| 754 | 24 | 234 |
| 下载次数 | 被引频次 | 阅读次数 |
针对番茄病害识别中深度神经网络参数过多、识别精度较低的问题，从网络轻量化和提取特征精准化的角度出发，对SqueezeNet结构进行改进。为精简fire模块，对其中Expand层的卷积核大小、网络层数以及通道数进行调整。同时，将模型与ECA模块结合，利用局部跨通道交互的方式获得各通道的注意力值，强化网络对关键特征的学习能力。实验结果表明，与LeNet、MobileNet和SqueezeNet模型相比，改进型SqueezeNet的模型大小和识别准确率均具有明显优势，为嵌入式设备在实际生产中的番茄病害识别提供一种技术方法。
Abstract：Aiming at the problems of too many parameters and low recognition accuracy in deep neural network for tomato disease recognition, the SqueezeNet structure was improved from the perspective of simplified network and precise features recognition. In order to simplify the fire module, the size of the convolution kernel, the number of network layers and channels in the Expand layer were adjusted. At the same time, the model was combined with the ECA module to obtain the attention value of each channel by local cross-channel interaction, so as to strengthen the network′s learning ability for key features. The experimental results showed that the improved SqueezeNet had obvious advantages in model size and recognition accuracy compared with the LeNet, MobileNet and SqueezeNet models, which provided a technical method for tomato disease recognition in application of embedded devices.
[1] MOHANTY S P,HUGHES D P,SALATHE M.Using deep learning for image-based plant disease detection[J].Frontiers in plant science,2016,7:1419.
[2] 佘颢，吴伶，单鲁泉.基于SSD网络模型改进的水稻害虫识别方法[J].郑州大学学报(理学版),2020,52(3):49-54.SHE H,WU L,SHAN L Q.Improved rice pest recognition based on SSD network model[J].Journal of Zhengzhou university (natural science edition),2020,52(3):49-54.
[3] FUENTES A,IM D H,YOON S,et al.Spectral analysis of CNN for tomato disease identification[M]//Artificial Intelligence and Soft Computing.Cham:Springer International Publishing,2017:40-51.
[4] 孙奥，吴冬燕，吴阳江，等.深度学习在农作物病虫害识别中应用初探[J].电子测试，2019(6):68-70.SUN A,WU D Y,WU Y J,et al.Preliminary study on the application of deep learning in the identification of crop diseases and insect pests[J].Electronic test,2019(6):68-70.
[5] 刘子记，杜公福，牛玉，等.番茄主要病害的发生与防治技术[J].长江蔬菜，2019(19):59-62.LIU Z J,DU G F,NIU Y,et al.Occurrence and control techniques of tomato main diseases[J].Journal of Changjiang vegetables,2019(19):59-62.
[6] 吴健宇.基于深度卷积神经网络的农作物病虫害识别及实现[D].哈尔滨：哈尔滨工业大学，2019.WU J Y.Identification of crop pests and diseases based on deep convolutional neural network and implementation of algorithm[D].Harbin:Harbin Institute of Technology,2019.
[7] 杨胜明.基于图像识别的农作物虫害检测技术研究[D].哈尔滨：哈尔滨工程大学，2018.YANG S M.Research on crop pest detection technology based on image recognition[D].Harbin:Harbin Engineering University,2018.
[8] 陈天娇，曾娟，谢成军，等.基于深度学习的病虫害智能化识别系统[J].中国植保导刊，2019,39(4):26-34.CHEN T J,ZENG J,XIE C J,et al.Intelligent identification system of disease and insect pests based on deep learning[J].China plant protection,2019,39(4):26-34.
[9] DURMUS H,GüNES E O,KIRCI M.Disease detection on the leaves of the tomato plants by using deep learning[C]//Proceedings of the 6th International Conference on Agro-Geoinformatics.Piscataway:IEEE Press,2017:1-5.
[10] ELHASSOUNY A,SMARANDACHE F.Smart mobile application to recognize tomato leaf diseases using convolutional neural networks[C]//Proceedings of the International Conference of Computer Science and Renewable Energies.Piscataway:IEEE Press,2019:1-4.
[11] 李晓振，徐岩，吴作宏，等.基于注意力神经网络的番茄叶部病害识别系统[J].江苏农业学报，2020,36(3):561-568.LI X Z,XU Y,WU Z H,et al.Recognition system of tomato leaf disease based on attentional neural network[J].Jiangsu journal of agricultural sciences,2020,36(3):561-568.
[12] IANDOLA F N,HAN S,MOSKEWICZ M W,et al.SqueezeNet:AlexNet-level accuracy with 50x fewer parameters and < 0.5 MB model size[EB/OL].[2021-04-15].https://arxiv.org/pdf/1602.07360.pdf.
[13] WANG Q L,WU B G,ZHU P F,et al.ECA-net:efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE Press,2020:11531-11539.
[14] HUGHES D P,SALATHE M.An open access repository of images on plant health to enable the development of mobile disease diagnostics[EB/OL].[2021-03-15].https://arxiv.org/ftp/arxiv/papers/1511/1511.08060.pdf.
[15] 浮盼盼.大规模不均衡数据分类方法研究[D].大连：辽宁师范大学，2014.FU P P.Research on classification methods for large-scale imbalanced data[D].Dalian:Liaoning Normal University,2014.
基本信息:
DOI：10.13705/j.issn.1671-6841.2021311
中图分类号:S436.412.1;TP391.41;TP183
引用信息:
[1]胡玲艳,周婷,许巍,等.面向番茄病害识别的改进型SqueezeNet轻量级模型[J],2022,54(04):71-77.DOI:10.13705/j.issn.1671-6841.2021311.
基金信息:
国家自然科学青年基金项目(61601076);; 大连市科技创新基金项目(2020JJ26SN058)
2021-07-15
2021
2022-04-02
2022
2