Biodiversity conservation is a critical environmental challenge, with accurate assessment being essential for conservation efforts. This study addresses the limitations of current plant diversity assessment methods, particularly in recognizing mixed and stunted grass species, by developing an enhanced species recognition approach using unmanned aerial vehicle (UAV) hyperspectral data and deep learning models in the steppe region of Xilinhot, Inner Mongolia. We compared five models—support vector machine (SVM), two-dimensional convolutional neural network (2D-CNN), three-dimensional convolutional neural network (3D-CNN), hybrid spectral CNN (HybridSN), and the improved HybridSN+—for grass species identification. The results show that SVM and 2D-CNN models have relatively poor recognition effects on mixed distribution and stunted individuals, while HybridSN and HybridSN+ models can effectively identify important grass species in the region, and the recognition accuracy of the HybridSN+ model can reach 96.45 (p < 0.05). Notably, the 3D-CNN model’s recognition performance was inferior to the HybridSN model, especially for densely populated and smaller grass species. The HybridSN+ model, optimized from the HybridSN model, demonstrated improved recognition performance for smaller grass species individuals under equivalent conditions, leading to a discernible enhancement in overall accuracy (OA). Diversity indices (Shannon–Wiener diversity, Simpson diversity, and Pielou evenness) were calculated using the identification results from the HybridSN+ model, and spatial distribution maps were generated for each index. A comparative analysis with diversity indices derived from ground survey data revealed a strong correlation and consistency, with minimal differences between the two methods. This study provides a feasible technical approach for efficient and meticulous biodiversity assessment, offering crucial scientific references for regional biodiversity conservation, management, and restoration.

As a crucial grassland management practice, mowing plays a key role in maintaining the stability, productivity, and economic value of grassland ecosystems. The development of large-scale monitoring techniques for detecting whether mowing has occurred is of significant scientific and practical importance for improving the understanding of grassland ecosystem response mechanisms and optimizing management strategies. This study focuses on the concentrated grassland area of the Xilingol League in Inner Mongolia, restricted to the SAR-covered western sub-region. All classification accuracies reported here are obtained under spatially random train/test splits and represent an upper bound; generalization to geographically disjoint blocks remains unverified. By utilizing Sentinel-1, Sentinel-2, and Landsat-8 remote sensing images during the mowing season (August to September 2023) along with field survey data, we first applied the random forest-SHAP algorithm to select the optimal features from 70 texture features and construct a multimodal remote sensing dataset. Subsequently, we proposed the MAD-Net (Multi-Modal Attention Fusion Network with Dynamic Weighting) model to fully exploit information related to mowing identification from both optical and SAR data and conducted comparative analyses with other models. The results indicate that the CNN_LSTM_Attention model, which integrates convolutional neural networks, long short-term memory networks, and convolutional block attention modules, performed best in terms of capturing spatiotemporal variations in time series NDVI data. The U-Net model achieved the highest performance on the optimized texture dataset, while the MAD-Net model, which consists of three subnetworks that target different feature data, reached an identification accuracy of 92.59% in the SAR-covered western sub-region under a spatially random train/test split. This result represents an optimistic upper bound, as generalization to geographically independent blocks has not been evaluated. Ablation studies reveal that NDVI time series is the most informative single modality, while texture and SAR features provide complementary information; the proposed dynamic weighting module outperforms conventional fusion strategies. This study provides a new perspective for the large-scale binary classification of mown vs. non-mown grassland and effectively combines multimodal remote sensing data with deep learning models. Thus, this work not only offers a comparative basis for timely and effective identification of mowed grasslands but also provides insights for formulating optimized regional grassland management policies.