Accurate weed segmentation in Unmanned Aerial Vehicle (UAV) imagery remains a significant challenge in precision agriculture due to environmental variability, weak contextual representation, and inaccurate boundary detection. To address these limitations, we propose the Multi-Scale and Edge-Aware Network (MSEA-Net), a lightweight and efficient deep learning framework designed to enhance segmentation accuracy while maintaining computational efficiency. Specifically, we introduce the Multi-Scale Spatial-Channel Attention (MSCA) module to recalibrate spatial and channel dependencies, improving local–global feature fusion while reducing redundant computations. Additionally, the Edge-Enhanced Bottleneck Attention (EEBA) module integrates Sobel-based edge detection to refine boundary delineation, ensuring sharper object separation in dense vegetation environments. Extensive evaluations on publicly available datasets demonstrate the effectiveness of MSEA-Net, achieving a mean Intersection over Union (IoU) of 87.42% on the Motion-Blurred UAV Images of Sorghum Fields dataset and 71.35% on the CoFly-WeedDB dataset, outperforming benchmark models. MSEA-Net also maintains a compact architecture with only 6.74 M parameters and a model size of 25.74 MB, making it suitable for UAV-based real-time weed segmentation. These results highlight the potential of MSEA-Net for improving automated weed detection in precision agriculture while ensuring computational efficiency for edge deployment.

The escalating impacts of climate change necessitate enhanced monitoring techniques for debris covered glaciers (DCG) , a critical yet challenging subject in environmental studies. This urgency underscores the importance of developing advanced methodologies that improve precision in satellite imagery analysis. Our study introduces the Physics-Informed cascaded Swin-Unet (PICSw-UNet) model, a substantial upgrade over traditional segmentation methods such as UNet-ResNet34. By integrating domain-specific physical knowledge with deep learning, our model refines segmentation tasks effectively. Leveraging the architecture of Swin-UNet cascades, this hybrid approach incorporates physical constraints into the training process, guiding the learning algorithm towards more realistic and accurate predictions. Experimental results demonstrate the superiority of our model with improvements in key performance metrics, including Intersection over Union (IoU) and Area Under the Curve (AUC). Specifically, the Physics-Informed cascaded Swin-Unet (PICSw-UNet) achieved an IoU of 91.65\% and an AUC of 96.4\%, compared to 81.24\% and 87.6\%, respectively, for the UNet-ResNet34. Moreover, our model also shows enhanced efficiency with a reduced inference time (IT) of 0.2795 seconds. These advancements indicate promising implications for future remote sensing applications and glaciological studies, offering a promising direction for advancing image analysis methodologies.