河南农业大学学报2025，Vol.59Issue(5)：767-775,9.DOI:10.16445/j.cnki.1000-2340.20250828.001
机器学习改进卷积神经网络在作物病害识别中的研究进展
Research progress of machine learning-improved convolutional neural network in crop disease recognition
摘要
Abstract
This review summarizes the integrated improvement methods for crop disease identification,evolving from machine learning to convolutional neural network(CNN).It systematically outlines the key technologies involved in both machine learning and CNN-based crop disease indentification,including six major application processes:data acquisition,data preprocessing,model training,net-work architecture selection,feature extraction and fusion,and model validation.The core reasons for the performance differences between the two methods are analyzed,and the shared technical chal-lenges(such as high data requirements,high computational demands,and limited generalization capa-bility)are identified.Corresponding strategies for using machine learning to enhance CNN-based crop disease identification are also summarized.Finally,the review highlights current research challenges and discusses potential future research directions.关键词
卷积神经网络/机器学习/深度学习/作物病害/病害识别Key words
convolutional neural networks/machine learning/deep learning/crop diseases/disease recognition分类
农业科技引用本文复制引用
汪强,李美琳,马新明,乔红波,郭伟,时雷,熊淑萍,樊泽华,郑光..机器学习改进卷积神经网络在作物病害识别中的研究进展[J].河南农业大学学报,2025,59(5):767-775,9.基金项目
"十四五"国家重点研发计划(2023YFD2301503) （2023YFD2301503）
国家自然科学基金项目(32271993) （32271993）
河南省科技研发计划联合基金优势学科培育项目(222301420114) （222301420114）
河南省科技攻关国际科技合作项目(242102521027) （242102521027）
