Review of automatic row alignment technology for intelligent agricultural machinery in the field

• A systematic overview is provided from three key technical aspects: environmental perception, row information extraction, and tracking control. • The advantages and disadvantages of different technologies in terms of accuracy, robustness, real-time performance, cost, and adaptability to different crops were compared. • The applicability of advanced technologies such as the Transformer architecture model, end-to-end navigation, visual large language models, and embodied intelligence row alignment has been analyzed. • The main challenges and future development directions of automatic row alignment technology for agricultural machinery were discussed. Automatic row alignment technology for agricultural machinery is an important foundation for achieving precision agriculture and smart operations, but the complex field environment poses significant challenges to perception stability and control accuracy. This paper systematically reviews the research progress and development trends of automated row alignment technology for agricultural machinery. It analyzed the characteristics of environmental perception technologies: contact, vision, LiDAR, and their fusion strategies. The focus was on comparing the advantages and disadvantages of crop line recognition methods from traditional image processing to deep learning in terms of real-time performance, robustness, and generalization ability. In addition, tracking control models and methods are reviewed, with emphasis on the applicability of model-based and data-driven strategies under complex operating conditions. Finally, emerging trends and key engineering challenges including Transformer architecture, end-to-end navigation, vision-language models, and embodied intelligence were discussed. This paper aims to provide technical references for research and engineering applications of automated row alignment systems.