摘要
农作物病虫害是农业生产中所要面临的重要问题之一,快速、有效的病虫害识别手段对保障农产品质量安全,提升经济效益具有重要意义。近年来基于图像的农作物病虫害自动诊断技术得到了广泛研究,并取得了一系列进步。本文主要分析了我国农作物病虫害自动识别技术的发展现状,阐述了利用卷积神经网络建立病虫害识别模型的关键技术和实施步骤,介绍了当前主流图像识别深度学习模型算法和改进思路,并对存在的技术瓶颈和发展趋势进行分析,以期为图像识别技术在实际生产中的应用提供参考。
Crop disease and insect pest is one of the important issue facing agricultural production.It is of important significance to ensure the quality safety of agricultural products and improve economic benefits by rapid and effective identification of crop disease and insect pest.In recent years,the image-based automatic diagnosis technology of crop disease and pest has been widely studied and a series of progress has been made.This paper mainly analyzed the current development status of automatic identification technology for crop disease and pest in China;expounded the key technology and implementation steps of using convolution neural network to establish the identification model of disease and pest;introduced the current mainstream deep learning model algorithms and idea improvement for image recognition;and analyzed the existing technical bottlenecks and development trends,hoping to provide references for the application of image recognition technology in actual production.
作者
王明
张倩
WANG Ming;ZHANG Qian(Institute of Data Science and Agricultural Economics,Beijing Academy of Agriculture and Forestry Sciences,Beijing 100097)
出处
《中国蔬菜》
北大核心
2023年第3期22-28,共7页
China Vegetables
基金
北京市数字农业创新团队项目(BAIC10-2022)
北京市农林科学院青年科研基金项目(QNJJ202213)。
关键词
病虫害
数据集
预测模型
深度学习
卷积神经网络
研究进展
disease and pest
data set
prediction model
deep learning
convolutional neural network
research progress
