International journal of applied earth observation and geoinformation : ITC journal最新文献

High spatial resolution XCO₂ data is key to investigating the mechanisms of carbon sources and sinks. However, current carbon satellites have a narrow swath and uneven observation points, making it difficult to obtain seamless and full-coverage data. We propose a novel method combining extreme gradient boosting (XGBoost) with particle swarm optimization (PSO) to construct the relationship between OCO-2 XCO₂ data and auxiliary data (i.e., vegetation, meteorological, anthropogenic emissions, and LST data), and to map the seamless monthly XCO₂ concentration in East Asia from 2015 to 2020. Validation results based on TCCON ground station data demonstrate the high accuracy of the model with an average R² of 0.93, Root Mean Square Error (RMSE) of 1.33 and Mean Absolute Percentage Error (MAPE) of 0.24 % in five sites. The results show that the average atmospheric XCO₂ concentration in East Asia shows a continuous increasing trend from 2015 to 2020, with an average annual growth rate of 2.21 ppm/yr. This trend is accompanied by clear seasonal variations, with the highest XCO₂ concentration in winter and the lowest in summer. Additionally, anthropogenic activities contributed significantly to XCO₂ concentrations, which were higher in urban areas. These findings highlight the dynamics of regional XCO₂ concentrations over time and their association with human activities. This study provides a detailed examination of XCO₂ distribution and trends in East Asia, enhancing our comprehension of atmospheric CO₂ dynamics.

Geo-text data, which combine geographical locations with textual information (e.g., geo-tagged tweets), are typically visualized using tag maps. Since tags are rich in attribute information, tag maps are an intuitive method of visualizing how attribute domains carried by tags vary across space. However, users may be interested not only in the overall spatial distribution of tags but also in exploring detailed attributes-in-space analyses, such as examining how a subclass of attribute domains is distributed globally or checking whether all attribute subclasses exhibit the same global distribution pattern. To date, the methods for representing tags with visual encoding (e.g., size, color) to extend various attributes-in-space tasks to support exploratory analysis remain unclear. In this work, we extended tag maps to support exploratory analysis by distinguishing space searching into local or global spaces and attribute domains into within or between attribute classes, supporting four types of attributes-in-space tasks: global-within, local-within, global-between, and local-between tasks. We evaluated our exploratory tag map through two case studies: investigating major disaster occurrences from 1981 to 2020 and examining the leading causes of death in 2000 and 2019 for Spain, France, Germany and Italy. We used eye-tracking and a questionnaire to evaluate our exploratory tag map for comparison. Both methods had similar self-reported usability scores in terms of aesthetics, density, layout, and legibility. However, our exploratory tag map was more effective and efficient and had a lower cognitive load.

The successful launch of the Sentinel-1 satellite in 2014 brought a large amount of free SAR images to researchers and scholars, and its application in the fields of ocean monitoring, land use change, natural disaster monitoring and emergency response is becoming increasingly mature and precise. The main applications of InSAR can be categorized into surface deformation monitoring and DEM generation. Sentinel-1 was initially designed for surface deformation monitoring; thus, there are fewer relevant studies on the use of Sentinel-1 data for DEM extraction. However, as the only SAR satellite whose data are currently free and openly available and whose data are constantly updated, it is highly important to study its sources of error in the DEM generation process and the accuracy of its products. In addition, the SAR data provided by the Sentinel-1 satellite has the advantages of high resolution, all-day, all-weather, providing a large data source for DEM production. Taking the Ankang area as an example, this paper analyzes the influence of the InSAR spatiotemporal baseline, ground cover, terrain factors, SAR imaging and other factors on the accuracy of the Sentinel-1-extracted DEM using multisource ground observation data to validate its feasibility for terrain mapping in complex terrain. Finally, we look forward to how to effectively improve the quality of Sentinel-1 DEM products to provide guidance and a reference for subsequent research on DEM extraction using Sentinel-1 SAR images and designation of Sentinel-1 C satellite’s parameters.

High-resolution Digital Elevation Models (DEMs) are essential for precise geographic analysis. However, obtaining high-resolution DEMs in regions with dense vegetation, complex terrain, or satellite imagery voids presents substantial challenges. This study introduces a deep learning approach using three-dimensional (3D) terrain features combined with Conditional Generative Adversarial Networks (CGANs) to reconstruct DEMs. The 3D terrain features, such as valley and ridge lines, exhibit topographic relief patterns and provide constraints for CGANs to reconstruct DEMs. Experiments conducted in the Loess Plateau of Shaanxi confirmed the performance of the proposed method, demonstrating marked improvements in the accuracy of DEM reconstruction compared to models based on two-dimensional (2D) terrain features. The elevation accuracy of the reconstructed DEMs by the proposed method is 5.30 m, which is higher than that of the 2D terrain features method (18.90 m) by 71.96 %. Meanwhile, the proposed method shows a 15.78 % and 17.64 % improvement in elevation accuracy and slope accuracy, respectively, when reconstructing a 5 m high-resolution DEM from a 30 m low-resolution DEM. The proposed method can be flexibly used for reconstructing, repairing, and filling voids in DEM data.

With the rapid development of LiDAR and artificial intelligence technologies, 3D point cloud semantic segmentation has become a highlight research topic. This technology is able to significantly enhance the capabilities of building information modeling, navigation and environmental perception. However, current deep learning-based methods primarily rely on voxelization or multi-layer convolution for feature extraction. These methods often face challenges in effectively differentiating between homogeneous objects or structurally adherent targets in complex real-world scenes. To this end, we propose a Graph Transformer point cloud semantic segmentation network (ASGFormer) tailored for structurally adherent objects. Firstly, ASGFormer combines Graph and Transformer to promote global correlation understanding in the graph. Secondly, spatial index and position embedding are constructed based on distance relationships and feature differences. Through a learnable mechanism, the structural weights between points are dynamically adjusted, achieving adaptive spatial structure within the graph. Finally, dummy nodes are introduced to facilitate global information storage and transmission between layers, effectively addressing the issue of information loss at the terminal nodes of the graph. Comprehensive experiments are conducted on the various real-world 3D point cloud datasets, analyzing the effectiveness of proposed ASGFormer through qualitative and quantitative evaluations. ASGFormer outperforms existing approaches with of 91.3% for OA, 78.0% for mAcc, and 72.3% for mIoU on S3DIS dataset. Moreover, ASGFormer achieves 72.8%, 45.5%, 81.6%, 70.1% mIoU on ScanNet, City-Facade, Toronto 3D and Semantic KITTI dataset, respectively. Notably, the proposed method demonstrates effective differentiation of homogeneous structurally adherent objects, further contributing to the intelligent perception and modeling of complex scenes.

Assessing the quality of UAV-HSIs (Unmanned aerial vehicle hyperspectral images) is crucial for evaluating sensor performance, identifying distortion types, and measuring data inversion accuracy. Due to the absence of reference images, UAV-HSI quality assessment leans towards no-reference image quality assessment (NR-IQA), offering versatile applications. NR-IQA methods of remote sensing images using machine learning techniques have emerged, however, NR-IQA methods for UAV-HSIs containing multi-type and multiple distortions have not been developed. This paper introduces an NR-IQA method for UAV-HSI, employing machine learning techniques. We summarize and simulate distortion types in UAV-HSIs, constructing a quality assessment dataset based on 23 original high-quality and 806 simulated degraded UAV-HSIs. Extracting 129 features encompassing texture, color, transform domain, structural, and statistical aspects, we form seven feature sets through random and filtered feature selection algorithms. Ten machine learning quality assessment models are trained using this dataset and feature sets. The results showed that the model with the highest evaluation accuracy was extra trees (ET) (R² = 0.928, RMSE = 0.326, RPD = 3.601), using feature set 1 that fuses Tamura texture, color, wavelet transform, and mean subtracted contrast normalized (MSCN) coefficient for a total of 11 features, the PLCC and SROCC of its predicted and true quality scores reached 0.963 and 0.925, respectively. In addition, the random forest (RF), gradient boosting decision tree (GBDT), generalized regression neural network (GRNN), and extreme learning machine (ELM) also had high evaluation accuracies (R² > 0.9 and RPD > 2.5). These findings underscore the applicability of our proposed machine learning-based NR-IQA method to assess the quality of the UAV-HSIs containing noise, blur, strip noise, and multiple distortions. Additionally, this study serves as a reference for selecting features and models for other hyperspectral image quality assessments.

The North Pacific Ocean plays a pivotal role as a carbon sink within the global carbon cycle. However, a comprehensive understanding of the spatiotemporal dynamics of carbon dioxide concentration and its determinants in this domain remains elusive due to its vast dimensions and the intricacies of influencing factors, with previous research on carbon dioxide partial pressure in the North Pacific Ocean also being relatively scarce. While prevalent machine learning methodologies have been extensively applied to predict the partial pressure of ocean carbon dioxide (pCO₂), their limited interpretability has impeded substantial progress in elucidating the underlying mechanisms. This study introduces a gridded spatiotemporal neural network weighted regression (GSTNNWR) model to illuminate temporal and spatial relationships among relevant environmental variables and pCO₂. The GSTNNWR model achieves high-precision and high-resolution forecasts of surface pCO₂ in the North Pacific Ocean, demonstrating commendable performance (R² = 0.863 and RMSE=15.123 µatm). Simultaneously, we obtain a quantitative characterization of how various environmental factors influence pCO2 across different temporal and spatial scales. Results show a dominant positive effect of temperature on the pCO2, with an averaged normalized coefficient of 0.28, and variability in the effects of chlorophyll and salinity on the pCO₂ at different spatial and temporal locations and temperatures, whose average normalized coefficients are −0.10 and −0.04.The findings of our study will provide insights into the mechanisms and interactions within the North Pacific carbon cycle, contributing to a better understanding of ocean carbon sink formation and the dynamic regulation of the North Pacific carbon cycle.

The hierarchical structure of geographic or urban space can be well-characterized by the concept of living structure, a term coined by Christopher Alexander. All spaces, regardless of their size, possess certain degrees of livingness that can be mathematically quantified. While previous studies have successfully quantified the livingness of small spaces such as images or artworks, the livingness of geographic space has not yet been characterized in a recursive manner. Zipf’s law has been observed in urban systems and intra-urban structures. However, whether Zipf’s law is applicable to the hierarchical substructures of geographic space has rarely been investigated. In this study, we recursively extract the substructures of geographic space using global nighttime light imagery. We quantify the livingness of global cities considering the number of substructures (S) and their inherent hierarchy (H). We further investigate the scaling properties of the extracted substructures across scales and the relationships between livingness and population for global cities. The results demonstrate that all substructures of global cities form a living structure that conforms to Zipf’s law. The degree of livingness better captures population distribution than nighttime light intensity values for the global cities. This study contributes in three aspects: First, it considers global cities as a whole to quantify spatial livingness. Second, it applies the concept of livingness to cities to better capture the spatial structure of the population using nighttime light data. Third, it introduces a novel method to recursively extract substructures from nighttime images, offering a valuable tool to investigate urban structures across multiple spatial scales.

Accurate lane maps with semantics are crucial for various applications, such as high-definition maps (HD Maps), intelligent transportation systems (ITS), and digital twins. Manual annotation of lanes is labor-intensive and costly, prompting researchers to explore automatic lane extraction methods. This paper presents an end-to-end large-scale lane mapping method that considers both lane geometry and semantics. This study represents lane markings as polylines with uniformly sampled points and associated semantics, allowing for adaptation to varying lane shapes. Additionally, we propose an end-to-end network to extract lane polylines from mobile laser scanning (MLS) data, enabling the inference of vectorized lane instances without complex post-processing. The network consists of three components: a feature encoder, a column proposal generator, and a lane information decoder. The feature encoder encodes textual and structural information of lane markings to enhance the method’s robustness to data imperfections, such as varying lane intensity, uneven point density, and occlusion-induced incomplete data. The column proposal generator generates regions of interest for the subsequent decoder. Leveraging the embedded multi-scale features from the feature encoder, the lane decoder effectively predicts lane polylines and their associated semantics without requiring step-by-step conditional inference. Comprehensive experiments conducted on three lane datasets have demonstrated the performance of the proposed method, even in the presence of incomplete data and complex lane topology. Furthermore, the datasets used in this work, including source ground points, generated bird’s eye view (BEV) images, and annotations, will be publicly available with the publication of the paper. The code and dataset will be accessible through here.

Increasing population and the consequent rise in the demand for food and water resources pose significant challenges for the future of agriculture in Egypt. Rapid large-scale agricultural expansion has occurred in the country to meet the growing demand, but agricultural loss from urban infringement and field abandonment remains prevalent. Documenting the full spectrum of changes within Egypt’s agricultural systems is crucial for developing effective land-use policies that improve food security. Here we map and estimate the areal extent of multiple types of agricultural change in Egypt (i.e., agricultural gain, agricultural abandonment, and agricultural loss from urban growth) by applying the Landsat-based detection of trends in disturbance and recovery (LandTrendr) algorithm, a widely used time series temporal segmentation algorithm. First, we used LandTrendr to identify areas of agricultural gain and loss throughout Egypt from 1987 to 2019. Second, we combined land-cover maps and the LandTrendr results to create a comprehensive land-cover change map. Lastly, we evaluated the accuracy of our findings and estimated per-class areas with quantified uncertainty using high-quality reference data. Our results reveal a notable expansion in Egypt’s agricultural land area. However, this growth is accompanied by the widespread loss of prime agricultural land, a consequence of urban development and agricultural abandonment. This study emphasizes the pressing need for the implementation of sustainable land-use policies in Egypt, particularly as climate change will exacerbate pressures on the agricultural sector in the future.