To address the challenges of low detection accuracy and weak generalization in UAV aerial imagery caused by complex ground environments, significant scale variations among targets, dense small objects, and background interference, this paper proposes an improved lightweight multi-scale small-object detection model, MBD-YOLO (MBFF module, BiMS-FPN, and Dual-Stream Head). Specifically, to enhance multi-scale feature extraction capabilities, we introduce the Multi-Branch Feature Fusion (MBFF) module, which dynamically adjusts receptive fields through parallel branches and adaptive depthwise convolutions, expanding the receptive field while preserving detail perception. We further design a lightweight Bidirectional Multi-Scale Feature Aggregation Pyramid Network (BiMS-FPN), integrating bidirectional propagation paths and a Multi-Scale Feature Aggregation (MSFA) module to mitigate feature spatial misalignment and improve small-target detection. Additionally, the Dual-Stream Head with NMS-free architecture leverages a task-aligned architecture and dynamic matching strategies to boost inference speed without compromising accuracy. Experiments on the VisDrone2019 dataset demonstrate that MBD-YOLO-n surpasses YOLOv8n by 6.3% in mAP50 and 8.2% in mAP50–95, with accuracy gains of 17.96–55.56% for several small-target categories, while increasing parameters by merely 3.1%. Moreover, MBD-YOLO-s achieves superior detection accuracy, efficiency, and generalization with only 12.1 million parameters, outperforming state-of-the-art models and proving suitable for resource-constrained embedded deployment scenarios. The superior performance of MBD-YOLO, which harmonizes high precision with low computational demand, fulfills the critical requirements for real-time deployment on resource-limited UAVs, showing great promise for applications in traffic monitoring, urban security, and agricultural surveying.

The study of vegetation net primary productivity (NPP) is essential in the Yanshan–Taihang Mountain Ecological Conservation Zone (YTECZ). Serving as an ecological security barrier for the Beijing–Tianjin–Hebei region, understanding the spatiotemporal dynamics and drivers of NPP in the YTECZ is fundamental for supporting effective sustainable development policies. Utilizing MODIS NPP, climatic data (temperature and precipitation), and the Human Footprint Index (HFP, a comprehensive metric of anthropogenic pressure), this study employed univariate linear regression, ArcGIS spatial analysis, and the Geographical Detector to investigate the spatiotemporal patterns and drivers of vegetation NPP in the YTECZ from 2004 to 2023 and to project its future trends through time series analysis. Our findings reveal a significant fluctuating upward trend in vegetation NPP over the 21-year period (mean annual increase: 4.58 g C·m⁻²), displaying a distinct spatial gradient characterized by higher values in western and northern sectors relative to eastern and southern areas. The interannual variability of vegetation NPP was primarily dominated by precipitation fluctuation, while its spatial heterogeneity was jointly driven by vapor pressure deficit (VPD) and temperature. Notably, human activities exhibited significant explanatory power on NPP’s spatial pattern, and their interaction with climatic factors (e.g., VPD) resulted in non-linear enhancements. Future projections suggest that the current increasing trend is unlikely to be sustained in the long term, indicating substantial uncertainty in vegetation carbon sequestration patterns. This study provides critical insights into vegetation response mechanisms to global change drivers, offering a scientific foundation for ecological management strategies toward sustainable development in the YTECZ.

Vegetation is an important component of an ecosystem, fulfilling various ecological functions in areas such as soil and water conservation, climate regulation, and water source maintenance. This study focuses on the Yongding River Basin as a research area. This study used vegetation indices with long time series as a data source in combination with Landsat land use data. This study applied linear trend estimation to analyze the interannual variation trend in vegetation greenness from 2002 to 2022 in the Yongding River Basin and quantitatively analyzed the impact of land use changes on vegetation greenness. The results show that, from 2002 to 2022, the vegetation greenness in the Yongding River Basin has shown an overall increasing trend. The average growth season and the maximum annual normalized difference vegetation index (NDVI) growth rates were 0.006/10a and 0.008/10a, respectively, and the area of increased vegetation greenness accounted for 90% of the total area. During the main growth season (April to October) in the Yongding River Basin, the NDVI generally showed a spatial pattern of being higher in mountainous areas and lower in water areas, with the largest coefficient of variation in vegetation in the river water areas, and the most stable vegetation in forest land. In terms of the changes in vegetation greenness, the contribution rate of arable land was between 36.73% and 38.63%, followed by grassland and forest land, with contribution rates of 26.86% to 27.11% and 23.94% to 26.43%, respectively. The total contribution rate of water areas, construction land, and unused land was around 10.18%. This study can provide a theoretical basis for environmental protection and rational land use in the Yongding River Basin.

The net primary productivity (NPP) of vegetation—a critical component of ecosystem carbon cycling and a key indicator of the quality and functionality of ecosystems—is jointly influenced by natural and anthropogenic factors. As NPP is a vital agricultural and ecological region in China, understanding the spatiotemporal dynamics and driving mechanisms of vegetation NPP in the North China Plain (NCP) has significant implications for regional sustainable development. Utilizing MODIS NPP, temperature, precipitation, and human activity data from 2003 to 2023, this study employs univariate linear regression, ArcGIS spatial analysis, and the Hurst index to investigate the spatiotemporal characteristics, driving factors, and future trends in vegetation NPP. The results indicate that vegetation NPP exhibited a fluctuating upward trend over the 21-year period, with an annual increase of 2.60 g C/m². Spatially, NPP displayed a “high in the south, low in the north” pattern. There is significant spatial heterogeneity between temperature, precipitation, and vegetation NPP in the study area, with natural factors generally exerting a greater influence than human activities; however, the coupling of human activities with other factors significantly amplify their impact. The Hurst index (mean: 0.43) revealed an anti-persistent future trend in vegetation NPP, suggesting substantial uncertainties regarding its long-term dynamics. These findings enhance our understanding of the responses of vegetation to global change and provide a scientific basis for balancing food security and ecological conservation in the NCP.

To address the challenges of low detection accuracy and weak generalization in identifying mining-induced ground fissures from UAV imagery, caused by their slender and discontinuous morphology, complex background clutter, and multi-scale surface features, this paper proposes an enhanced deep semantic segmentation model, SAG-DeepLabV3+ (with Spatial Vision Transformer, Attention mechanisms, and Adaptive Gated Fusion). Specifically, to enhance global context modeling and fine boundary delineation, we introduce a Spatial Vision Transformer (SVT) branch within the Atrous Spatial Pyramid Pooling (ASPP) module. We further employ a dual attention mechanism, sequentially combining Squeeze-and-Excitation (SE) and a Convolutional Block Attention Module (CBAM), for progressive channel and spatial feature refinement. Moreover, an Adaptive Gated Fusion (AGF) module is designed to dynamically optimize the fusion of multi-level decoder features. Experiments on a dedicated UAV-based mining fissure dataset comprising 1280 annotated images show that SAG-DeepLabV3+ achieves a state-of-the-art mean Intersection over Union (mIoU) of 79.52% (with Xception backbone) and 79.19% (with lightweight MobileNetV2 backbone), surpassing DeepLabV3+, U-Net, and PSPNet by a significant margin. Furthermore, by leveraging transfer learning (pre-training on the public CrackVision12K dataset and fine-tuning on our mining fissure dataset), the model’s mIoU is further elevated to 82.04%, demonstrating superior generalization capability. The proposed SAG-DeepLabV3+ effectively balances high accuracy with operational efficiency, fulfilling the potential demand for lightweight automated fissure monitoring under resource-limited field deployments, and lays a foundation for subsequent real-time on-site deployment verification.