High resolution remote sensing imageries contain a wealth of information about ground object features. Semantic segmentation of high resolution remote sensing imageries has emerged as a prominent area of research. However, many existing models have become complex in pursuit of accuracy, making it challenging to achieve a balance between efficiency and precision. To address this issue, we propose LiteSeger, a lightweight semantic segmentation model that combines the strengths of convolutional neural network (CNN) and transformer architectures. Initially, the MobileViT backbone network, which integrates CNN and Transformer, is employed to rapidly extract multi-level features. Subsequently, we devised a Global Attention mechanism and a Local Attention mechanism, respectively, for low-level and high-level features, which we incorporated into the decoder as the Main Feature Fusion (MFF) module. Local Attention enhances the representation of detailed features in both spatial and channel dimensions. Global Attention strengthens global interconnectivity through long-range semantic information. Particularly, Our proposed Distance-aware Grouped Attention demonstrates a significant reduction in the computational complexity of attention computation, without compromising segmentation accuracy. In comparison to other light-weight models, our proposed LiteSeger achieves the best MIoU and Ave.F1 on both the Vaihingen and Potsdam datasets.On the Vaihingen dataset, it achieves an MIoU of 71.22 and an Ave.F1 of 82.97. On the Potsdam dataset, the MIoU and Ave.F1 reach 72.00 and 83.34, respectively. At the same time, it has significant advantages in inference efficiency. LiteSeger contains only 7.0 M parameters and under the experimental settings of this study, it achieves an inference speed of 50 FPS. The source codes are available at https://github.com/HIGISX/LiteSeger.

This study introduces a novel deep learning-based model, UniU-Net, designed to achieve the high-precision segmentation of areca palms in remote sensing imagery. UniU-Net incorporates an auxiliary encoder and a unified attention fusion module (UAFM), enhancing the model’s anti-overfitting capabilities to improve its overall segmentation performance. Specifically, the primary and auxiliary encoders, through isomorphic parallel processing, leverage the principles of structural reparameterization to enhance the model’s effective learning of areca palm features while reducing the risk of overfitting. The UAFM utilizes a spatial attention mechanism to facilitate the effective fusion of multi-scale features. This architecture enables the model to capture intricate morphological details and accurately delineate the boundaries of areca palms, even under complex and heterogeneous environmental conditions such as mixed vegetation and varying illumination. To validate the effectiveness of UniU-Net, comprehensive experiments were conducted on a specialized areca palm dataset, demonstrating superior performance compared to several state-of-the-art semantic segmentation models. The proposed method achieves significant improvements in key evaluation metrics, such as the F1-score and intersection over union (IoU), highlighting its robustness and precision in automated areca palm extraction tasks. The integration of advanced attention mechanisms not only enhances the model’s ability to focus on relevant regions but also improves the segmentation accuracy in challenging scenarios. Beyond the specific application of areca palm segmentation, the methodologies introduced in this study hold substantial practical significance for broader agricultural applications, such as precision farming and crop monitoring.