AUTHOR=Wang Lining , Wang Guanping , Yang Sen , Liu Yan , Yang Xiaoping , Feng Bin , Sun Wei , Li Hongling 

TITLE=Research on improved YOLOv8n based potato seedling detection in UAV remote sensing images

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 15 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2024.1387350

DOI=10.3389/fpls.2024.1387350

ISSN=1664-462X

ABSTRACT=Accurate detection of potato seedlings is crucial for obtaining information on potato seedlings and ultimately increasing potato yield. In this study, we established a dataset of drone-captured im-ages of potato seedlings and proposed a lightweight model, VBGS-YOLOv8n, based on an en-hanced version of YOLOv8n. This model employs a lighter VanillaNet as the backbone network in-stead of the original YOLOv8n model. To address the small target features of potato seedlings, we introduced a weighted bidirectional feature pyramid network to replace the path aggregation network, reducing information loss between network layers, facilitating rapid multi-scale feature fusion, and enhancing detection performance. Additionally, we incorporated GSConv and Slim-neck designs at the Neck section to balance accuracy while reducing model complexity. Experimental results on pota-to seedling detection in the test set demonstrate outstanding performance across various model evalu-ation metrics. VBGS-YOLOv8n, with 1,524,943 parameters and 4.2 billion FLOPs, achieves a preci-sion of 97.1%, a mean average precision of 98.4%, and an inference time of 2.0ms. Comparative tests reveal that VBGS-YOLOv8n strikes a balance between detection accuracy, speed, and model effi-ciency compared to YOLOv8 and other mainstream networks. Specifically, compared to YOLOv8, the model parameters and FLOPs are reduced by 51.7% and 52.8% respectively, while precision and a mean average precision are improved by 1.4% and 0.8% respectively, and the inference time is re-duced by 31.0%. Comparative experiments with mainstream models such as YOLOv7, YOLOv5, RetinaNet and QueryDet demonstrate that VBGS-YOLOv8n outperforms these models overall, highlighting the effectiveness of this model. Therefore, the research on VBGS-YOLOv8n enables efficient detection of potato seedlings in drone remote sensing images, providing a reference for sub-sequent efficient identification of potato seedlings and deployment on mobile devices.