%0 Journal Article
%T 基于深度学习的田间害虫自动识别技术综述<br>Survey of Automatic Identification of Field Pests Based on Deep Learning
%A 赵雪如
%A 李晖
%A 胡欣仪
%A 李超然
%A 唐栩燃
%A 赵泽华
%A 罗伟
%A 谭廷俊
%J Journal of Image and Signal Processing
%P 77-88
%@ 2325-6745
%D 2023
%I Hans Publishing
%R 10.12677/JISP.2023.122008
%X 中国是世界上最大的农业国之一，虫害是农业发展面临的一大问题，而害虫种类的精确识别是预测虫害、进行防治的重要基础。在过去，传统的人工害虫识别方法效率低，难以满足害虫防治的要求。随着科技的进步，害虫自动识别技术在农业实践中被广泛应用。在此背景下，本文探究了近年来基于深度学习的害虫自动识别技术的研究进展与发展状况，总结并比较了相关研究方法的异同与有关算法的创新点，以及在不同数据集的识别效果。最后讨论了深度学习在该领域面临的问题与挑战，得出今后的研究重点是建立综合性的害虫识别技术体系，将有关神经网络与优化算法结合起来，以提高害虫识别的准确率。本文可为深入研究基于深度学习的田间害虫自动识别技术提供参考。<br />
China is one of the largest agricultural countries in the world, and insect pest is a major problem faced by agricultural development, and accurate identification of insect species is an important basis for forecasting insect pests and controlling them. In the past, traditional artificial pest identification methods were inefficient and difficult to meet the requirements of pest control. With the progress of science and technology, automatic pest identification technology is widely used in agricultural practice. In this context, this paper explores the research progress and development of automatic pest identification technology based on deep learning in recent years, summarizes and compares the similarities and differences of relevant research methods and the innovation of relevant algorithms, as well as the recognition effect in different data sets. Finally, the problems and challenges faced by deep learning in this field are discussed, and it is concluded that the focus of future research is to establish a comprehensive pest identification technology system, and combine relevant neural networks with optimization algorithms to improve the accuracy of pest identification. This paper can provide a reference for further research on automatic identification technology of field pests based on deep learning.
%K 深度学习，田间害虫识别，自动识别技术<br>Machine Learning
%K Identification of Field Pests
%K Automatic Identification Technology
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=63853