Remote sensing semantic segmentation is driven by land-use monitoring, urban planning, and ecological assessment, yet progress is hampered by scarce pixel-level labels. To address this issue, we present HyperR3SNet, which is an efficient framework for remote sensing semantic segmentation that tackles data scarcity and scale variations in overhead imagery. HyperR3SNet transfers self-supervised vision foundation models (VFMs) to remote sensing, providing strong feature generalization with minimal labeled data. Building on this, a multiscale cross-attention (MSCA) module is inserted into the backbone, Vision Transformer layer, enabling the network to extract richer features across widely varying object scales. To keep the model lightweight, a matrix-factorized task head is employed, sharply reducing parameters and computation while sustaining accuracy. HyperR3SNet is the first model to integrate a hyperbolic pixel-level loss with VFM adaptation for cross-domain and low-annotation remote sensing segmentation, leveraging the exponential expansion property of hyperbolic geometry to capture latent interclass relations and preserve structural consistency under weak supervision. Evaluated on the widely adopted remote sensing segmentation datasets (e.g., iSAID, LoveDA, Potsdam, and Vaihingen), HyperR3SNet achieves mean intersection over union (mIoU) of 67.60%, 55.86%, 80.07%, and 96.46%, respectively, and on average surpasses a broad range of state-of-the-art (SOTA) methods.

In multi-resolution remote sensing imagery, roads typically exhibit sparse, elongated, and structurally complex morphological characteristics, posing formidable connectivity modeling challenges for semantic segmentation models. Existing approaches predominantly focus on pixel-level accuracy, often neglecting the topological integrity of road networks, which leads to frequent discontinuities and omissions in predicted results. To address this, this paper proposes an end-to-end road extraction framework equipped with multi-receptive field modeling and structural connectivity preservation capabilities. The model incorporates a multi-receptive-field module to capture road patterns across varying spatial scales, a connectivity-aware decoding mechanism to strengthen structural coherence, and a topology-aware loss that explicitly guides the restoration of continuous road networks during training. On the DeepGlobe-Road dataset, TopoRF-Net achieves OA 98.57%, IoU 69.76%, F1-score 82.18%, Precision 85.50%, and Recall 79.12%; on the Massachusetts dataset, TopoRF-Net similarly achieved outstanding results: OA 96.65%, IoU 59.68%, F1-score 74.75%, Precision 77.98%, and Recall 71.77%. These results conclusively demonstrate that the proposed method significantly outperforms existing approaches in both precision and connectivity metrics, whilst exhibiting favorable parameter efficiency and inference performance.

Geo-registration is a fundamental process seamlessly integrating digital information within the physical world in Mobile Augmented Reality (MAR). Achieving high precision, real-time capability, and strong adaptability in georegistration is crucial for the effective functioning of MAR applications, especially in outdoor environments. However, existing methods frequently struggle with inaccuracies in long-distance positioning and latency of pose estimation, compounded by their sensitivity to scale changes of outdoor environment. This study addresses these challenges by proposing a novel continuous and real-time MAR geo-registration method for outdoor applications. Our approach integrates real-time kinematic Global Navigation Satellite System (RTK-GNSS) fusion with geodesic equations and rotation invariance estimation. This method substantially surpasses traditional methods, achieving 0.05 m virtual-real position accuracy (approximately six times better) and under 0.2 degrees pose accuracy (nearly a fivefold improvement). Additionally, it exhibits superior robustness in complex MAR scenarios. Beyond improved accuracy, this method reduces the reliance on high-quality sensor hardware and precise calibration, making it suitable for various AR systems, including smartphones and tablets.

Digital maps, serving as pivotal instruments in the realm of geographic information services, adeptly encapsulate spatial data through refined graphics and symbols. Widely used across various domains such as smart devices, in-vehicle navigation systems, unmanned aerial vehicles (UAV), electronic gaming, and virtual reality, these maps face the ongoing challenge of delivering real-time, precise, and efficacious map rendering and interaction experiences. In response to this challenge, this study introduces a Stylized Hierarchical Symbol structure (SHS) and a Geographic Feature Color-weighted Rendering technique (GFCR). The approach is further complemented by incorporating Adaptive Hierarchy-Weighted caching (AHie), Catalan Number-based Caching access (CatNCa), and Grid-based Dynamic caching loading (GriD). Experimental results demonstrate the method's effectiveness in addressing issues related to timeliness, accuracy, and logical correctness in mobile map environments, enhancing the real-time interaction experience of mobile digital maps. Notably, this method imposes modest hardware and storage requirements, making it applicable to a wide range of scenarios including natural resource planning, intricate road landscapes, route navigation challenges, and emerging fields such as autonomous driving, virtual reality, and the metaverse.

Remote sensing indices are widely used in various fields of geoscience research. However, there are limits to how effectively the knowledge of indices can be managed or analyzed. One of the main problems is the lack of ontology models and research on indices, which makes it difficult to acquire and update knowledge in this area. Additionally, there is a lack of techniques to analyze the mathematical semantics of indices, making it difficult to directly manage and analyze their mathematical semantics. This study utilizes an ontology and mathematical semantics integration method to offer a novel knowledge graph for a remote sensing index knowledge graph (RSIKG) so as to address these issues. The proposed semantic hierarchical graph structure represents the indices of knowledge with an entity-relationship layer and a mathematical semantic layer. Specifically, ontologies in the entity-relationship layer are constructed to model concepts and relationships among indices. In the mathematical semantics layer, index formulas are represented using mathematical semantic graphs. A method for calculating similarity for index formulas is also proposed. The article describes the entire process of building RSIKG, including the extraction, storage, analysis, and inference of remote sensing index knowledge. Experiments provided in this article demonstrate the intuitive and practical nature of RSIKG for analyzing indices knowledge. Overall, the proposed methods can be useful for knowledge queries and the analysis of indices. And the present study lays the groundwork for future research on analysis techniques and knowledge processing related to remote sensing indices.

Geographic registration (geo-registration) is a crucial foundation for augmented reality (AR) map applications. However, existing methods encounter difficulties in aligning spatial data with the ground surface in complex outdoor scenarios. These challenges make it difficult to accurately estimate the geographic north orientation. Consequently, the accuracy and robustness of these methods are limited. To overcome these challenges, this paper proposes a rotation-invariant estimation method for high-precision geo-registration in AR maps. The method introduces several innovations. Firstly, it improves the accuracy of generating heading data from low-cost hardware by utilizing Real-Time Kinematic GPS and visual-inertial fusion. This improvement contributes to the increased stability and precise alignment of virtual objects in complex environments. Secondly, a fusion method combines the true-north direction vector and the gravity vector to eliminate alignment errors between geospatial data and the ground surface. Lastly, the proposed method dynamically combines the initial attitude relative to the geographic north direction with the motion-estimated attitude using visual-inertial fusion. This approach significantly reduces the requirements on sensor hardware quality and calibration accuracy, making it applicable to various AR precision systems such as smartphones and augmented reality glasses. The experimental results show that this method achieves AR geo-registration accuracy at the 0.1-degree level, which is about twice as high as traditional AR geo-registration methods. Additionally, it exhibits better robustness for AR applications in complex scenarios.

Increasingly complex vector map applications and growing multi-source spatial data pose a serious challenge to the accuracy and efficiency of vector map visualization. It is true especially for real-time and dynamic scene visualization in mobile augmented reality, with the dramatic development of spatial data sensing and the emergence of AR-GIS. Such issues can be decomposed into three issues: accurate pose representation, fast and precise topological relationships computation and high-performance acceleration methods. To solve these issues, a novel quaternion-based real-time vector map visualization approach is proposed in this paper. It focuses on precise position and orientation representation, accurate and efficient spatial relationships calculation and acceleration parallel rendering in mobile AR. First, a quaternion-based pose processing method for multi-source spatial data is developed. Then, the complex processing of spatial relationships is mapped into simple and efficient quaternion-based operations. With these mapping methods, spatial relationship operations with large computational volumes can be converted into efficient quaternion calculations, and then the results are returned to respond to the interaction. Finally, an asynchronous rendering acceleration mechanism is also presented in this paper. Experiments demonstrated that the method proposed in this paper can significantly improve vector visualization of the AR map. The new approach, when compared to conventional visualization methods, provides more stable and accurate rendering results, especially when the AR map has strenuous movements and high frequency variations. The smoothness of the user interaction experience is also significantly improved.

High-performance spatial target snapping is an essential function in 3D scene modeling and mapping that is widely used in mobile augmented reality (MAR). Spatial data snapping in a MAR system must be quick and accurate, while real-time human–computer interaction and drawing smoothness must also be ensured. In this paper, we analyze the advantages and disadvantages of several spatial data snapping algorithms, such as the 2D computational geometry method and the absolute distance calculation method. To address the issues that existing algorithms do not adequately support 3D data snapping and real-time snapping of high data volumes, we present a new adaptive dynamic snapping algorithm based on the spatial and graphical characteristics of augmented reality (AR) data snapping. Finally, the algorithm is experimented with by an AR modeling system, including the evaluation of snapping efficiency and snapping accuracy. Through the experimental comparison, we found that the algorithm proposed in this paper is substantially improved in terms of shortening the snapping time, enhancing the snapping stability, and improving the snapping accuracy of vector points, lines, faces, bodies, etc. The snapping efficiency of the algorithm proposed in this paper is 1.6 times higher than that of the traditional algorithm on average, while the data acquisition accuracy based on the algorithm in this paper is more than 6 times higher than that of the traditional algorithm on average under the same conditions, and its data accuracy is improved from the decimeter level to the centimeter level.

With the extensive application of big spatial data and the emergence of spatial computing, augmented reality (AR) map rendering has attracted significant attention. A common issue in existing solutions is that AR-GIS systems rely on different platform-specific graphics libraries on different operating systems, and rendering implementations can vary across various platforms. This causes performance degradation and rendering styles that are not consistent across environments. However, high-performance rendering consistency across devices is critical in AR-GIS, especially for edge collaborative computing. In this paper, we present a high-performance, platform-independent AR-GIS rendering engine; the augmented reality universal graphics library (AUGL) engine. A unified cross-platform interface is proposed to preserve AR-GIS rendering style consistency across platforms. High-performance AR-GIS map symbol drawing models are defined and implemented based on a unified algorithm interface. We also develop a pre-caching strategy, optimized spatial-index querying, and a GPU-accelerated vector drawing algorithm that minimizes IO latency throughout the rendering process. Comparisons to existing AR-GIS visualization engines indicate that the performance of the AUGL engine is two times higher than that of the AR-GIS rendering engine on the Android, iOS, and Vuforia platforms. The drawing efficiency for vector polygons is improved significantly. The rendering performance is more than three times better than the average performances of existing Android and iOS systems.

We propose a fundamental theorem for eco-environmental surface modelling (FTEEM) in order to apply it into the fields of ecology and environmental science more easily after the fundamental theorem for Earth’s surface system modeling (FTESM). The Beijing-Tianjin-Hebei (BTH) region is taken as a case area to conduct empirical studies of algorithms for spatial upscaling, spatial downscaling, spatial interpolation, data fusion and model-data assimilation, which are based on high accuracy surface modelling (HASM), corresponding with corollaries of FTEEM. The case studies demonstrate how eco-environmental surface modelling is substantially improved when both extrinsic and intrinsic information are used along with an appropriate method of HASM. Compared with classic algorithms, the HASM-based algorithm for spatial upscaling reduced the root-mean-square error of the BTH elevation surface by 9 m. The HASM-based algorithm for spatial downscaling reduced the relative error of future scenarios of annual mean temperature by 16%. The HASM-based algorithm for spatial interpolation reduced the relative error of change trend of annual mean precipitation by 0.2%. The HASM-based algorithm for data fusion reduced the relative error of change trend of annual mean temperature by 70%. The HASM-based algorithm for model-data assimilation reduced the relative error of carbon stocks by 40%. We propose five theoretical challenges and three application problems of HASM that need to be addressed to improve FTEEM.

在高精度曲面建模方法和地球表层系统建模基本定理研究结果基础上,演绎提出了生态环境曲面建模基本定理.以京津冀地区为案例,对基于生态环境曲面建模基本定理的空间升尺度、空间降尺度、空间插值、数据融合和模型-数据同化等算法进行了实证研究,与传统算法精度进行了比较分析.结果表明,由于基于生态环境曲面建模基本定理的各种算法综合了外蕴量信息和内蕴量信息,同时运用了理论上完善的信息综合方法,使海拔高度曲面的升尺度均方根误差至少降低了9m,年平均气温未来情景的降尺度精度至少提高16％,年平均气温过去变化趋势的数据融合精度至少提高70％,年平均降雨量过去变化趋势的空间插值精度至少提高0.2％,碳储量的模型-数据同化精度提高了40％.文章最后讨论了生态曲面建模基本定理亟待解决的五大理论问题和四大应用基础问题.

作为人工智能的代表性技术,深度学习已经成为大数据等各个领域中最具有突破性发展的新技术.深度学习的成功主要得益于其新颖的数据驱动的特征表示学习能力,这种能力成功地替代了传统建模中基于领域知识人为设计特征的方式.在这些技术推动下,人工智能技术在新一代GIS基础软件技术的研究与应用中发挥着极为重要的作用,而现有人工智能GIS(AI GIS)技术研究整体仍处于初步探索阶段,距离成熟阶段尚有较大距离.作为新一代GIS基础软件的方法和技术,AI GIS已经广泛应用在遥感数据分析、水资源研究、空间流行病学和环境健康等地学领域,与传统GIS模型相比大大提高了对非结构化的遥感或街景影像和文本的地理信息提取和特征理解能力,显示出巨大的价值和发展潜力,但现有研究对AI GIS软件技术体系的梳理和总结尚不够全面.大部分研究只关注地理空间人工智能算法的研究及其特定场景下的应用研究,而对相关的AI GIS软件技术体系关注较少.本文分析了地理智慧的几个层次,并讨论了其与AI GIS的关系,总体介绍了国内外现有人工智能技术与GIS软件相结合的发展现状,进而提出了AI GIS软件技术体系.根据AI与GIS的结合关系提出了AI GIS由地理空间智能算法、AI赋能GIS和GIS赋能AI三部分组成.此外,为深入介绍AI GIS各部分组成,本文以SuperMap为例,探讨了AI GIS软件的设计与实现.最后,探讨了AI GIS的未来发展中亟需解决的问题.本文基于AI GIS软件技术的初步探索,尝试为地理智能的基础GIS软件技术体系的构建提供理论基础,以促进人工智能技术与GIS技术的进一步融合和发展,为实现地理智能提供一个可行的研究方向.

The General Circulation Models and Coupled Model Intercomparison Project Phase 5 (CMIP5) datasets used for the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) painted future opportunities and challenges afforded by climate change with broad strokes. The model outputs incorporate substantial uncertainty due to the relatively coarse spatial scales and the complexity of the processes incorporated in these models and as a consequence, it is difficult to predict future trends and infrastructure needs in large countries like China at local and regional scales over the next century. The work reported in this article describes several statistical transfer functions that were used to downscale CMIP5 climate predictions in China. The original and new downscaled CMIP5 predictions are compared with observations from 735 meteorological stations scattered across China for the period 2006-2015 to show the various improvements achieved with downscaling. Comparing the three RCP scenarios (2.6, 4.5 and 8.5) during the period 2006-2015 with observations from 735 meteorological stations indicates that MAEs of mean annual temperature were 1.9 degrees C for China on average and that the actual temperature was under-estimated at 87% of the meteorological stations under all three scenarios. After the downscaling process using a High Accuracy Surface Modeling (HASM)-based method, the MAEs for mean annual temperature under the three scenarios were reduced to 0.62 degrees C for China on average. The MAEs of annual mean precipitation were 317.29, 315.24 and 315.49 mm under the RCP2.6, RCP4.5, and RCP8.5 scenarios, respectively for China on average and the actual precipitation was over-estimated by all three scenarios at approximately 75% of the meteorological stations. The HASM-based downscaling process meant that the MAEs for the three scenarios were reduced to 80-85 mm for China on average. The downscaled predictions are used to show how temperature and precipitation are likely to vary by region in China from 2011 to 2100. The downscaled results suggest that most of China will become warmer and wetter on average under all three scenarios over the next 30 years and provide improved information to guide the investments and actions that will be needed to improve climate change resilience across China's varied landscapes in the 21st century.

A comparison between the Coupled Model Intercomparison Project Phase 5 (CMIP5) data and observations at 735 meteorological stations indicated that mean annual temperature (MAT) was underestimated about 1.8°C while mean annual precipitation (MAP) was overestimated about 263mm in general across the whole of China. A statistical analysis of China-CMIP5 data demonstrated that MAT exhibits spatial stationarity, while MAP exhibits spatial non-stationarity. MAT and MAP data from the China-CMIP5 dataset were downscaled by combining statistical approaches with a method for high accuracy surface modeling (HASM). A statistical transfer function (STF) of MAT was formulated using minimized residuals output by HASM with an ordinary least squares (OLS) linear equation that used latitude and elevation as independent variables, abbreviated as HASM-OLS. The STF of MAP under a BOX-COX transformation was derived as a combination of minimized residuals output by HASM with a geographically weight regression (GWR) using latitude, longitude, elevation and impact coefficient of aspect as independent variables, abbreviated as HASM-GB. Cross validation, using observational data from the 735 meteorological stations across China for the period 1976 to 2005, indicates that the largest uncertainty occurred on the Tibet plateau with mean absolute errors (MAEs) of MAT and MAP as high as 4.64°C and 770.51mm, respectively. The downscaling processes of HASM-OLS and HASM-GB generated MAEs of MAT and MAP that were 67.16% and 77.43% lower, respectively across the whole of China on average, and 88.48% and 97.09% lower for the Tibet plateau.

In this paper, a combination of a novel interpolation method and a local regression method was employed to improve the estimation accuracy of monthly precipitation over China. After the normalized processing and Box-Cox transformation of the data, we used the geographically weighted regression (GWR) method to describe the spatial precipitation trend, and then interpolated the residual by using a modified high accuracy surface modeling method (HASM-PRE). A high quality database of monthly precipitation with a resolution of 1 km(2) was constructed based on the meteorological stations. Results showed that wet years and dry years appear alternatively, and trend analysis of precipitation data series from 1981 to 2010 showed that the probability of years with extreme precipitation has increased in recent years. Precipitation in winter is rather uncertain and more dynamic from year to year compared to precipitation in summer.

Surfaces of mean annual temperature and mean annual precipitation during the period from 1951 to 2010 in Jiang-Xi province (Poyang Lake Basin) of China are simulated by means of a method for high accuracy surface modeling (HASM), using data from 106 weather observation stations scattered over and around Poyang Lake Basin. Methodologically, we analyzed errors of HASM by comparing with classical methods. The analysis indicated that HASM has a much higher accuracy than the classical methods. The simulation results from HASM show that mean annual temperature was respectively 17.05 degrees C and 17.46 degrees C in the periods from 1951 to 1980 (P1) and from 1981 to 2010 (P2). Mean annual precipitation was 1602 mm and 1718 mm, respectively in the P1 and P2. In other words, climate has become warmer and wetter in recent 60 years in general. The ecosystems are very sensitive to relatively small changes in surface meteorology. Warm temperate moist forest and subtropical moist forest were the dominant HLZ types, accounted for 94.99% of total area of Jiang-Xi province. The proportion of warm temperate moist forest decreased from 30.81% to 16.84%, while subtropical moist forest increased from 64.79% to 76.73%. The mean centers of the rare HLZ types, cool temperate wet forest, cool temperate rain forest and subtropical wet forest, moved respectively 291 km, 104 km and 122 km. Warm temperate wet forest and subtropical moist forest had a little shift, moved respectively 21 km and 17 km during the period from P1 to P2. (C) 2014 Elsevier B.V. All rights reserved.

In order to better regionalize and discuss the rationality/irrationality of the spatial patterns China' food provision, food production and population data was collected and GIS spatial analysis and modeling methods were used. Multi-level spatial analysis and contrast between North and South China was carried out from three aspects: (1) Ecosystem food provision potential (EFPP). Step-by-step-modifying models were constructed to assess EFPP, parameters including solar radiation, temperature, humidity, topography, soil, and landuse. (2) Conversion ratio of the EFPP (CRFP), representing the ratio of actual food production to the EFPP. High EFPP and low CRFP means high remaining food potential for future exploration (or protecting, increasing). (3) Population pressure of food provision (PPFP). PPFP was calculated based on food production, population, nutrition ingredient, and consumption standards. High PPFP means food deficiency. Results: (I) The EFPP in South and Southeast China is much higher than in the North regions, while the CRFP is the opposite; this means the South and Southeast China has more remaining food potential to explore (or to protect). CRFP in Northeast China is the highest (81%), indicating the food provision in Northeast China is approaching its maximum potential. In the future it is not wise to rely solely on food provision increases in North China, which may aggravate some problems like water shortage and ecosystem deterioration. (2) PPFP in the South and Southeast of China is much greater than in the North and has been rising, indicating that South and Southeast China have deficiency in food supply and is more and more dependent on food transportation from North China. It is necessary to preserve the fertile and high-yielding croplands as well as reclaim new food resources in the southern and eastern to improve its food self-sufficiency. From the above results, we can derive that the "North Grain to South" (NGS) pattern of China is irrational. This is in opposition to the present pattern of NGS but consistent with some other studies of domain experts, who also claim the NGS pattern may need adjustment. (C) 2015 Published by Elsevier B.V.

Resources and environmental problems directly related to the survival and development of mankind constitute hot issues of concern in the global community. Because of the complexity and diversity of concrete is-sues, research object is usually abstracted to a complicated model which consists of several sub-models in the do-main of resources and environment research. Experts and scholars divided big models into a number of subsys-tems on the base of scientific hypotheses of complex environmental and resource system generated by a specific mechanism or statistical analysis methods. How to integrate these models effectively constitutes the chief chal-lenge of the research of resource and environment model integration. With the rapid development of complexity and diversity of geo-sciences related issues, integrated modeling methods which include traditional modeling lan-guages and user-friendly graphical-modeling have been unable to meet the need of the large-scale highly compli-cated modeling. Based on the current theories, this paper presented a formal language definition termed resource and environment model-flow (REM). REM is a useful supplement to the theory of resources and environment model integration. The build process of model-flow is described by the abstract formal symbolic language avoid-ing the concrete graphical modeling process. REM is abstracted from the nature of compound model to solve the contradiction between the flexibility and complexity of integrating model. It is suitable for building more com-plex composite models. We implemented the prototype modeling environment of REM. According to the applica-tion of the prototype system of REM, it was shown that model-flow based modeling process can build a compli-cated model by using simple symbols to express semantic meanings. The introduction of REM can provides a hard foundation for the further study of performance optimization and intelligent development about model inte-gration.

The Angstrom-Prescott formula is commonly used in climatological calculation methods of solar radiation simulation. Fitting the coefficients is carried out using linear regression and in recent years it has been found that these coefifcients have obvious spatial variability. A common solution is to divide the study area into several subregions and ift the coefifcients one by one. Here, we use ground observation data for sunshine hours and solar radiation from 1961 to 2010. Adopting extraterrestrial radiation as the initial value, Angstrom-Prescott coefifcients are obtained by Geographically Weighted Regression at a national scale. The surfaces of solar radiation are obtained on the basis of the surfaces of sunshine hours interpolated by high accuracy surface modeling and astronomical radiation;results from spatial y nonstationary and error comparison tests show that Angstrom-Prescott coefifcients have signiifcant spatial nonstationarity. Compared to existing research methods, the method presented here achieves a better simulation effect.

A fast effective method(HASM-SSOR),which is easily realized in parallel environment,is first developed to solve the low computational speed and large data storage of HASM.In addition,the meteorological data from 752 meteorological stations in China during 1951-2010 are chosen to simulate the average temperature and average precipitation using multiple linear regression equation and polynomial regression equation combined with HASM-SSOR methods.That is,the residual is interpolated using HASM-SSOR method based on the assumption of data stationary and the error introduced by non-stationary is overcome.The multiple correlation coefficients are 0.9923 for temperature and 0.9987 for precipitation.Last,trends of inter-decadal average temperature and precipitation in China from 1951 to 2010 are analyzed and the performance of HASM in climate research is verified.Results show that average temperature of China has increased and the change ratio is 0.35℃/10 a.No obvious trend is identified for average precipitation over China.The change ratio for precipitation is 9.4 mm/10 a.