Crop diseases and pests,which cause great loss in grain yield and quality,seriously affect the sustainable development of agriculture and environment. Deep learning provides a new method for identification and control of crop diseases and pests and presents unique advantage in detection accuracy and efficiency. On the basis of summarizing the development history and algorithm advantages and disadvantages of deep learning, the application, problems and development trends in crop diseases and pests were discussed in the paper. It is known that color images collected by professional camera and mobile phone are the most important data source, CNN(convolutional neural networks)is the basic structure of artificial neural networks, and deep learning based on transfer learning is the research focus at present. To accelerate the application of deep learning techniques and promote the development of smart agriculture, several aspects should be strengthened, which include constructing disease and pest datasets as soon as possible, optimizing deep learning network structure, building mobile platform, studying single and mixed diseases and pests and integrating unmanned aerial vehicle and satellite remote sensing and in situ observations. It was point out in the paper that the application of deep learning techniques to crop disease and pest identification would strengthen the control of crop diseases and pests on the basis of the protection of ecological environment, hence to ensure the yield and quality of crops.
KRIZHEVSKYA, SUTSKEVERI, HINTONG E.ImageNet classification with deep convolutional neural networks: NIPS'12[C].New York:Curran Associates Incoporated,2012:1097-1105.
[14]
KHANA, SOHAILA, ZAHOORAU,et al.A survey of the recent architectures of deep convolutional neural networks[J].Artificial Intelligence Review,2020,53(8):5455-5516.
SALEEMM H, POTGIETERJ, MAHMOOD ARIFK.Plant disease detection and classification by deep learning[J]. Plants,2019,8(11):468.
[17]
LUY, YOUNGS.A survey of public datasets for computer vision tasks in precision agriculture[J].Computers and Electronics in Agriculture,2020,178:105760.
ARSENOVICM, KARANOVICM, SLADOJEVICS,et al.Solving current limitations of deep learning based approaches for plant disease detection[J].Symmetry,2019,11:939.
SIMONYANK, ZISSERMANA.Very deep convolutional networks for large-scale image recognition[J].arXiv preprint arXiv,2014:1409.1556.
[31]
SHAFIQM, GUZ.Deep residual learning for image recognition:A survey[J].Applied Sciences,2022,12(18):8972.
[32]
YANGL, YUX, ZHANGS,et al.GoogLeNet based on residual network and attention mechanism identification of rice leaf diseases[J].Computers and Electronics in Agriculture,2023,204:107543.
TOOE C, YUJIANL, NJUKIS,et al.A comparative study of fine-tuning deep learning models for plant disease identification[J].Computers and Electronics in Agriculture,2019,161:272-279.
[38]
ISHENGOMAF S, RAII A, SAIDR N.Identification of maize leaves infected by fall armyworms using UAV-based imagery and convolutional neural networks[J].Computers and Electronics in Agriculture,2021,184:106124.
[39]
THENMOZHIK, SRINIVASULU REDDYU.Crop pest classification based on deep convolutional neural network and transfer learning[J].Computers and Electronics in Agriculture,2019,164:104906.
[40]
LIY, WANGH, DANGL M,et al.Crop pest recognition in natural scenes using convolutional neural networks[J]. Computers and Electronics in Agriculture,2020,169:105174.
[41]
HOWARDA G, ZHUM, CHENB,et al.MobileNets: Efficient convolutional neural networks for mobile vision applications[J].arXiv preprint arXiv,2017:1704.04861.
DECHANTC, WIESNER-HANKST, CHENS,et al.Automated identification of northern leaf blight-infected maize plants from field imagery using deep learning[J].Phytopathology,2017,107(11):1426-1432.
HASANR I, YUSUFS M, ALZUBAIDIL.Review of the state of the art of deep learning for plant diseases:A broad analysis and discussion[J].Plants,2020,9(10):1302.
[60]
马铮.基于深度-迁移学习的玉米叶部病害识别方法研究[D].哈尔滨:东北农业大学, 2021.
[61]
BARBEDOJGA.A review on the main challenges in automatic plant disease identification based on visible range images[J].Biosystems Engineering,2016,144:52-60.