摘要
Abstract
As an agricultural development model that uses modern information technology to achieve intelligence and sustainability,smart agriculture has an increasing demand for efficient and accurate analysis and modeling of multimodal agricultural big data.As a model architecture that has had great influence in the field of deep learning in recent years,Transformer can effectively solve various tasks in smart agricultural application scenarios and improve agricultural production efficiency.This paper first introduces the mathematical principles and model architecture of the Transformer model,and deeply discusses the Transformer's innovative application in agricultural scenarios such as crop disease detection,pest and insect identification and monitoring,crop breeding and genetic improvement,agricultural remote sensing image analysis,livestock management,aquaculture management,and crop yield prediction.Finally,it summarizes the challenges and future development directions faced by Transformer in smart agriculture applications.关键词
智慧农业/Transformer/深度学习/人工智能/多模态数据Key words
smart agriculture/Transformer/deep learning/artificial intelligence/multimodal data分类
计算机与自动化
