Land surface temperature (LST) is a critical parameter of surface energy fluxes and has become the focus of numerous studies. LST downscaling is an effective technique for supplementing the limitations of the coarse-resolution LST data. However, the relationship between LST and other land surface parameters tends to be nonlinear and spatially nonstationary, due to spatial heterogeneity. Nonlinearity and spatial nonstationarity have not been considered simultaneously in previous studies. To address this issue, we propose a multi-factor geographically weighted machine learning (MFGWML) algorithm. MFGWML utilizes three excellent machine learning (ML) algorithms, namely extreme gradient boosting (XGBoost), multivariate adaptive regression splines (MARS), and Bayesian ridge regression (BRR), as base learners to capture the nonlinear relationships. MFGWML uses geographically weighted regression (GWR), which allows for spatial nonstationarity, to fuse the three base learners’ predictions. This paper downscales the 30 m LST data retrieved from Landsat 8 images to 10 m LST data mainly based on Sentinel-2A images. The results show that MFGWML outperforms two classic algorithms, namely thermal image sharpening (TsHARP) and the high-resolution urban thermal sharpener (HUTS). We conclude that MFGWML combines the advantages of multiple regression, ML, and GWR, to capture the local heterogeneity and obtain reliable and robust downscaled LST data.

Building change detection is a prominent topic in remote sensing applications. Scholars have proposed a variety of fully-convolutional-network-based change detection methods for high-resolution remote sensing images, achieving impressive results on several building datasets. However, existing methods cannot solve the problem of pseudo-changes caused by factors such as “same object with different spectrums” and “different objects with same spectrums” in high-resolution remote sensing images because their networks are constructed using simple similarity measures. To increase the ability of the model to resist pseudo-changes and improve detection accuracy, we propose an improved method based on fully convolutional network, called multitask difference-enhanced Siamese network (MDESNet) for building change detection in high-resolution remote sensing images. We improved its feature extraction ability by adding semantic constraints and effectively utilized features while improving its recognition performance. Furthermore, we proposed a similarity measure combining concatenation and difference, called the feature difference enhancement (FDE) module, and designed comparative experiments to demonstrate its effectiveness in resisting pseudo-changes. Using the building change detection dataset (BCDD), we demonstrate that our method outperforms other state-of-the-art change detection methods, achieving the highest F1-score (0.9124) and OA (0.9874), indicating its advantages for high-resolution remote sensing image building change detection tasks.

Dark ships, vessels deliberately disabling their AIS signals, constitute a grave maritime safety hazard, with detection efforts hindered by issues like over-reliance on AIS, inadequate surveillance coverage, and significant mismatch rates. This paper proposes an improved multi-feature association method that integrates satellite remote sensing and AIS data, with a focus on oriented bounding box course estimation, to improve the detection of dark ships and enhance maritime surveillance. Firstly, the oriented bounding box object detection model (YOLOv11n-OBB) is trained to break through the limitations of horizontal bounding box orientation representation. Secondly, by integrating position, dimensions (length and width), and course characteristics, we devise a joint cost function to evaluate the combined significance of multiple features. Subsequently, an advanced JVC global optimization algorithm is employed to ensure high-precision association in dense scenes. Finally, by integrating data from Gaofen-6 (optical) and Gaofen-3B (SAR) satellites, a day-and-night collaborative monitoring framework is constructed to address the blind spots of single-sensor monitoring during night-time or adverse weather conditions. Our results indicate that the detection model demonstrates a high average precision (AP50) of 0.986 on the optical dataset and 0.903 on the SAR dataset. The association accuracy of the multi-feature association algorithm is 91.74% in optical image and AIS data matching, and 91.33% in SAR image and AIS data matching. The association rate reaches 96.03% (optical) and 74.24% (SAR), respectively. This study provides an efficient technical tool for maritime safety regulation through multi-source data fusion and algorithm innovation.

Assessing the supply and demand of urban green space (UGS) ecosystem services (ESs) can provide relevant insights for urban planning. This study presents an analysis method for the spatial distribution of UGS ES supply and demand at administrative unit and 1-m grid scales and directly compares the matches of ES supply and demand in spatially explicit maps at two scales. Based on the analysis results at administrative unit scale, administrative units with an unbalanced UGS ES supply and demand were divided into three types: (Ⅰ) lack of green space; (Ⅱ) unreasonable green space structure; (Ⅲ) comprehensive, and different optimization schemes were put forward. According to the analysis results at 1-m scale, the regions with an unbalanced ES supply and demand of an administrative unit were divided into the following: (1) severe ES shortage area; (2) moderate ES shortage area; (3) mild ES shortage area, and the severe ES shortage area was taken as the UGS optimization area. We take the UGS within the 5th Ring Road of Beijing as an example and propose suggestions for optimizing the UGS pattern based on the evaluation of the supply and demand of UGS carbon sequestration services and purification services for particulate matter with an aerodynamic diameter <2.5 µm (PM_2.5). This study provides an easy-to-use evaluation method for the spatial distribution of UGS ES supply and demand and proposes different optimization suggestions for the unbalanced area, thus playing a role in UGS construction activities and green space structure optimization.

As an important ecological barrier and a crucial base for animal husbandry in China, the forage–livestock balance in the Three-River-Source Region (TRSR) directly impacts both the degradation and recovery of grassland. This study examines the forage–livestock balance in the TRSR over the past 13 years (2010–2022) by calculating both the theoretical and actual livestock carrying capacity, thereby providing a scientific basis for regional animal husbandry policies. Firstly, the Carnegie–Ames–Stanford Approach (CASA) model was improved to fit the specific characteristics of alpine grassland ecosystem in the TRSR. This enhanced model was subsequently used to calculate the net primary productivity (NPP) of the grassland, from which the regional grassland yield and theoretical livestock carrying capacity were derived. Secondly, the actual livestock carrying capacity was calculated and spatialized based on the number of regional year-end livestock. Finally, the livestock carrying pressure index was determined using both the theoretical and actual livestock carrying capacity. The results revealed several key findings: (1) The average grassland NPP in the TRSR was 145.44 gC/m², the average grassland yield was 922.7 kg/hm², and the average theoretical livestock carrying capacity was 0.55 SU/hm² from 2010 to 2022. Notably, all three metrics showed an increasing trend over the past 13 years, which indicates the rise in grassland vegetation activities. (2) The average actual livestock carrying capacity over the 13-year period was 0.46 SU/hm², showing a decreasing trend on the whole. The spatial distribution displayed a pattern of higher capacity in the east and lower in the west. (3) Throughout the 13 years, the TRSR generally maintained a forage–livestock balance, with an average livestock carrying pressure index of 0.96 (insufficient). However, the trend of livestock carrying pressure is on the rise, with serious overloading observed in the western part of Qumalai County and the northern part of Tongde County. Slight overloading was also noted in Zhiduo, Maduo, and Zeku Counties. Notably, Tanggulashan Town, Zhiduo, Qumalai, and Maduo Counties showed significant increases in livestock carrying pressure, while Zaduo County and the eastern regions experienced significant decreases. In conclusion, this study not only provides feasible technical methods for assessing and managing the forage–livestock balance in the TRSR but also contributes significantly to the sustainable development of the region’s grassland ecosystem and animal husbandry industry.

Maritime unmanned aerial vehicle (UAV) ship detection faces challenges including variations in ship pose and appearance under multiple viewpoints, occlusion and confusion in dense scenes, complex backgrounds, and the scarcity of ship datasets from UAV tilted perspectives. To overcome these obstacles, this study introduces a high-quality dataset named Marship-OBB9, comprising 11,268 drone-captured images and 18,632 instances spanning nine typical ship categories. The dataset systematically reflects the characteristics of maritime scenes under diverse scales, viewpoints, and environmental conditions. Based upon this dataset, we propose a novel detection network named YOLO11-UAVShip. First, an oriented bounding box detection mechanism is incorporated to precisely fit ship contours and reduce background interference. Second, a newly designed CK_DCNv4 module, integrating deformable convolution v4 (DCNv4) and a C3k2 backbone structure, is developed to enhance geometric feature extraction under aerial oblique view. Additionally, for ships with large aspect ratios, SGKLD effectively addresses the localization challenges in dense environments, achieving robust position regression. Comprehensive experimental evaluation demonstrates that the proposed method yields a 2.1% improvement in mAP@0.5 and a 2.3% increase in recall relative to baseline models on the Marship-OBB9 dataset. While maintaining real-time inference speed, our approach greatly enhances detection accuracy and robustness. This work provides a practical and deployable solution for intelligent ship detection in UAV imagery.