Lin Zhou

To address the low accuracy and limited generalization of intelligent weed recognition under complex paddy-field conditions involving illumination variation, vegetation occlusion, multi-scale targets, and complex background interference, this study proposes an improved YOLOv9-based weed detection model, termed BR-YOLOv9. First, a weighted bidirectional feature pyramid network (BiFPN) was introduced into the neck network of YOLOv9 to strengthen cross-scale feature fusion and improve the representation of small and multi-scale weed targets. Subsequently, the original RepNCSPELAN4 module was reconstructed, and a novel RepFEL module was designed by integrating feature enhancement and a large selective kernel mechanism to improve fine-grained feature extraction and suppress background interference. Experimental results showed that BR-YOLOv9 achieved mAP@0.5 and mAP@0.5:0.95 values of 98.21% and 95.97%, respectively, representing improvements of 2.03 and 3.71 percentage points over the original YOLOv9 model. In the generalization experiments, BR-YOLOv9 was further evaluated under multiple representative field scenarios, including close-up, long-distance, high-illumination, low-illumination, and multi-target conditions. The results showed that BR-YOLOv9 maintained stable detection performance across different imaging distances, illumination conditions, and target densities, demonstrating stronger robustness and generalization capability in complex agricultural environments. The proposed model provides a robust visual perception method for paddy-field weed detection and offers technical support for precision weed management and intelligent weeding equipment.

Accurate estimation of winter wheat aboveground biomass (AGB) is essential for crop growth monitoring and precision agricultural management. To reduce the effects of canopy structural complexity and spectral saturation on AGB estimation, this study evaluated winter wheat grown under different compost substitution ratios and planting densities. Based on unmanned aerial vehicle (UAV) multispectral and RGB imagery acquired over two growing seasons at four key growth stages, spectral vegetation indices, colour vegetation indices, and canopy structural features were extracted and integrated. Recursive feature elimination, Elastic Net, and support vector regression were used to construct stage-specific AGB estimation models. The optimal feature strategy varied among growth stages, indicating that AGB estimation requires stage-specific feature selection rather than a single fixed feature combination. The proposed framework achieved validation R² values of 0.872, 0.898, 0.867, and 0.895 at the jointing, booting, flowering, and grain-filling stages, respectively, and the corresponding RRMSE values were 12.5%, 12.1%, 14.3%, and 12.0%, respectively. Additional comparisons with PLSR, RF, and XGBoost based on the stage-specific optimal feature sets further confirmed the competitive performance of SVR under the present small-sample and multi-source feature conditions. Model improvement was more evident at the flowering and grain-filling stages. At these stages, the integration of selected spectral, colour, and structural features better represented canopy closure, spike-layer formation, and late-season biomass variation. Under the treatment combining 20% compost substitution with a planting density of 4.5 million plants ha⁻¹, winter wheat maintained relatively high AGB levels across growth stages. The novelty of this study lies in demonstrating that the effectiveness of multi-source UAV feature fusion for winter wheat AGB estimation is growth-stage dependent and is enhanced when coupled with feature selection. These findings provide a methodological reference for multi-temporal AGB monitoring and precision cultivation management under similar field conditions.