基于卷积神经网络的农作物病虫害检测研究进展  

Research Progress of Crop Disease and Pest Detection Based on Convolutional Neural Network

作　　者：蔡国庆 吴建军[1] 祝玉华[1] 甄彤[1] 李智慧[1] 连一萌 Cai Guoqing;Wu Jianjun;Zhu Yuhua;Zhen Tong;Li Zhihui;Lian Yimeng(School of Information Science&Engineering,Henan University of Technology,Zhengzhou 450001,China)

机构地区：[1]河南工业大学信息科学与工程学院,河南郑州450001

出　　处：《山东农业科学》2024年第11期170-180,共11页Shandong Agricultural Sciences

基　　金：国家重点研发计划项目(2022YFD2100202);国家重大科技专项(2018YFD0401404)。

摘　　要：农作物病虫害是全球农业生产的严重威胁之一,易造成巨大的经济损失。引入机器视觉和机器学习方法进行农作物病虫害检测,不仅可以提高病虫害检测的效率,而且有助于及时采取防治措施,降低损失。卷积神经网络(CNN)作为深度学习的代表技术之一,在计算机视觉领域的图像识别、物体识别等方面应用广泛,在农作物病虫害检测方面也取得了一些成果。本文概述了基于CNN检测农作物病虫害的技术要点、发展历程,综述了该技术的主要研究方向与进展,总结了目前研究中存在的主要问题并提出相应的解决策略,旨在为CNN在农业上的应用提供理论依据,并为农业生产管理的智能化提供技术支撑。Crop pests and diseases pose a significant threat to global agricultural production,often lead-ing to substantial economic losses.The application of machine vision and machine learning methods for crop pest and disease monitoring not only enhances detection efficiency,but also facilitates timely preventive meas-ures to reduce losses.Convolutional neural network(CNN),as a prominent deep learning technique,has been widely used in computer vision for image and object recognition,and also has achieved promising a-chieves in crop disease and pest detection.This paper outlined the key aspects and development history of CNN-based techniques for crop disease and pest detection,reviewed the major research directions and ad-vancements in this area,summarized the main issues in current research,and proposed corresponding solu-tions.The aim of this paper was to provide a theoretical basis for CNN application in agriculture and to offer technical supports for the intelligent management of agricultural production.

关 键 词：农作物病虫害检测 卷积神经网络 深度学习 计算机视觉 

分 类 号：S431.9[农业科学—农业昆虫与害虫防治]