Highlights Utilize the SAM (Segment Anything Model) framework to automatically generate precise segmentation masks. Conduct a comprehensive evaluation of the accuracy and inference efficiency between U-Net and YOLOv11 architectures in weed identification. Abstract. In agriculture, weeds compete with crops for essential resources, often reducing crop yields. Accurate and efficient discrimination between crops and weeds is therefore essential for improving agricultural productivity and enabling effective weed management. Deep learning-based segmentation models such as U-Net and YOLOv11 have shown promise for weed classification tasks, but they differ in terms of segmentation accuracy and computational efficiency. In this study, we used pixel-level masks generated by the Segment Anything Model (SAM) to train both U-Net and YOLOv11, and conducted a comparative evaluation of their performance in weed segmentation. The models were assessed for segmentation accuracy and inference time. The results offered insights into the trade-offs between these approaches and provided guidance for selecting suitable models for real-time weed segmentation in agricultural applications. Compared with YOLOv11, U-Net achieves higher segmentation accuracy of up to 3.26%, and a lighter U-Net model offers faster inference times of up to 1.76×, making it suitable for real-time tasks. Keywords: Classification, Image processing, Machine vision, Weed management.
Zhang et al. (Thu,) studied this question.