alfalfa leaf diseases based on deep learning and support vector machine[J]. Journal of China Agricultural University, 2017, 22(7): 123-133.
[9] 牛群峰, 刘江鹏, 王莉, 等. 基于残差网络与迁移学习的柑橘类型识别模型研究[J]. 现代农业科技, 2023(8): 73-76.
NIU Q F, LIU J P, WANG L, et al. Research on citrus species recognition model based on residual network and transfer learning[J]. Modern Agricultural Science and Technology, 2023(8): 73-76.
[10] 梁秀满, 肖寒. 基于SDD-YOLO的轻量级带钢缺陷实时检测算法[J]. 中国测试: 1-8[2023-05-05]. http://kns.cnki.net/kcms/detail/51.1714.TB.20230109.1648.002.html.
LIANG X M, XIAO H. Lightweight strip defect real-timedetection algorithm based on SDD-YOLO[J]. China Measurement & Test: 1-8[2023-05-05]. http://kns.cnki.net/kcms/detail/51.1714.TB.20230109.1648.002.html.
[11] TENG S, CHEN G F. Vibration signal-based structural damage detection through deep learning and digital image correlation[C]//Proceedings of the 7th International Conference on Environmental Science and Civil Engineering (ESCE 2021), 2021: 414-419.
[12] 刘广, 胡国玉, 古丽巴哈尔·托乎提, 等. 基于改进YOLOv3的葡萄叶部病虫害检测方法[J]. 微电子学与计算机, 2023, 40(2): 110-119.
LIU G, HU G Y, GULIBAHAER Tuohuti, et al. Detection of grape leaf diseases and insect pests based on improved YOLOv3[J]. Microelectronics & Computer, 2023, 40(2): 110-119.
[13] 孙道宗, 丁郑, 刘锦源, 等. 基于改进SqueezeNet模型的多品种茶树叶片分类方法[J]. 农业机械学报, 2023, 54(2): 223-230.
SUN D Z, DING Z, LIU J Y, et al. Classification method of multi-variety tea leaves based on improved SqueezeNet model[J]. Transactions of the Chinese Society of Agricultural Machinery, 2023, 54(2): 223-230.
[14] 王志强, 于雪莹, 杨晓婧, 等. 基于WGAN和MCA-MobileNet的番茄叶片病害识别[J]. 农业机械学报, 2023, 54(5): 244-252.
WANG Z Q, YU X Y, YANG X J, et al. Tomato leaf diseases recognition based on WGAN and MCA-MobileNet[J]. Transactions of the Chinese Society of Agricultural Machinery, 2023, 54(5): 244-252.
[15] CHENG X, ZHANG Y H, CHEN Y Q, et al. Pest identification via deep residual learning in complex background[J]. Computers and Electronics in Agriculture, 2017, 141: 351-356.
[16] LI C F, WANG B P. Hand gesture recognitionin complex background based on improved deep residuallearning network[C]//2021 International Symposium on Computer Technology and Information Science (ISCTIS), Guilin, China, 2021: 239-245.
[17] 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: 104906.
[18] LIU D M, YANG H X, GONG Y J, et al. A recognition method of crop diseases and insect pests based on transfer learning and convolution neural network[J]. Mathematical Problems in Engineering, 2022(3): 1-10.
[19] ERUL E, I?IN A. ChatGPT ile Sohbetler: Turizmde ChatGPT nin ?nemi (Chats with ChatGPT: Importance of ChatGPT in Tourism)[J]. Journal of Tourism and Gastronomy Studies, 2023.
[20] 李哲, 潘维, 张赟. 基于深度卷积神经网络的植物病虫害识别[J]. 电子技术与软件工程, 2020(23): 194-195.
LI Z, PAN W, ZHANG Y. Plant disease and pest identification based on deep convolutional neural networks[J]. Electronic Technology and Software Engineering, 2020(23): 194-195.
[21] XU L X, CAO B X, ZHAO F J, et al. Wheat leaf disease identification based on deep learning algorithms[J]. Physiological and Molecular Plant Pathology, 2023, 123: 101940.
[22] MIQUE E L, PALAOAG T D. Rice pest and disease detection using convolutional neural network[C]//Proceedings of the 1st International Conference on Information Science and Systems. New York, NY, USA: ACM, 2018: 147-151.
[23] 胡文艺, 王洪坤, 杜育佳. 基于SE模块和ResNet的番茄病虫害识别方法[J]. 农业工程, 2022, 12(9): 33-40.
HU W Y, WANG H K, DU Y J. Tomato pest identification based on the SE module and ResNet[J]. Agricultural Engineering, 2022, 12(9): 33-40.
[24] QIAO Z N, YUAN X H, ZHANG R M, et al. Efficientdeep-narrow residual networks using dilated pooling foscene recognition[J]. Expert Systems with Applications, 2023, 234: 121018.
[25] 方晨晨, 石繁槐. 基于改进深度残差网络的番茄病害图像识别[J]. 计算机应用, 2020, 40(S1): 203-208.
FANG C C, SHI F H. Image recognition of tomato diseases based on improved deep residual network[J]. Journal of Computer Applications, 2020, 40(S1): 203-208.