Alfalfa-Yolo: a boundary-aware lightweight framework for real-time weed detection in alfalfa fields
- 1College of Grassland Science, Qingdao Agricultural University
- 2Shandong Key Laboratory for Germplasm Innovation of Saline-alkaline Tolerant Grasses and Trees
Abstract
Accurate weed detection in alfalfa fields is hindered by morphological similarity between crops and weeds at the seedling stage, substantial scale variation across weed targets, and degradation of fine spatial information during feature resampling. To address these challenges, Alfalfa-Yolo is proposed as a boundary-aware compact detection framework integrating three complementary components. The Edge-Perception Global Transfer Network (EPGTNet) uses bidirectional Sobel responses and cross-layer transmission to preserve edge-related high-frequency cues in deeper semantic features. The Multi-Scale Adaptive Feature Pyramid Network (MSAFPN) replaces static concatenation with attention-guided aggregation to adaptively regulate multi-scale contributions. The Enhanced Channel Shuffle Upsampling Module (ECSM) combines depthwise convolution, channel shuffle, and spatial shift to enhance local spatial representation after upsampling. On the self-constructed alfalfa-field dataset comprising 1,100 independent original field images, Alfalfa-Yolo achieves 77.6% [email protected] and 55.9% [email protected]:0.95 with 2.67 M parameters. Independent-dataset benchmarking on CottonWeedDet12 and ImageWeeds confirms that the architecture remains competitive after target-dataset retraining, while a separate zero-shot cross-dataset protocol directly evaluates transfer without target-domain training. Deployment tests on NVIDIA RTX 4090, NVIDIA Orin NX, and RK3588 further characterize the accuracy-efficiency trade-off under different hardware and quantization settings. These results indicate that Alfalfa-Yolo provides a compact and practically deployable solution for real-time weed localization in alfalfa fields.
Related
- HECA-MobileNet: an explainable lightweight attention-based deep learning framework for custard apple disease detection
- JassidNet as a quantization-aware lightweight phenotyping framework for high-throughput cotton jassid (Amrasca biguttula) detection and counting toward objective resistance screening
- LTK-YOLO: Lightweight Keypoint Detection for Robotic Tea Bud Picking
- LeafXNet: A Lightweight Explainable Model With Integrated Attention Mechanism for Fruit Leaf Disease Detection
- Agri-PIDNet-S: Lightweight semantic segmentation for field-structure and obstacle perception in seedling-stage cotton fields
BibTeX
@article{wang2026alfalfayolo,
title = {{Alfalfa-Yolo: a boundary-aware lightweight framework for real-time weed detection in alfalfa fields}},
author = {Jingyi Wang and Kai Li and Tao Zhou and Yuanjun Xu and Yiwen Gao and Changzhong Li and Yanhua Zhao},
journal = {Frontiers in Plant Science},
year = {2026},
month = oct,
doi = {10.3389/fpls.2026.1955970},
url = {https://www.frontiersin.org/articles/10.3389/fpls.2026.1955970/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Alfalfa-Yolo: a boundary-aware lightweight framework for real-time weed detection in alfalfa fields
AU - Jingyi Wang
AU - Kai Li
AU - Tao Zhou
AU - Yuanjun Xu
AU - Yiwen Gao
AU - Changzhong Li
AU - Yanhua Zhao
PY - 2026
DA - 2026/10/15
JO - Frontiers in Plant Science
DO - 10.3389/fpls.2026.1955970
UR - https://www.frontiersin.org/articles/10.3389/fpls.2026.1955970/full
AB - Accurate weed detection in alfalfa fields is hindered by morphological similarity between crops and weeds at the seedling stage, substantial scale variation across weed targets, and degradation of fine spatial information during feature resampling. To address these challenges, Alfalfa-Yolo is proposed as a boundary-aware compact detection framework integrating three complementary components. The Edge-Perception Global Transfer Network (EPGTNet) uses bidirectional Sobel responses and cross-layer transmission to preserve edge-related high-frequency cues in deeper semantic features. The Multi-Scale Adaptive Feature Pyramid Network (MSAFPN) replaces static concatenation with attention-guided aggregation to adaptively regulate multi-scale contributions. The Enhanced Channel Shuffle Upsampling Module (ECSM) combines depthwise convolution, channel shuffle, and spatial shift to enhance local spatial representation after upsampling. On the self-constructed alfalfa-field dataset comprising 1,100 independent original field images, Alfalfa-Yolo achieves 77.6% [email protected] and 55.9% [email protected]:0.95 with 2.67 M parameters. Independent-dataset benchmarking on CottonWeedDet12 and ImageWeeds confirms that the architecture remains competitive after target-dataset retraining, while a separate zero-shot cross-dataset protocol directly evaluates transfer without target-domain training. Deployment tests on NVIDIA RTX 4090, NVIDIA Orin NX, and RK3588 further characterize the accuracy-efficiency trade-off under different hardware and quantization settings. These results indicate that Alfalfa-Yolo provides a compact and practically deployable solution for real-time weed localization in alfalfa fields.
SN - 1664-462X
ER -