论文概览 |《IJAEOG》2024.09 Vol.133（下）

-----------关注公众号----------

Characterizing the livingness of geographic space across scales using global nighttime light data

【摘要】

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.

【摘要翻译】

地理或城市空间的层次结构可以通过“生活结构”这一概念很好地表征，该术语由克里斯托弗·亚历山大提出。所有空间，无论其大小，均具有一定程度的生活性，并且这种生活性可以通过数学方法进行量化。虽然之前的研究成功量化了小空间（如图像或艺术品）的生活性，但地理空间的生活性尚未以递归方式进行表征。在城市系统和城市内部结构中，齐夫定律（Zipf's law）已被观察到。然而，齐夫定律是否适用于地理空间的层次子结构的研究却较为稀少。在本研究中，我们利用全球夜间光影像递归提取地理空间的子结构。我们在考虑子结构数量（S）及其固有层次（H）的基础上，量化全球城市的生活性。我们进一步研究提取的子结构在不同尺度上的缩放特性以及生活性与全球城市人口之间的关系。结果表明，全球城市的所有子结构形成了一个符合齐夫定律的生活结构。与夜间光强度值相比，生活性的程度更好地捕捉了全球城市的人口分布。该研究在三个方面作出了贡献：首先，它将全球城市整体视为一个整体，量化空间生活性；其次，它将生活性的概念应用于城市，以便更好地利用夜间光数据捕捉人口的空间结构；第三，它引入了一种新方法，从夜间影像中递归提取子结构，为多尺度研究城市结构提供了有价值的工具。

【doi】

https://doi.org/10.1016/j.jag.2024.104136

【作者信息】

Zheng Ren, 瑞典盖夫莱大学工程与可持续发展学院，地理信息科学系

Bin Jiang, 香港科技大学（广州）社会中心，城市治理与设计研究方向

Chris de Rijke, 瑞典盖夫莱大学工程与可持续发展学院，地理信息科学系

Stefan Seipel，瑞典乌普萨拉大学信息技术系视觉信息与交互组

PesRec: A parametric estimation method for indoor semantic scene reconstruction from a single image

【摘要】

Reconstructing semantic indoor scenes is a challenging task in augmented and virtual reality. The quality of scene reconstruction is limited by the complexity and occlusion of indoor scenes. This is due to the difficulty in estimating the spatial structure of the scene and insufficient learning for object location inference. To address these challenges, we have developed PesRec, an end-to-end multi-task scene reconstruction network for parameterizing indoor semantic information. PesRec incorporates a newly designed spatial layout estimator and a 3D object detector to effectively learn scene parameter features from RGB images. We modify an object mesh generator to enhance the robustness of reconstructing indoor occluded objects through point cloud optimization in PesRec. Using the analyzed scene parameters and spatial structure, the proposed PesRec reconstructs an indoor scene by placing object meshes scaled to 3D detection boxes in an estimated layout cuboid. Extensive experiments on two benchmark datasets demonstrate that PesRec performs exceptionally well for object reconstruction with an average chamfer distance of 5.24 × 10-3 on the Pix3D dataset including 53.61 % mAP for 3D object detection and 79.7 % 3D IoU for the estimation of layout on the commonly-used SUN RGB-D datasets. The proposed computing network breaks through the limitations caused by complex indoor scenes and occlusions, showing optimization results that improve the quality of reconstruction in the fields of augmented reality and virtual reality.

【摘要翻译】

重建语义室内场景是在增强现实和虚拟现实中的一项挑战性任务。场景重建的质量受到室内场景复杂性和遮挡的限制。这主要源于对场景空间结构的估计困难以及对物体位置推断的学习不足。为了解决这些挑战，我们开发了PesRec，这是一种端到端的多任务场景重建网络，用于参数化室内语义信息。PesRec结合了新设计的空间布局估计器和3D物体检测器，从RGB图像中有效学习场景参数特征。我们修改了物体网格生成器，通过在PesRec中进行点云优化来增强室内遮挡物体重建的鲁棒性。利用分析得到的场景参数和空间结构，PesRec通过将物体网格放置在按3D检测框缩放的估计布局立方体中来重建室内场景。在两个基准数据集上的大量实验表明，PesRec在物体重建方面表现出色，在Pix3D数据集上平均切线距离为5.24 × 10^-3，在3D物体检测中达到53.61%的mAP，在常用的SUN RGB-D数据集上布局估计的3D IoU为79.7%。该计算网络突破了复杂室内场景和遮挡造成的局限性，展示了改善增强现实和虚拟现实领域重建质量的优化结果。

【doi】

https://doi.org/10.1016/j.jag.2024.104135

【作者信息】

Xingwen Cao, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国，中国科学院大学，北京，中国，北京大学城市规划与设计学院，深圳，中国

Xueting Zheng, 南京大学生命科学学院，南京，中国

Hongwei Zheng, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国；中国科学院大学，北京，中国

Xi Chen, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国；中国科学院大学，北京，中国

Anming Bao, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国；中国科学院大学，北京，中国

Ying Liu, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国；中国科学院大学，北京，中国

Tie Liu, 新疆生态与地理研究所，干旱区生态安全与可持续发展重点实验室，中国科学院，乌鲁木齐，中国；中国科学院大学，北京，中国

Haoran Zhang, 北京大学城市规划与设计学院，深圳，中国

Muhua Zhao, 北京大学城市规划与设计学院，深圳，中国

Zichen Zhang，北京大学城市规划与设计学院，深圳，中国

Incorporating of spatial effects in forest canopy height mapping using airborne, spaceborne lidar and spatial continuous remote sensing data

在利用机载、空间激光雷达和空间连续遥感数据进行森林冠层高度映射中融入空间效应

【摘要】

Forest canopy height (FCH) is crucial for monitoring forest structure and aboveground biomass. Light detecting and ranging (LiDAR), as a promising remote sensing technology, provides various forms of data for measuring and mapping FCH. Airborne laser scanning (ALS) could accurately measure FCH at the plot-level. Spaceborne lidar system (SLS) allows for global sampling of FCH at the footprint-level. However, ALS data has limited spatial coverage, while SLS data has relatively lower estimation accuracy. To this end, we proposed a two-step FCH mapping framework by combining ALS, SLS and auxiliary data. Firstly, using the ALS-derived FCH as reference, the SLS-derived relative height metrics were calibrated at the footprint-level using a regression method. Secondly, to further address the spatial discontinuities in SLS-derived FCH maps, a site-level FCH model was built using a weighted ensemble multi-machine learning model incorporating spatial effects (WEML_SE). The calibrated footprint-level calibration FCH model was used as a reference, and multiple remote sensing data metrics were selected and subjected to important variable selection. Specifically, a spatial adjacency matrix was established based on the spatial locations of SLS footprints, and spatial feature vectors were extracted. The result indicated that the correlation coefficient between the SLS-derived FCH and the ALS-derived FCH (r = 0.39–0.73, MRE=10.6–25.9 %, and RMSE=2.58–9.37 m) improved at footprint-level (r = 0.71–0.84, MRE=7.7–18.7 %, RMSE=1.96–7.68 m). Moreover, the WEML_SE exhibited better performance (r = 0.59–0.75, MRE=8.8–14.8 %, RMSE=2.12–5.4 m) compared to the model without incorporating spatial effects (r = 0.45–0.71, MRE=9.4–15.8 %, RMSE=2.28–5.89 m). This study emphasizes the advantages of integrating spaceborne and airborne LiDAR data to construct footprint-level estimation of FCH. The proposed WEML_SE model provides new possibilities for accurately generating wall-to-wall estimates of forest biomass.

【摘要翻译】

森林冠层高度（FCH）对于监测森林结构和地上生物量至关重要。光探测和测距（LiDAR）作为一种前景可期的遥感技术，提供了多种形式的数据来测量和绘制FCH。机载激光扫描（ALS）能够在地块水平上精确测量FCH，而卫星激光系统（SLS）则允许在脚印水平上进行全球FCH采样。然而，ALS数据的空间覆盖有限，而SLS数据的估计准确性相对较低。为此，我们提出了一个两步FCH映射框架，结合了ALS、SLS和辅助数据。首先，利用ALS衍生的FCH作为参考，通过回归方法在脚印水平上校准SLS衍生的相对高度指标。其次，为了进一步解决SLS衍生的FCH图中的空间不连续性，构建了一个使用加权集成多机器学习模型（WEML_SE）并结合空间效应的场地级FCH模型。校准后的脚印级FCH模型被用作参考，选择了多个遥感数据指标并进行重要变量选择。具体来说，基于SLS脚印的空间位置建立了空间邻接矩阵，并提取了空间特征向量。结果表明，SLS衍生的FCH与ALS衍生的FCH之间的相关系数（r = 0.39–0.73，MRE=10.6–25.9%，RMSE=2.58–9.37 m）在脚印级上得到了改善（r = 0.71–0.84，MRE=7.7–18.7%，RMSE=1.96–7.68 m）。此外，WEML_SE的性能表现优于未结合空间效应的模型（r = 0.45–0.71，MRE=9.4–15.8%，RMSE=2.28–5.89 m），其结果为（r = 0.59–0.75，MRE=8.8–14.8%，RMSE=2.12–5.4 m）。本研究强调了整合空间和机载LiDAR数据在构建FCH脚印级估计中的优势。所提出的WEML_SE模型为准确生成全覆盖森林生物量估计提供了新的可能性。

【doi】

https://doi.org/10.1016/j.jag.2024.104123

【作者信息】

Wankun Min, 武汉大学资源与环境科学学院，中国武汉

Yumin Chen, 武汉大学资源与环境科学学院，中国武汉

Wenli Huang, 武汉大学资源与环境科学学院，中国武汉

John P. Wilson, 美国南加州大学空间科学研究所，洛杉矶，加州90089-0374

Hao Tang, 新加坡国立大学艺术与社会科学学院地理系，新加坡

Meiyu Guo, 香港浸会大学地理系，香港

Rui Xu，武汉大学资源与环境科学学院，中国武汉

Stability analysis of continuous operating reference stations on Vancouver Island with a permanent GPS deformation array based on GAMIT/GLOBK

基于GAMIT/GLOBK的温哥华岛连续运行参考站的稳定性分析，结合永久GPS变形阵列

【摘要】

Continuous operating reference station (CORS) networks, underpinned by GNSS technology, support various technological services including comprehensive surveying, navigation, and remote sensing. The stability of these CORS networks is crucial for maintaining the high accuracy and reliability of data services. In this paper, we focus on analyzing the GPS deformation array of Vancouver Island, Canada, and present a thorough stability analysis scheme for CORS networks that is adaptable to other areas with frequent tectonic activity. For this study, we first conducted a data quality assessment based on three primary indices — namely, multipath effect on L1 and L2 frequency, signal-to-noise ratio in both frequency and cycle slip ratio — using the Transfer Edit Quality Check (TEQC) software. Then, we achieved high-precision data processing using the GAMIT/GLOBK software, in order to quantify the changes at each station over the specific periods. Seasonality analysis was also incorporated into the pipeline. Considering the potential variances affecting the deformation array over time, the coordinates were transformed under a unified reference frame, enabling the study of changes with station coordinates from different epochs under the same systematic conditions. Additionally, the impact of plate tectonic was examined through correlation analysis. The results showed that, while the observation data from the QUAD station in the island deformation array are of poor quality, relative stability was maintained within the deformation array. The PTRF and CLRS stations exhibited vertical instability in the fall season, while the QUAD station showed a weaker instability trend. The remaining stations maintained good stability with closer stations demonstrating stronger consistency in their displacement. In terms of plate tectonic, the deformation array stations exhibit a similar southwestward displacement trend as the North American Plate in the horizontal direction, but no significant trend in the vertical direction.

【摘要翻译】

连续运行参考站（CORS）网络基于全球导航卫星系统（GNSS）技术，支持包括综合测量、导航和遥感在内的各种技术服务。这些CORS网络的稳定性对于维护数据服务的高精度和可靠性至关重要。本文集中分析了加拿大温哥华岛的GPS变形阵列，并提出了一种适用于其他频繁发生构造活动区域的CORS网络稳定性分析方案。为此，我们首先基于三个主要指标进行了数据质量评估，即对L1和L2频率的多路径效应、频率的信噪比以及周期滑移比，使用了传输编辑质量检查（TEQC）软件。然后，我们使用GAMIT/GLOBK软件实现了高精度的数据处理，以量化特定时间段内各站的变化。季节性分析也纳入了该流程中。考虑到可能影响变形阵列随时间变化的变量，坐标在统一的参考框架下进行了变换，使得在相同的系统条件下研究不同时间点站点坐标的变化。此外，通过相关分析考察了板块构造的影响。结果显示，尽管岛屿变形阵列中的QUAD站观测数据质量较差，但变形阵列内相对保持了稳定性。PTRF和CLRS站在秋季表现出垂直不稳定性，而QUAD站则显示出较弱的不稳定趋势。其余站点保持良好的稳定性，且相邻站点的位移一致性更强。在板块构造方面，变形阵列站点在水平方向上表现出与北美板块相似的西南位移趋势，但在垂直方向上没有显著趋势。

【doi】

https://doi.org/10.1016/j.jag.2024.104118

【作者信息】

Chen Liu, 同济大学测绘与地理信息学院，中国上海市四平路1239号，邮政编码200092

Xiangtong Liu, 东华理工大学测绘与地理信息工程学院，中国江西省广兰大道418号，邮政编码330013

Rong Huang, 同济大学测绘与地理信息学院，中国上海市四平路1239号，邮政编码200092；同济大学上海行星测绘与深空探索遥感重点实验室，中国上海市四平路1239号，邮政编码200092

Lingxiao Zhang, 同济大学测绘与地理信息学院，中国上海市四平路1239号，邮政编码200092

Zhen Ye, 同济大学测绘与地理信息学院，中国上海市四平路1239号，邮政编码200092；同济大学上海行星测绘与深空探索遥感重点实验室，中国上海市四平路1239号，邮政编码200092

Xiaohua Tong，同济大学测绘与地理信息学院，中国上海市四平路1239号，邮政编码200092；同济大学上海行星测绘与深空探索遥感重点实验室，中国上海市四平路1239号，邮政编码200092

Advanced Post-earthquake Building Damage Assessment: SAR Coherence Time Matrix with Vision Transformer

【摘要】

Rapid and accurate assessment of affected areas is crucial for post-earthquake rescue efforts, as earthquakes can lead to extensive damage and casualties. The post-earthquake damage assessment method based on SAR coherence is widely utilized, but issues such as inadequate consideration of decorrelation factors and underutilization of preseismic coherence can negatively impact assessment outcomes. To address these limitations and enhance accuracy while reducing false alarms, we propose a novel approach for post-earthquake building damage assessment utilizing a SAR coherence time matrix. The proposed method involves constructing time matrices by computing preseismic image coherence to maximize the utilization of preseismic coherence information. By developing a Vision Transformer model within the realm of deep learning, we aimed to extract features from these time matrices based on their unique characteristics. Through the use of predicted values obtained from the trained model to simulate coseismic coherence, a scoring metric was established as a proxy for damage. This novel method was successfully applied to evaluate the damage caused by the 2016 Italy earthquake and the 2023 Turkey earthquake, yielding improved accuracy and reduced false alarm rates. The research findings demonstrate the transferability and reliability of this method, presenting it as an accurate and dependable tool for post-earthquake building damage assessment.

【摘要翻译】

快速和准确地评估受影响地区对于地震后的救援工作至关重要，因为地震可能导致广泛的破坏和人员伤亡。基于合成孔径雷达(SAR)相干性的后地震损害评估方法被广泛使用，但诸如对去相干因素考虑不足和对震前相干性利用不充分等问题，可能对评估结果产生负面影响。为了解决这些局限性，并在提高准确性的同时减少误报率，我们提出了一种利用SAR相干性时间矩阵进行后地震建筑损害评估的新方法。该方法通过计算震前图像相干性构建时间矩阵，以最大限度地利用震前相干性信息。我们在深度学习领域开发了一个视觉变换器模型，旨在基于时间矩阵的独特特征提取特征。通过使用从训练模型中获得的预测值来模拟同震相干性，建立了一种评分指标作为损害的代理。该新方法成功应用于评估2016年意大利地震和2023年土耳其地震造成的损害，取得了更高的准确性和更低的误报率。研究结果表明该方法的可转移性和可靠性，呈现出作为后地震建筑损害评估的准确可靠工具的潜力。

【doi】

https://doi.org/10.1016/j.jag.2024.104133

【作者信息】

Yanchen Yang, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Chou Xie, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Bangsen Tian, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Yihong Guo, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Yu Zhu, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Shuaichen Bian, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Ying Yang, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Ming Zhang, 中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Yimin Ruan，中国科学院航空航天信息研究所，北京，中国

中国科学院大学，北京，中国

浙江德清卫星应用研究院目标微波特性实验室

Continuous change detection outperforms traditional post-classification change detection for long-term monitoring of wetlands

连续变化检测优于传统的后分类变化检测，适用于湿地的长期监测

【摘要】

Accurate long-term monitoring of wetlands using satellite archives is crucial for effective conservation. While new methods based on temporal profile classification have been useful for long-term monitoring of wetlands, their advantages over traditional classification methods have not yet been demonstrated. This study aimed to compare continuous change detection (using the continuous change detection and classification (CCDC) algorithm) to traditional post-classification change detection for monitoring wetland changes between 1984 and 2022 in a temperate coastal marsh (Marais Poitevin, France) from the Landsat archive. The reference dataset was collected mainly from field observations and used to train and test a random forest classifier. The accuracy of the resulting change map was then assessed for both methods using validation points collected via visual interpretation of historical aerial photographs and Landsat temporal profiles. The change map derived from CCDC had much higher unbiased overall accuracy (0.86 ± 0.02) than that derived from post-classification change detection (0.51 ± 0.03). In addition, wetland loss was much higher than wetland gain (18 % and 2 % of the area, respectively) and was due mainly to conversion of grassland to cropland and urbanization. The study demonstrated that, unlike traditional post-classification change detection, continuous change detection provides maps of wetland changes sufficiently accurate for operational use by managers. The study also confirmed the ongoing impact of agricultural intensification and artificialization on wetland degradation in Europe.

【摘要翻译】

使用卫星档案对湿地进行准确的长期监测对于有效的保护至关重要。虽然基于时间序列分类的新方法对湿地的长期监测很有帮助，但尚未证明其相较于传统分类方法的优势。本研究旨在比较连续变化检测（使用连续变化检测与分类（CCDC）算法）与传统后分类变化检测在1984年至2022年期间监测温带沿海湿地（法国波伊特文沼泽）变化的效果，数据来源于Landsat档案。参考数据集主要通过实地观察收集，并用于训练和测试随机森林分类器。然后，使用通过历史航空照片和Landsat时间序列的视觉解译收集的验证点评估两个方法生成的变化图的准确性。由CCDC生成的变化图的无偏总体准确性（0.86 ± 0.02）远高于后分类变化检测（0.51 ± 0.03）。此外，湿地损失远高于湿地增益（分别为18%和2%），主要是由于草地转为农田和城市化。本研究表明，与传统的后分类变化检测不同，连续变化检测提供的湿地变化图足够准确，可以供管理者进行日常操作使用。研究还确认了农业集约化和人工化对欧洲湿地退化的持续影响。

【doi】

https://doi.org/10.1016/j.jag.2024.104142

【作者信息】

Quentin Demarquet, 法国，雷恩大学2，LETG UMR 6554 CNRS，亨利·勒莫尔广场，35000雷恩

Sébastien Rapinel, 法国，雷恩大学2，LETG UMR 6554 CNRS，亨利·勒莫尔广场，35000雷恩

Olivier Gore, 法国，普瓦捷沼泽公共机构，里舍利街1号，85400卢松

Simon Dufour, 法国，雷恩大学2，LETG UMR 6554 CNRS，亨利·勒莫尔广场，35000雷恩

Laurence Hubert-Moy，法国，雷恩大学2，LETG UMR 6554 CNRS，亨利·勒莫尔广场，35000雷恩

Universal adversarial defense in remote sensing based on pre-trained denoising diffusion models

基于预训练去噪扩散模型的遥感中的通用对抗防御

【摘要】

Deep neural networks (DNNs) have risen to prominence as key solutions in numerous AI applications for earth observation (AI4EO). However, their susceptibility to adversarial examples poses a critical challenge, compromising the reliability of AI4EO algorithms. This paper presents a novel Universal Adversarial Defense approach in Remote Sensing Imagery (UAD-RS), leveraging pre-trained diffusion models to protect DNNs against various adversarial examples exhibiting heterogeneous adversarial patterns. Specifically, a universal adversarial purification framework is developed utilizing pre-trained diffusion models to mitigate adversarial perturbations through the introduction of Gaussian noise and subsequent purification of the perturbations from adversarial examples. Additionally, an Adaptive Noise Level Selection (ANLS) mechanism is introduced to determine the optimal noise level for the purification framework with a task-guided Fréchet Inception Distance (FID) ranking strategy, thereby enhancing purification performance. Consequently, only a single pre-trained diffusion model is required for purifying various adversarial examples with heterogeneous adversarial patterns across each dataset, significantly reducing training efforts for multiple attack settings while maintaining high performance without prior knowledge of adversarial perturbations. Experimental results on four heterogeneous RS datasets, focusing on scene classification and semantic segmentation, demonstrate that UAD-RS outperforms state-of-the-art adversarial purification approaches, providing universal defense against seven commonly encountered adversarial perturbations. Codes and the pre-trained models are available online (https://github.com/EricYu97/UAD-RS).

【摘要翻译】

深度神经网络（DNNs）在许多地球观测（AI4EO）应用中作为关键解决方案而崭露头角。然而，它们对对抗样本的脆弱性构成了一个关键挑战，影响了AI4EO算法的可靠性。本文提出了一种新颖的遥感影像通用对抗防御方法（UAD-RS），利用预训练的扩散模型来保护DNNs免受不同对抗样本的攻击，这些样本具有异质对抗模式。具体而言，开发了一种通用对抗净化框架，利用预训练的扩散模型通过引入高斯噪声并随后净化对抗样本中的扰动，从而减轻对抗扰动。此外，引入了一种自适应噪声水平选择（ANLS）机制，以确定净化框架的最佳噪声水平，并采用任务导向的Fréchet Inception Distance（FID）排名策略，从而增强净化性能。因此，只需要一个预训练的扩散模型就可以净化各数据集中具有异质对抗模式的各种对抗样本，显著减少了针对多个攻击设置的训练工作，同时在没有对抗扰动先验知识的情况下保持高性能。在四个异质遥感数据集上的实验结果显示，UAD-RS在场景分类和语义分割任务中优于最先进的对抗净化方法，为七种常见的对抗扰动提供了通用防御。代码和预训练模型可在线获取（https://github.com/EricYu97/UAD-RS）。

【doi】

https://doi.org/10.1016/j.jag.2024.104131

【作者信息】

Weikang Yu, 慕尼黑工业大学 (TUM)，德国慕尼黑，80333；德累斯顿-罗斯多夫赫尔默茨中心 (HZDR)，德国弗赖贝格，09599

Yonghao Xu, 林雪平大学计算机视觉实验室，瑞典林雪平，58183

Pedram Ghamisi，德累斯顿-罗斯多夫赫尔默茨中心 (HZDR)，德国弗赖贝格，09599

Point cloud semantic segmentation with adaptive spatial structure graph transformer

采用自适应空间结构图变换器的点云语义分割

【摘要】

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.

【摘要翻译】

随着激光雷达和人工智能技术的快速发展，3D点云语义分割已成为一个重要的研究课题。这项技术能够显著增强建筑信息建模、导航和环境感知的能力。然而，目前基于深度学习的方法主要依赖于体素化或多层卷积进行特征提取。这些方法在复杂的现实场景中，常常难以有效区分同质物体或结构相连目标。为此，我们提出了一种针对结构相连物体的图变换点云语义分割网络（ASGFormer）。首先，ASGFormer结合了图和变换器，以促进图中的全局关联理解。其次，基于距离关系和特征差异构建空间索引和位置嵌入。通过可学习的机制，点之间的结构权重动态调整，实现图中的自适应空间结构。最后，引入虚拟节点以促进层之间的全局信息存储和传输，有效解决图终端节点的信息丢失问题。在各种真实世界的3D点云数据集上进行了全面的实验，通过定性和定量评估分析所提ASGFormer的有效性。在S3DIS数据集上，ASGFormer的整体准确率（OA）达到91.3%，平均准确率（mAcc）为78.0%，而平均交并比（mIoU）为72.3%。此外，在ScanNet、City-Facade、Toronto 3D和Semantic KITTI数据集上，ASGFormer的mIoU分别为72.8%、45.5%、81.6%和70.1%。值得注意的是，所提出的方法有效区分了同质的结构相连物体，为复杂场景的智能感知和建模做出了进一步贡献。

【doi】

https://doi.org/10.1016/j.jag.2024.104105

【作者信息】

Ting Han, 中山大学，地理空间工程与科学学院，珠海，519082，中国

Yiping Chen, 中山大学，地理空间工程与科学学院，珠海，519082，中国

Jin Ma, 中山大学，地理空间工程与科学学院，珠海，519082，中国

Xiaoxue Liu, 厦门大学，智能城市传感与计算福建省重点实验室，厦门，361005，中国

Wuming Zhang, 中山大学，地理空间工程与科学学院，珠海，519082，中国

Xinchang Zhang, 广州大学，地理与遥感学院，广州，510006，中国；新疆大学，地理与遥感科学学院，乌鲁木齐，830046，中国；广东省城乡规划建设智能服务工程技术研究中心，广州，511300，中国

Huajuan Wang，珠海市测绘院，珠海，519000，中国

A hierarchical downscaling scheme for generating fine-resolution leaf area index with multisource and multiscale observations via deep learning

【摘要】

Leaf area index (LAI) is one of key variables for depicting vegetation structures in land ecosystems. Land surface models necessitate uniform LAI inputs at varying spatial scales to ensure accurate outputs at multiscale levels, however, operational satellite LAI products are acquired only at low spatial resolutions, inhibiting their application at finer spatial scales. Spatial downscaling methods are beneficial for the spatial enhancement of LAI products, and the emergence of deep learning methods has provided promising options for land surface parameter downscaling. However, the potential of deep learning has not been well explored in LAI downscaling. To address this research gap, this study designed an original hierarchical downscaling approach facilitated by generative adversarial network (GAN), transfer learning (TL), and data augmentation techniques to retrieve LAI at fine spatial resolutions, leveraging multiscale satellite images, and cascading from 500-m to 250-m and then to 30-m scales. First, an improved super-resolution GAN (ISRGAN) model was pre-trained using the GLASS LAI and MOD09Q1 products to bridge the general non-linear relationships of LAI between the 500-m and 250-m resolutions. Subsequently, limited reference LAI images were applied to fine-tune this pre-trained ISRGAN model to address the domain shift in the 250-m resolution LAI estimations. Then, the fine-tuned LAI values and the 30-m resolution LAI reference images were utilized as the ISRGAN inputs to produce fine-resolution LAI maps. Finally, the downscaled LAI values derived from the proposed approach were separately validated against reference LAI maps and field measurements across the 250-m and 30-m resolutions. Results show that the fine-tuned transfer learning technique outperforms the pre-trained ISRGAN model and GLASS LAI, with a lower RMSE (0.78) and higher R2 (0.83) at the 250-m resolution. Moreover, the proposed hierarchical downscaling framework achieves better performances for 30-m resolution LAI estimations, regardless of the validation accuracy (R2 = 0.76; RMSE=0.95) and spatiotemporal distributions, than the ISRGAN model which was directly trained by the 500-m and 30-m resolution images. This study highlights that a hierarchical downscaling is valuable for fine-resolution LAI estimations, which leverages multiscale and multisource satellite observations via deep learning.

【摘要翻译】

叶面积指数（LAI）是描绘陆地生态系统中植被结构的关键变量之一。陆地表面模型需要在不同空间尺度上提供均匀的LAI输入，以确保在多尺度水平上获得准确的输出，然而，操作性卫星LAI产品仅在低空间分辨率下获取，这限制了它们在更精细空间尺度上的应用。空间下采样方法对LAI产品的空间增强非常有益，而深度学习方法的出现为陆地表面参数的下采样提供了有前景的选择。然而，深度学习在LAI下采样中的潜力尚未得到充分探索。为了解决这一研究空白，本研究设计了一种原创的分层下采样方法，利用生成对抗网络（GAN）、迁移学习（TL）和数据增强技术来恢复细空间分辨率下的LAI，借助多尺度卫星图像，并从500米逐步降到250米，然后到30米的尺度。首先，利用GLASS LAI和MOD09Q1产品对改进的超分辨率GAN（ISRGAN）模型进行预训练，以桥接500米和250米分辨率之间LAI的一般非线性关系。随后，采用有限的参考LAI图像对该预训练的ISRGAN模型进行微调，以解决250米分辨率LAI估计中的领域转移。然后，微调后的LAI值和30米分辨率的LAI参考图像被用作ISRGAN的输入，以生成细分辨率的LAI地图。最后，从提议的方法中得出的下采样LAI值在250米和30米分辨率下分别与参考LAI地图和现场测量进行了验证。结果显示，微调的迁移学习技术在250米分辨率下的表现优于预训练的ISRGAN模型和GLASS LAI，具有更低的均方根误差（RMSE）（0.78）和更高的决定系数（R²）（0.83）。此外，所提的分层下采样框架在30米分辨率LAI估计方面的表现优于直接使用500米和30米分辨率图像训练的ISRGAN模型，无论是验证准确性（R² = 0.76；RMSE=0.95）还是时空分布。这项研究突出了分层下采样在细分辨率LAI估计中的价值，利用深度学习借助多尺度和多源卫星观测进行分析。

【doi】

https://doi.org/10.1016/j.jag.2024.104152

【作者信息】

Huaan Jin, 中国科学院山地灾害与环境研究所数字山地与遥感应用研究中心，成都 610299，中国

Yuting Qiao, 中国科学院山地灾害与环境研究所数字山地与遥感应用研究中心，成都 610299，中国；中国科学院大学，北京 100049，中国

Tian Liu, 中国重庆市大足区交通委员会，402360，中国

Xinyao Xie, 中国科学院山地灾害与环境研究所数字山地与遥感应用研究中心，成都 610299，中国

Hongliang Fang, 中国科学院大学，北京 100049，中国；中国科学院地理科学与资源研究所资源与环境信息系统重点实验室（LREIS），北京 100101，中国

Qingchun Guo, 聊城大学地理与环境学院，聊城 252000，中国

Wei Zhao，中国科学院山地灾害与环境研究所数字山地与遥感应用研究中心，成都 610299，中国

Integrating spatial modeling-assisted InSAR phase unwrapping with temporal analysis for advanced mine subsidence time series mapping

将空间建模辅助的InSAR相位解缠与时间分析结合，用于先进的矿山沉降时间序列制图

【摘要】

This study introduces an alternate spatial-temporal modeling-assisted InSAR time-series analysis method for mine subsidence mapping, aiming to address the large deformation gradients and decorrelation issues. The approach employs the iterative Modeling-Assisted Phase Unwrapping (MA-PU) algorithm for spatial phase unwrapping, and integrates it with temporal models to derive the deformation time-series. Four different inversion procedures are implemented to derive the time-series based on pixel types. The MA-PU method’s effectiveness in handling phase gradients is validated with simulations and real data analysis, showing superiority over non-model-based methods. The incorporation with temporal modeling and time-series inversion demonstrates advantages over time-series inversion alone in dealing with rank-deficiency issues within subsidence zones. The approach is applied to the West Cliff Colliery, Australia, using 23 ALOS-1 data. Results have been compared with GNSS data for validation. Obtained accuracy is approximately 16 mm, with a correlation of 0.99 between the two measurements, showing generally better performances compared to other methods. The comparison result suggest that this approach provides a more robust solution for monitoring mine subsidence in complex scenarios.

【摘要翻译】

本研究介绍了一种用于矿山沉降测图的替代时空建模辅助InSAR时间序列分析方法，旨在解决大变形梯度和去相关问题。该方法采用迭代的建模辅助相位展开（MA-PU）算法进行空间相位展开，并将其与时间模型相结合以推导变形时间序列。根据像素类型实施了四种不同的反演过程以推导时间序列。通过仿真和实际数据分析验证了MA-PU方法在处理相位梯度方面的有效性，显示出优于非模型基方法的性能。与时间建模和时间序列反演结合相比，证明其在处理沉降区秩亏问题上优于仅采用时间序列反演的方法。该方法应用于澳大利亚West Cliff煤矿，使用23景ALOS-1数据。结果与GNSS数据进行了比对验证。所得精度约为16毫米，两者的相关性为0.99，与其他方法相比表现出总体更好的性能。比较结果表明，该方法为复杂场景下的矿山沉降监测提供了更稳健的解决方案。

【doi】

https://doi.org/10.1016/j.jag.2024.104143

【作者信息】

Alex Hay-Man Ng, 广东工业大学测绘工程系，中国广州510006；新南威尔士大学土木与环境工程学院，澳大利亚悉尼，新南威尔士2052；教育部城市群环境安全与绿色发展重点实验室，广东工业大学，中国广州510006

Bangjie Wen, 广东工业大学测绘工程系，中国广州510006

Yurong Ma, 广东工业大学测绘工程系，中国广州510006

Li Guo, 自然资源部国土卫星遥感应用中心，中国北京100048

Yiwei Dai, 广东工业大学测绘工程系，中国广州510006

Hua Wang, 华南农业大学资源环境学院，中国广州510000

Linlin Ge, 新南威尔士大学土木与环境工程学院，澳大利亚悉尼，新南威尔士2052

Zheyuan Du，澳大利亚地球科学研究所，堪培拉，澳大利亚

Reconstructing high-resolution DEMs from 3D terrain features using conditional generative adversarial networks

【摘要】

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.

【摘要翻译】

高分辨率数字高程模型（DEMs）对于精确的地理分析至关重要。然而，在植被茂密、地形复杂或卫星影像存在空白的地区获取高分辨率DEMs面临巨大挑战。本研究引入了一种结合三维（3D）地形特征和条件生成对抗网络（CGANs）的深度学习方法来重建DEMs。3D地形特征，如山谷线和山脊线，展现了地形起伏的模式，并为CGANs重建DEMs提供了约束条件。在陕西黄土高原进行的实验验证了所提出方法的性能，表明与基于二维（2D）地形特征的模型相比，DEMs重建的精度有了显著提高。该方法重建的DEMs的高程精度为5.30米，比基于2D地形特征方法的18.90米提高了71.96%。同时，在从30米低分辨率DEM重建5米高分辨率DEM时，所提方法在高程精度和坡度精度上分别提高了15.78%和17.64%。该方法可灵活应用于DEMs数据的重建、修复和空白填充。

【doi】

https://doi.org/10.1016/j.jag.2024.104115

【作者信息】

Mengqi Li, 南京信息工程大学地理科学学院，中国南京

苏黎世大学地理系，瑞士苏黎世

Wen Dai, 南京信息工程大学地理科学学院，中国南京

Guojie Wang, 南京信息工程大学遥感与测绘工程学院，中国南京

Bo Wang, 南京信息工程大学遥感与测绘工程学院，中国南京

Kai Chen, 南京信息工程大学地理科学学院，中国南京

Yifei Gao, 江苏师范大学地理、测绘与规划学院，中国徐州

Solomon Obiri Yeboah Amankwah，南京信息工程大学遥感与测绘工程学院，中国南京

Predicting plants in the wild: Mapping arctic and boreal plants with UAS-based visible and near infrared reflectance spectra

【摘要】

Biophysical changes in the Arctic and boreal zones drive shifts in vegetation, such as increasing shrub cover from warming soil or loss of living mat species due to fire. Understanding current and future responses to these factors requires mapping vegetation at a fine taxonomic resolution and landscape scale. Plants vary in size and spectral signatures, which hampers mapping of meaningful functional groups at coarse spatial resolution. Fine spatial grain of remotely sensed data (<10 cm pixels) is often necessary to resolve patches of many Arctic and boreal plant groups, such as bryophytes and lichens, which are significant components of terrestrial vegetation cover. Separation of co-occurring small vegetation patches in images also requires high spectral resolution. Our goal here was to test the capabilities of UAS-based imaging spectroscopy for mapping plant functional types (PFT) using high spatial and spectral resolution data over Arctic and boreal vegetation at four sites in central Alaska. We then tested several Machine and Deep learning models of PFT cover using the reflectance spectra. The best models were very simple, balancing both bias (overfitting caused by imbalance sample sizes) and variance (fit to the independent validation data), explaining > 50 % of the independent ground cover estimation and > 84 % accuracy in estimating validation pixels. We explored the impact of spectral resolution on PFT mapping by including vegetation indices and a gradient of narrow (5 nm) to wide (50 nm) band features in our classification models across. Vegetation indices were the most important predictors for classifying PFTs, while including band features improved models, with narrow and wide bandwidths having similar importance but models with wide bandwidths performing slightly better. We conclude that Arctic and boreal PFT reflectance can be pooled across sites for mapping with relatively few labeled pixels. Underfit, simple algorithms outperformed deep learning, at least with these small sample sizes, in classifying PFTs by balancing bias and variance. Future work should aim to increase the number of labeled pixels and the detail of labels to further improve mapping taxonomic precision.

【摘要翻译】

北极和寒带地区的生物物理变化推动了植被的变化，如由于土壤变暖导致灌木覆盖增加或由于火灾导致地表生物垫物种的损失。为了理解当前和未来对这些因素的响应，需在精细的分类层次和景观尺度上对植被进行制图。植物在大小和光谱特征上存在差异，影响了在粗糙空间分辨率下对有意义的功能群的制图。遥感数据的精细空间粒度（<10厘米像素）通常对于分辨北极和寒带植物群（如苔藓植物和地衣）的斑块是必要的，这些植物是陆地植被覆盖的重要组成部分。图像中共存的小植被斑块的分离也需要高光谱分辨率。本研究的目标是测试基于无人机系统（UAS）的成像光谱技术在使用高空间和光谱分辨率数据制图北极和寒带植被功能类型（PFT）方面的能力，研究地点为阿拉斯加中部的四个站点。我们随后测试了几种基于反射光谱的机器学习和深度学习模型对PFT覆盖率的预测能力。最佳模型非常简单，平衡了样本量不平衡导致的偏差（过拟合）和独立验证数据的方差，解释了超过50%的独立地表覆盖估算，并在验证像素的估算中达到了84%以上的准确率。我们还通过在分类模型中加入植被指数和从窄（5纳米）到宽（50纳米）波段特征的梯度，探索了光谱分辨率对PFT制图的影响。植被指数是PFT分类的最重要预测因子，而引入波段特征则进一步改进了模型，窄带和宽带特征的重要性相似，但宽带模型的表现略胜一筹。我们得出的结论是，北极和寒带PFT的反射光谱可以在不同站点间汇总，只需较少标记像素即可用于制图。在分类PFT时，简单的算法通过平衡偏差和方差，至少在样本量较小的情况下，表现优于深度学习。未来的工作应增加标记像素的数量和标记的细节，以进一步提高分类精度。

【doi】

https://doi.org/10.1016/j.jag.2024.104156

【作者信息】

Peter R. Nelson, 缅因大学，美国；生态光谱实验室 – lecospec, LLC，美国

Kenneth Bundy, 缅因大学，美国

Kevaughn. Smith, 缅因大学，美国

Matt. Macander, ABR公司，美国

Catherine Chan，内布拉斯加大学，美国

Verification of the accuracy of Sentinel-1 for DEM extraction error analysis under complex terrain conditions

验证Sentinel-1在复杂地形条件下的DEM提取精度分析

【摘要】

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.

【摘要翻译】

2014年Sentinel-1卫星的成功发射为研究人员和学者带来了大量免费SAR图像，其在海洋监测、土地利用变化、自然灾害监测和应急响应等领域的应用日益成熟和精确。InSAR的主要应用可分为地表形变监测和DEM生成。Sentinel-1最初是为地表形变监测设计的，因此关于利用Sentinel-1数据进行DEM提取的研究相对较少。然而，作为目前唯一一个数据免费开放且不断更新的SAR卫星，研究其在DEM生成过程中的误差来源及其产品的精度具有重要意义。此外，Sentinel-1卫星提供的SAR数据具有高分辨率、全天候、全时段的优势，为DEM生产提供了丰富的数据源。以安康地区为例，本文使用多源地面观测数据分析了InSAR时空基线、地表覆盖、地形因素、SAR成像等因素对Sentinel-1提取的DEM精度的影响，验证了其在复杂地形中进行地形制图的可行性。最后，我们展望了如何有效提高Sentinel-1 DEM产品质量，为后续使用Sentinel-1 SAR图像提取DEM的研究以及Sentinel-1 C卫星参数的设计提供指导和参考。

【doi】

https://doi.org/10.1016/j.jag.2024.104157

【作者信息】

Shuangcheng Zhang, 长安大学地质工程与测绘学院，陕西西安 710054，中国；地理信息工程国家重点实验室，陕西西安 710054，中国；西部矿产资源与地质工程教育部重点实验室，陕西西安 710054，中国

Jie Wang, 长安大学地质工程与测绘学院，陕西西安 710054，中国

Zhijie Feng, 长安大学地质工程与测绘学院，陕西西安 710054，中国

Tao Wang, 长安大学地质工程与测绘学院，陕西西安 710054，中国

Jun Li, 长安大学地质工程与测绘学院，陕西西安 710054，中国

Ning Liu，长安大学地质工程与测绘学院，陕西西安 710054，中国

Using Physics-Encoded GeoAI to Improve the Physical Realism of Deep Learning′s Rainfall-Runoff Responses under Climate Change

使用物理编码的GeoAI提高深度学习在气候变化下的降雨径流响应的物理真实性

【摘要】

Recent research has shown that deep learning (DL) faces physical realism challenges in predicting runoff responses under climate change, mainly due to DL’s data dependence and lack of process understanding. In this study, a physics-encoded neural network model (dNN) was developed to adress this. dNN enables a fully process-based way to training and prediction by encoding process-based modeling knowledge into the DL architecture, including the water balance principle and causal linkages of catchment hydrological processes. To examine whether dNN can produce reliable runoff responses under warming scenarios, we first conducted regional training for dNN on daily runoff in 29 catchment in California. Two process-based models, EXP-HYDRO and HBV, were then developed as benchmarks. Both dNN and a pure data-driven LSTM were forced under warming scenarios, and the monthly hydrographs and total runoff ratios metrics were evaluated relative to the benchmarks. The results demonstrated: (1) For monthly hydrographs, dNN exhibited advantages in capturing cold-season runoff increase and warm-season recession than LSTM, effectively predicting the changes and trends in monthly runoff under warming scenarios; (2) For total runoff ratios, dNN predicted fewer catchments with increased runoff, indicating it can better maintain the total water budget under warming scenarios; (3) Through the synergy with physics, dNN was able to reasonably infer unobserved snowpack dynamics under warming scenarios. These results highlight the credibility and necessity of considering physics for DL in predicting runoff responses under climate change. Overall, this study provides a promising solution for considering physics in DL to further improve the process understanding in changing environments.

【摘要翻译】

最近的研究表明，深度学习（DL）在预测气候变化下的径流响应时面临物理真实性的挑战，主要是由于DL对数据的依赖性以及缺乏过程理解。在本研究中，开发了一种物理编码的神经网络模型（dNN）来解决这一问题。dNN通过将基于过程的建模知识嵌入到DL架构中，包括水量平衡原理和流域水文过程的因果联系，实现了完全基于过程的训练和预测。为检验dNN能否在变暖情景下产生可靠的径流响应，我们首先对加利福尼亚州29个流域的日径流进行区域训练。然后，开发了两个基于过程的模型EXP-HYDRO和HBV作为基准。dNN和纯数据驱动的LSTM模型均在变暖情景下进行测试，评估了月水文图和总径流比率相对于基准的表现。结果显示：（1）在月水文图方面，dNN在捕捉冷季径流增加和暖季径流衰退方面比LSTM具有优势，能够有效预测变暖情景下月径流的变化和趋势；（2）在总径流比率方面，dNN预测的径流增加的流域数量较少，表明其能更好地保持变暖情景下的水量平衡；（3）通过与物理原理的结合，dNN能够合理推断变暖情景下未观测到的积雪动态。研究结果强调了在预测气候变化下的径流响应时考虑物理学的可信度和必要性。总体而言，本研究为在DL中考虑物理过程提供了一种有前景的解决方案，有助于在不断变化的环境中进一步提升过程理解。

【doi】

https://doi.org/10.1016/j.jag.2024.104101

【作者信息】

Heng Li, 中国地质大学（北京）信息工程学院，北京，中国

Yuqian Hu, 中国地质大学（北京）信息工程学院，北京，中国

Chunxiao Zhang,中国地质大学（北京）信息工程学院，北京，中国；北京市房山区综合勘查观察研究站，自然资源部，北京，中国

Dingtao Shen, 湖北省地理过程分析与模拟重点实验室，中南财经政法大学，430079 武汉，中国；中南财经政法大学城市与环境科学学院，430079 武汉，中国

Bingli Xu, 陆军装甲兵学院信息与通信系，北京，中国

Min Chen, 虚拟地理环境重点实验室（中华人民共和国教育部），南京师范大学，江苏 南京，中国；可持续发展目标大数据国际研究中心，北京，中国；江苏省地理信息资源开发与应用协同创新中心，江苏 南京，中国

Wenhao Chu, 中国地质大学（北京）信息工程学院，北京，中国

Rongrong Li，香港中文大学太空与地球信息科学研究所，新界沙田，香港，中国

Assessment of global and regional UPD for BDS/GNSS PPP-AR at low latitudes during quiet and geomagnetic storm periods

在安静和磁暴期间评估全球和区域UPD在低纬度的BDS/GNSS PPP-AR

【摘要】

The resolution of ambiguity (AR) is of paramount importance for the precise point positioning (PPP) technique, as it enables the reduction of convergence duration and the improvement of positioning accuracy. Uncalibrated phase delay (UPD) is a key factor in ambiguity fixation. We investigate the difference between different scale UPD estimates which are obtained from MGEX and YNCORS stations in PPP-AR performance during quiet and geomagnetic storm periods. The conclusion is supported by two sets of experiments which indicate that regional UPD estimates offer higher stability and accuracy. In the quiet period, the average time to first fix (TTFF) of the regional UPD-based PPP-AR is enhanced by 35.75% compared with the global UPD-based PPP-AR, the N-direction accuracy is enhanced by 60.32%, the E-direction accuracy is decreased by 6.36%, the U-direction accuracy is decreased by 56.14%, and the convergence rate is enhanced by 7.5% to reach 100%. During the period of geomagnetic storm, the average initialization time of regional UPD-based PPP-AR is 32.05% faster than the average initialization time of global UPD-based PPP-AR. the N-direction accuracy is enhanced by 90.32%, the E-direction accuracy is decreased by 61.11%, the U-direction accuracy is enhanced by 55.18%, and the convergence rate is enhanced by 6.67%. Additionally, the TTFF of PPP-AR based on regional UPD is significantly shorter and more stable. The difference may be due to the regional UPD absorbing part of the common error caused by the geomagnetic storm, which results in a greater enhancement in the accuracy and stability of the regional UPD product during the storm.

【摘要翻译】

模糊解的解决（AR）对精确点定位（PPP）技术至关重要，因为它能够缩短收敛时间并提高定位精度。未校准相位延迟（UPD）是模糊固定的一个关键因素。我们研究了在安静期和地磁风暴期间，来自MGEX和YNCORS站点的不同规模UPD估计在PPP-AR性能中的差异。通过两组实验的结果支持了这一结论，表明区域UPD估计提供了更高的稳定性和准确性。在安静期，基于区域UPD的PPP-AR的首次修复平均时间（TTFF）比基于全球UPD的PPP-AR提高了35.75%；N方向的精度提高了60.32%；E方向的精度降低了6.36%；U方向的精度降低了56.14%；而收敛率提高了7.5%，达到100%。在地磁风暴期间，基于区域UPD的PPP-AR的平均初始化时间比基于全球UPD的PPP-AR的平均初始化时间快32.05%；N方向的精度提高了90.32%；E方向的精度降低了61.11%；U方向的精度提高了55.18%；而收敛率提高了6.67%。此外，基于区域UPD的PPP-AR的TTFF显著更短且更稳定。这一差异可能是由于区域UPD吸收了部分由地磁风暴引起的共同误差，从而在风暴期间显著提高了区域UPD产品的准确性和稳定性。

【doi】

https://doi.org/10.1016/j.jag.2024.104119

【作者信息】

Jun Tang, 昆明理工大学土地资源工程学院, 昆明, 中国

Wei Zhang, 昆明理工大学土地资源工程学院, 昆明, 中国

Yibin Yao, 武汉大学测绘与地理信息学院, 武汉, 中国

Chaoqian Xu, 武汉大学测绘与地理信息学院, 武汉, 中国

Liang Zhang, 武汉大学测绘与地理信息学院, 武汉, 中国

Youkun Wang，昆明测绘院, 昆明, 中国

Deformation mechanism-assisted deep learning architecture for predicting step-like displacement of reservoir landslide

变形机制辅助的深度学习架构，用于预测水库滑坡的阶梯位移

【摘要】

Reservoir landslides in the Three Gorges Reservoir, China, exhibit prolonged slow motion and the potential for catastrophic events due to fluctuations in reservoir levels and intense rainfall episodes. Their distinct step-like deformation characteristics, involving rapid transformation processes of different states, pose challenges for accurate early warning and prediction. Previous forecasting models have often struggled with limited accuracy. This study introduces a mechanism-assisted deep learning model, leveraging the Informer architecture, to predict prolonged step-like reservoir landslide displacement. Utilizing a 15-year continuous monitoring dataset of the Baishuihe landslide, this model investigates the landslide mechanism, identifies influencing conditions underlying the step-wise behavior, and customizes input features for the prediction model by integrating optimized variational mode decomposition and wavelet analysis. Additionally, the dynamic correlation and hysteresis analysis between triggering factors and displacement offer valuable physical insights into the model and enhance the interpretability of the model. The model is further tailored to accommodate features of the monitoring dataset associated with landslide evolution by integrating a global multi-head attention mechanism and pooling layers, enabling the capture of both globe dependencies and local critical features of the model inputs. Through rigorous model training, performance evaluation, and tuning, the proposed model efficiently predicts step-wise landslide displacement, particularly during short-term rapid transitions between creep-mutation states.

【摘要翻译】

中国三峡水库的滑坡现象表现出持续的缓慢运动和潜在的灾难性事件，这主要是由于水库水位波动和强降雨事件的影响。这些滑坡具有独特的阶梯状变形特征，涉及不同状态的快速转变过程，这给准确的预警和预测带来了挑战。以往的预测模型往往由于精度有限而难以奏效。本研究引入了一种机制辅助的深度学习模型，利用Informer架构来预测水库滑坡位移的持续阶梯状变化。该模型利用了对白水河滑坡进行的15年连续监测数据集，研究滑坡机制，识别导致阶梯状行为的影响条件，并通过集成优化的变分模态分解和小波分析，定制预测模型的输入特征。此外，触发因素与位移之间的动态相关性和滞后分析为模型提供了有价值的物理见解，并增强了模型的可解释性。该模型还通过集成全球多头注意力机制和池化层，进一步调整以适应与滑坡演变相关的监测数据集特征，从而能够捕捉模型输入的全局依赖性和局部关键特征。通过严格的模型训练、性能评估和调优，所提出的模型有效地预测了阶梯状滑坡位移，尤其是在爬行-突变状态之间的短期快速过渡期间。

【doi】

https://doi.org/10.1016/j.jag.2024.104121

【作者信息】

Yanan Jiang国家地质灾害防治与地质环境保护重点实验室，成都理工大学，610059 成都，中国；成都理工大学地球与行星科学学院，610059 成都，中国；四川省工业互联网智能监测与应用工程技术研究中心，610059 成都，中国

Linfeng Zheng, 成都理工大学地球与行星科学学院，610059 成都，中国

Qiang Xu，成都理工大学地球与行星科学学院，610059 成都，中国

Zhong Lu，南方卫理公会大学哈弗顿地球科学系，美国德克萨斯州达拉斯75275

Spatiotemporal weighted neural network reveals surface seawater pCO2 distributions and underlying environmental mechanisms in the North Pacific Ocean

时空加权神经网络揭示北太平洋海水pCO2分布及其潜在环境机制

【摘要】

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.

【摘要翻译】

北太平洋在全球碳循环中作为碳汇发挥着关键作用。然而，由于其广阔的规模和影响因素的复杂性，对该区域二氧化碳浓度及其决定因素的时空动态的全面理解仍然难以实现，先前关于北太平洋二氧化碳分压的研究也相对稀缺。尽管现有的机器学习方法已广泛应用于预测海洋二氧化碳的分压（pCO2），但其有限的可解释性妨碍了对潜在机制的深入阐释。本研究提出了一种加权回归的网格时空神经网络（GSTNNWR）模型，以阐明相关环境变量与pCO2之间的时间和空间关系。GSTNNWR模型实现了对北太平洋表面pCO2的高精度和高分辨率预测，表现出良好的性能（R² = 0.863，均方根误差 RMSE = 15.123 µatm）。同时，我们定量表征了不同环境因素如何在不同的时间和空间尺度上影响pCO2。结果显示，温度对pCO2有主导的正向影响，平均标准化系数为0.28，而氯ophyll和盐度对pCO2的影响在不同空间、时间位置和温度下有所变化，其平均标准化系数分别为−0.10和−0.04。我们的研究结果将为理解北太平洋碳循环中的机制和相互作用提供洞见，有助于更好地理解海洋碳汇的形成及北太平洋碳循环的动态调节。

【doi】

https://doi.org/10.1016/j.jag.2024.104120

【作者信息】

Yi Liu, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027

Yijun Chen, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027；浙江省地理信息科学重点实验室，中国杭州市310028

Zihang Huang, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027

Haoxuan Liang, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027

Jin Qi, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027；浙江省地理信息科学重点实验室，中国杭州市310028

Sensen Wu, 浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027；浙江省地理信息科学重点实验室，中国杭州市310028

Zhenhong Du，浙江大学地球科学学院，中国杭州市浙大路38号，邮政编码310027；浙江省地理信息科学重点实验室，中国杭州市310028

Salt marsh carbon stock estimation using deep learning with Sentinel-1 SAR of the Yangtze River estuary, China

利用Sentinel-1 SAR对中国长江入海口的盐沼碳储量估算

【摘要】

Salt marshes are pivotal in the global carbon cycle, serving as significant contributors to the blue carbon sink. Accurately estimating carbon stock in salt marshes relies on precise vegetation classification. Here, we developed the Salt Marsh Vegetation Classification Network (SVCN), a deep learning algorithm designed to classify three primary vegetation canopies (S. alterniflora, P. australis, and S. mariqueter) spanning from 2016 to 2023 over the Yangtze River estuary, China. The SVCN was initially trained using 412 vegetation samples and Sentinel-1 Synthetic Aperture Radar (SAR) data in 2018. Additionally, we trained three traditional machine learning models – Classification and Regression Trees, Random Forests, and K-Nearest Neighbors – to facilitate a comparative analysis of model performance. Leveraging the classified vegetation outcomes, we conducted salt marsh carbon stock estimations using the InVEST model. The results showed that the SVCN model outperformed the other three models, achieving an overall accuracy of 0.97. Salt marsh carbon stocks in the Yangtze River estuary exhibited an overall increasing trend from 2016 to 2023, with an average annual increase of 3.13 *103 Mg`year-1. However, there was a notable decrease of 10.36% in 2017, primarily attributed to reductions in the area covered by S. alterniflora and P. australis, which decreased by 11.18% and 10.11%, respectively. These findings highlight the potential of deep learning models and the incorporation of salt marshes in carbon stock estimates to enhance accuracy.

【摘要翻译】

盐沼在全球碳循环中扮演着关键角色，是蓝碳汇的重要贡献者。准确估计盐沼中的碳储量依赖于精确的植被分类。在此，我们开发了盐沼植被分类网络（SVCN），这是一种深度学习算法，旨在对2016年至2023年间长江口的三种主要植被冠层（S. alterniflora、P. australis 和 S. mariqueter）进行分类。SVCN最初使用412个植被样本和2018年的Sentinel-1合成孔径雷达（SAR）数据进行训练。此外，我们还训练了三种传统机器学习模型——分类与回归树、随机森林和K最近邻，以便进行模型性能的比较分析。利用分类的植被结果，我们使用InVEST模型进行了盐沼碳储量的估算。结果显示，SVCN模型的表现优于其他三个模型，整体准确率达到0.97。从2016年到2023年，长江口盐沼的碳储量呈现出整体上升的趋势，平均年增加量为3.13 × 10³ Mg·年⁻¹。然而，2017年出现了10.36%的显著下降，主要归因于S. alterniflora和P. australis覆盖面积的减少，分别下降了11.18%和10.11%。这些发现突显了深度学习模型的潜力，以及将盐沼纳入碳储量估算中的重要性，以提高准确性。

【doi】

https://doi.org/10.1016/j.jag.2024.104138

【作者信息】

Yuying Li, 华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国

Lina Yuan, 华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；华东师范大学地理空间人工智能学院，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国

Zijiang Song, 华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国

Shanshan Yu, 华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国

Xiaowen Zhang, 华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国

Min Liu，华东师范大学地理科学学院，地理信息科学教育部重点实验室，上海，200241，中国；上海市自然资源部大城市自然资源时空大数据分析与应用重点实验室，上海，200241，中国；华东师范大学地理空间人工智能学院，上海，200241，中国

Unlocking the potential of CYGNSS for pan-tropical inland water mapping through multi-source data and transformer

通过多源数据和变换解锁CYGNSS在热带内陆水域制图的潜力

【摘要】

Cyclone Global Navigation Satellite System (CyGNSS) data are widely recognized for their sensitivity to inland water bodies. However, the detection of water bodies using single CyGNSS data is subject to uncertainties, presenting challenges for large-scale and accurate water system detection. In this study, we employ CyGNSS data for regression estimation to map inland water bodies. In comparison to previous studies, we incorporate additional constraints, including topographic factors, vegetation information, soil moisture, and latitude and longitude data. Leveraging the U-shaped structure, Swin Transformer, and ContextModule, we effectively extract water body distribution information, referred to as CFRT. Through rigorous performance comparison with prevalent deep learning models, our method demonstrates remarkable accuracy. The generated water percent exhibits notable consistency with the reference data, achieving a root mean square error (RMSE) of 7.15% and a mean intersection over union of 0.778 within the reachable area of the CyGNSS data. Our approach emphasizes the significance of utilizing multi-source data to substantially enhance the accuracy of CyGNSS water system detection.

【摘要翻译】

气旋全球导航卫星系统（CyGNSS）数据因其对内陆水体的敏感性而广泛认可。然而，使用单一的CyGNSS 数据检测水体存在不确定性，这对大规模和准确的水系统检测提出了挑战。在本研究中，我们采用 CyGNSS 数据进行回归估计，以绘制内陆水体地图。与之前的研究相比，我们加入了额外的约束条件，包括地形因素、植被信息、土壤湿度以及经纬度数据。利用 U 形结构、Swin Transformer 和 ContextModule，我们有效提取了水体分布信息，称为 CFRT。通过与常见深度学习模型的严格性能比较，我们的方法表现出显著的准确性。生成的水体百分比与参考数据展现出显著一致性，在 CyGNSS 数据可达区域内实现了 7.15% 的均方根误差（RMSE）和 0.778 的平均交并比。我们的方法强调利用多源数据的重要性，从而显著提高 CyGNSS 水系统检测的准确性。

【doi】

https://doi.org/10.1016/j.jag.2024.104122

【作者信息】

Yuhan Chen, 环境科学与工程学院，南京信息工程大学，南京，210044，中国；自然资源部遥感与导航一体化应用技术创新中心，南京，210044，中国；哈尔滨工程大学青岛创新发展基地（中心），青岛，266000，中国

Qingyun Yan，环境科学与工程学院，南京信息工程大学，南京，210044，中国；自然资源部遥感与导航一体化应用技术创新中心，南京，210044，中国；南京信息工程大学遥感与测绘工程学院，南京，210044，中国

Estimating the expansion and reduction of agricultural extent in Egypt using Landsat time series

利用Landsat时间序列估算埃及农业范围的扩展和减少

【摘要】

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.

【摘要翻译】

增加的人口和随之而来的对食品和水资源的需求对埃及未来农业构成了重大挑战。

为了满足日益增长的需求，国家进行了快速的大规模农业扩张，但城市侵占和农田废弃导致的农业损失仍然普遍。记录埃及农业系统内的所有变化是制定有效的土地利用政策以改善粮食安全的关键。在本研究中，我们通过应用基于Landsat的干扰和恢复趋势检测算法（LandTrendr）绘制和估算了埃及多种类型的农业变化面积（即农业增益、农业废弃和由于城市增长导致的农业损失）。首先，我们使用LandTrendr识别了1987年至2019年间埃及农业增益和损失的区域。其次，我们结合土地覆盖图和LandTrendr的结果创建了一个全面的土地覆盖变化图。最后，我们使用高质量的参考数据评估了我们的发现的准确性，并估算了每个类别的面积和量化的不确定性。我们的结果显示，埃及的农业用地面积显著扩大。然而，这一增长伴随着优质农业用地的广泛损失，这是城市发展和农业废弃的结果。本研究强调了在埃及实施可持续土地利用政策的迫切需要，特别是气候变化将加剧对农业部门的压力。

【doi】

https://doi.org/10.1016/j.jag.2024.104141

【作者信息】

Kelsee H. Bratley, 波士顿大学地球与环境系，美国马萨诸塞州波士顿，邮政编码 02215

Curtis E. Woodcock，波士顿大学地球与环境系，美国马萨诸塞州波士顿，邮政编码 02215

Growing soil erosion risks and their role in modulating catastrophic floods in North Africa

不断增长的土壤侵蚀风险及其在调节北非灾难性洪水中的作用

【摘要】

Intensifying hydroclimatic changes in North Africa are causing unprecedented floods, droughts, and land degradation patterns that are increasingly associated with human casualties, socioeconomic instabilities, and outflow migrations. These patterns’ and their future forecasts remain largely unquantified, aggravating the impacts on several populous areas. To address this deficiency, we employ pixel-based remote sensing data correlation analysis and soil loss modeling to constrain the uncertainties on the decadal hydroclimatic and ecosystem changes in North Africa. Using cloud-based big data analysis in Google Earth Engine, we establish the convolution between precipitation patterns and surface textural characteristics, evaluating the spatial distribution of soil erosion risks at the continental scale. Our investigation uses a multi-step approach, integrating risk areas derived from soil erosion with high-resolution population data, offering critical insights into zones of different vulnerabilities. Our results unveiled a significant escalation in soil erosion anomalies over the past two decades. In particular, 15 % of the areas receiving precipitation in all of North Africa are currently at medium to high risk of soil erosion versus only 7 % in 2002. These risks are concentrated in urban areas, where each year, ∼29,000 people become highly vulnerable to these hazards, up from ∼22,000 in 2002. These increases are primarily associated with the surge in semi-unformal urban settings and the rise in rain aggressiveness and storminess. These factors, combined with the poor public perception of the imminence of these risks, create hotspots where the impacts are becoming insurmountable, as considered herein for the case of the recent catastrophic floods in Derna, Libya, used as a validation site. We conclude that increased soil erosion will modulate the impacts of upcoming catastrophic floods. As such, a pressing change in urban and land use policies in expansive areas of North Africa is called for to increase their resilience to upcoming hydroclimatic fluctuations.

【摘要翻译】

北非的水文气候变化日益加剧，导致前所未有的洪水、干旱和土地退化模式，这些现象越来越与人类伤亡、社会经济不稳定和迁移潮流相关联。这些模式及其未来预测在很大程度上尚未量化，加剧了对多个人口稠密地区的影响。为了应对这一缺陷，我们采用基于像素的遥感数据相关分析和土壤流失建模，以限制北非十年水文气候和生态系统变化的 uncertainties。通过使用 Google Earth Engine 中的云计算大数据分析，我们建立了降水模式与地表纹理特征之间的卷积，评估了大陆范围内土壤侵蚀风险的空间分布。我们的研究采用了多步骤方法，将土壤侵蚀派生的风险区域与高分辨率人口数据相结合，提供了对不同脆弱性区域的关键见解。我们的结果揭示，过去二十年土壤侵蚀异常显著上升。具体而言，目前在整个北非接受降水的区域中，有 15% 的地区处于中等到高风险的土壤侵蚀状态，而在 2002 年仅为 7%。这些风险集中在城市地区，每年大约有 29,000 人变得高度脆弱，而在 2002 年这一数字约为 22,000。这些增加主要与半非正式城市环境的激增以及降雨强度和风暴频率的上升相关。这些因素，加上公众对这些风险紧迫性的认知不足，形成了影响日益严重的热点，以利比亚德尔纳最近发生的灾难性洪水案例作为验证点。我们得出的结论是，土壤侵蚀的增加将调节即将到来的灾难性洪水的影响。因此，呼吁北非广泛地区在城市和土地使用政策方面进行紧迫的变革，以提高其对即将到来的水文气候波动的抵御能力。

【doi】

https://doi.org/10.1016/j.jag.2024.104132

【作者信息】

Adil Salhi, 阿卜杜勒马雷克·埃萨阿迪大学，地理与发展组，摩洛哥马尔蒂尔

Sara Benabdelouahab, 巴塞罗那大学，经济、环境地质与水文学组，西班牙巴塞罗那

Essam Heggy，南加州大学，维特比工程学院，美国洛杉矶，邮政编码：90089；NASA喷气推进实验室，加州理工学院，美国帕萨迪纳，邮政编码：91109

A deeply supervised vertex network for road network graph extraction in high-resolution images

一种深度监督的顶点网络，用于高分辨率影像中的道路网络图提取

【摘要】

Extracting road network maps for high-resolution remote sensing images is a critical and challenging remote sensing topic, with significant importance for traffic navigation, disaster management, autonomous driving, and urban planning. Although deep learning has demonstrated its powerful feature extraction capabilities in image processing tasks, including road extraction, existing road extraction algorithms still have some limitations. In particular, semantic segmentation-based methods often lack supervision of the connection relationships between roads and topological correctness constraints, resulting in fragmented and poorly connected road network graphs. Moreover, graph-based methods lack effective supervision strategies for vertex misdetection issues. To deal with these issues, we propose a deeply supervised vertex network (DSVNet) for road network graph extraction. First, to effectively supervise road vertices, we design a road vertex supervision module that yields improved vertex prediction accuracy. Second, to merge the benefits of segmentation-based methods and graph-based methods, we establish a parallel semantic segmentation branch based on the vertex querying task, thereby achieving enhanced road extraction accuracy. Furthermore, we introduce deformable attention into the model to boost its performance and computational efficiency. We validate the effectiveness of DSVNet on two large-scale public datasets. A large number of experimental results show that our research approach achieves the best road detection performance.

【摘要翻译】

从高分辨率遥感图像中提取道路网络地图是一个关键且具有挑战性的遥感课题，对交通导航、灾害管理、自动驾驶和城市规划具有重要意义。尽管深度学习在图像处理任务中展示了强大的特征提取能力，包括道路提取，现有的道路提取算法仍然存在一些局限性。特别是，基于语义分割的方法往往缺乏对道路连接关系的监督和拓扑正确性约束，导致道路网络图碎片化和连接不良。此外，基于图的方法在顶点错误检测问题上缺乏有效的监督策略。为了解决这些问题，我们提出了一种深度监督顶点网络（DSVNet）用于道路网络图提取。首先，为了有效监督道路顶点，我们设计了一个道路顶点监督模块，以提高顶点预测的准确性。其次，为了融合基于分割的方法和基于图的方法的优点，我们建立了一个基于顶点查询任务的并行语义分割分支，从而提高道路提取的准确性。此外，我们在模型中引入了可变形注意力机制，以提升其性能和计算效率。我们在两个大规模公共数据集上验证了DSVNet的有效性。大量实验结果表明，我们的研究方法实现了最佳的道路检测性能。

【doi】

https://doi.org/10.1016/j.jag.2024.104082

【作者信息】

Yu Zhao, 遥感与数字地球重点实验室，航天信息研究院，中国科学院，北京 100094，中国；中国科学院大学，北京 100049，中国

Zhengchao Chen, 遥感与数字地球重点实验室，航天信息研究院，中国科学院，北京 100094，中国

Zhujun Zhao, 中国城市发展规划设计咨询有限公司，北京 100120，中国

Cong Li, 遥感与数字地球重点实验室，航天信息研究院，中国科学院，北京 100094，中国

Yongqing Bai, 遥感与数字地球重点实验室，航天信息研究院，中国科学院，北京 100094，中国

Zhaoming Wu, 遥感与数字地球重点实验室，航天信息研究院，中国科学院，北京 100094，中国；中国科学院大学，北京 100049，中国

Degang Wang, 中国科学院大学，北京 100049，中国；计算光学成像技术重点实验室，航天信息研究院，中国科学院，北京 100094，中国

Pan Chen，地理空间信息中心，深圳先进技术研究院，中国科学院，深圳 518055，中国

A benchmark approach and dataset for large-scale lane mapping from MLS point clouds

【摘要】

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.

【摘要翻译】

准确的车道地图及其语义对于各种应用至关重要，如高清地图（HD Maps）、智能交通系统（ITS）和数字双胞胎。手动标注车道费时且成本高，这促使研究人员探索自动车道提取方法。本文提出了一种端到端的大规模车道映射方法，考虑了车道的几何形状和语义。该研究将车道标记表示为具有均匀采样点和相关语义的多线段，从而适应不同的车道形状。此外，我们提出了一种端到端的网络，从移动激光扫描（MLS）数据中提取车道多线段，实现向量化车道实例的推断，无需复杂的后处理。该网络由三个组件组成：特征编码器、列提案生成器和车道信息解码器。特征编码器编码车道标记的文本和结构信息，以增强该方法对数据缺陷的鲁棒性，例如车道强度变化、不均匀点密度和因遮挡导致的不完整数据。列提案生成器生成后续解码器的感兴趣区域。利用特征编码器嵌入的多尺度特征，车道解码器有效预测车道多线段及其相关语义，而无需逐步条件推断。在三个车道数据集上进行的全面实验表明，所提出方法的性能，即使在不完整数据和复杂车道拓扑的情况下也表现良好。此外，本文使用的数据集，包括源地面点、生成的鸟瞩图（BEV）图像和注释，将在论文发布时公开可用。代码和数据集将通过此处获取。

【doi】

https://doi.org/10.1016/j.jag.2024.104139

【作者信息】

Xiaoxin Mi, 武汉科技大学计算机科学与人工智能学院，湖北武汉；武汉大学测绘遥感信息工程国家重点实验室（LIESMARS），武汉，罗邮路129号，430079，湖北，中国；代尔夫特大学城市数据科学，朱利安大道134号，代尔夫特，2628 BL，荷兰

Zhen Dong, 武汉大学测绘遥感信息工程国家重点实验室（LIESMARS），武汉，罗邮路129号，430079，湖北，中国

Zhipeng Cao, 腾讯数据智能中心，北京，中国

Bisheng Yang, 武汉大学测绘遥感信息工程国家重点实验室（LIESMARS），武汉，罗邮路129号，430079，湖北，中国

Zhen Cao, 武汉大学测绘遥感信息工程国家重点实验室（LIESMARS），武汉，罗邮路129号，430079，湖北，中国

Chao Zheng, 腾讯数据智能中心，北京，中国

Jantien Stoter, 代尔夫特大学城市数据科学，朱利安大道134号，代尔夫特，2628 BL，荷兰

Liangliang Nan，代尔夫特大学城市数据科学，朱利安大道134号，代尔夫特，2628 BL，荷兰

Revealing association rules within intricate ecosystems: A spatial co-location mining method based on Geo-Eco knowledge graph

揭示复杂生态系统中的关联规则：基于Geo-Eco知识图谱的空间共定位挖掘方法

【摘要】

The analysis of association rules within ecosystems is crucial for monitoring, managing, and conserving natural resources. As widely adopted approaches for this task, geospatial methods involving spatial co-location pattern mining can reveal distribution rules and inherent associations among diverse geographical elements. Rooted in Tobler’s first law of geography, these methods focus on the impact of spatial proximity. However, apart from proximity, heterogeneity of environmental attributes such as elevation, temperature and precipitation are also essential for the formation of associations. For environmental co-location (Eco-location) pattern detection, we propose a method based on the Geo-Eco Knowledge Graph (GEKG) to mine multi-impact association rules. Firstly, we introduce the Adaptive Threshold (AT) to constrain the Delaunay triangular network, dynamically regulating adjacency relationships to generate geo-eco knowledge graph’s skeleton. For comprehensive ecosystem representation, various environmental attributes are integrated as semantic information into GEKG. In the reasoning of Eco-location patterns, we innovate beyond the traditional co-location paradigm by considering both spatial proximity and semantic similarity. Under the impact of various environmental information, sub-sets of geographically proximate entities are extracted to detect Eco-location patterns. For effective management and efficient computation, we utilize the Neo4j graph database to manage large-scale GEKG and mine Eco-location patterns with its graph search function. Experiments conducted on simulated and real-world ecological datasets show that, compared to existing techniques, our GEKG-based method can detect Eco-location patterns with greater accuracy and efficiency.

【摘要翻译】

生态系统中关联规则的分析对于监测、管理和保护自然资源至关重要。作为广泛采用的任务方法，涉及空间共位置模式挖掘的地理空间方法可以揭示不同地理元素之间的分布规则和内在关联。这些方法根植于托布勒的第一地理法则，侧重于空间邻近性的影响。然而，除了邻近性外，环境属性的异质性，例如海拔、温度和降水量，对关联的形成也至关重要。为了进行环境共位置（生态位置）模式检测，我们提出了一种基于地理-生态知识图谱（GEKG）的方法，以挖掘多影响关联规则。首先，我们引入自适应阈值（AT）来约束德劳内三角网络，动态调节邻接关系以生成地理-生态知识图谱的骨架。为了全面表示生态系统，各种环境属性被整合为语义信息纳入GEKG。在生态位置模式的推理中，我们超越了传统的共位置范式，考虑了空间邻近性和语义相似性。在各种环境信息的影响下，提取地理上接近的实体子集以检测生态位置模式。为了有效管理和高效计算，我们利用Neo4j图形数据库来管理大规模的GEKG，并利用其图形搜索功能挖掘生态位置模式。针对模拟和真实生态数据集进行的实验表明，与现有技术相比，我们基于GEKG的方法能够以更高的准确性和效率检测生态位置模式。

【doi】

https://doi.org/10.1016/j.jag.2024.104116

【作者信息】

Jinghan Wang, 武汉大学资源与环境科学学院，中国武汉430079

Guangyue Li, 武汉大学测绘与遥感信息工程国家重点实验室，中国武汉430079

Tinghua Ai，武汉大学资源与环境科学学院，中国武汉430079

DDPM-SegFormer: Highly refined feature land use and land cover segmentation with a fused denoising diffusion probabilistic model and transformer

【摘要】

The semantic segmentation of land use and land cover (LULC) is a crucial and widely employed remote sensing task. Conventional convolutional neural networks and vision transformers have been extensively utilized for LULC segmentation. However, high-resolution remote sensing images contain a wealth of spatial and color texture information, which is not fully exploited by traditional deep learning approaches. The information bottleneck of CNNs and transformers results in the loss of a significant amount of texture detail information during the feature extraction process, which further limits the performance of LULC segmentation. We present DDPM-SegFormer, a new framework that merges a denoising diffusion probabilistic model (DDPM) and vision transformer for LULC segmentation. The aim is to address the difficulties arising from extraction in complex geographic landscapes and to alleviate information bottlenecks. The framework utilizes the ability of a DDPM to generate refined semantic features and that of vision transformer to model the global image context. Our framework introduces two main innovations. First, we use a DDPM for the first time in LULC segmentation to generate highly refined multiscale semantic features. This approach alleviates the information bottleneck caused by relying solely on a CNN or transformer architecture. Second, we develop an effective feature-level fusion strategy that utilizes multihead cross-attention between the DDPM and Transformer. This approach achieves the harmonious fusion of fine-scale semantic features, generating continuous and highly refined semantic features that enhance the segmentation accuracy. The results indicate that DDPM-SegFormer achieves an MIOU of 83.72% and an F1-score of 90.97% for the large-scale LoveDA dataset and an MIOU of 90.91% and an F1score of 93.30% for the Tarim Basin LULC dataset in a desert scenario. The research demonstrated that the refined and continuous semantic features produced by DDPM-SegFormer can significantly enhance LULC segmentation performance.

【摘要翻译】

土地利用和土地覆盖（LULC）的语义分割是一项重要且广泛应用的遥感任务。传统的卷积神经网络（CNN）和视觉变换器（vision transformers）已被广泛用于LULC分割。然而，高分辨率遥感图像包含大量空间和色彩纹理信息，而传统的深度学习方法并未充分利用这些信息。CNN和变换器的信息瓶颈导致在特征提取过程中大量纹理细节信息的丢失，进一步限制了LULC分割的性能。我们提出了DDPM-SegFormer，这是一个将去噪扩散概率模型（DDPM）和视觉变换器结合起来的新框架，用于LULC分割。其目的是解决在复杂地理景观中提取时所面临的困难，并缓解信息瓶颈。该框架利用DDPM生成精细语义特征的能力，以及视觉变换器建模全局图像上下文的能力。我们的框架引入了两个主要创新。首先，我们首次在LULC分割中使用DDPM生成高度精细的多尺度语义特征。这种方法缓解了仅依赖CNN或变换器架构所造成的信息瓶颈。其次，我们开发了一种有效的特征级融合策略，利用DDPM与变换器之间的多头交叉注意力（multihead cross-attention）。这种方法实现了精细尺度语义特征的和谐融合，生成连续且高度精细的语义特征，从而提高了分割精度。结果表明，DDPM-SegFormer在大规模LoveDA数据集上实现了83.72%的平均交并比（MIOU）和90.97%的F1分数；在沙漠场景下的塔里木盆地LULC数据集上实现了90.91%的MIOU和93.30%的F1分数。研究表明，DDPM-SegFormer产生的精细和连续的语义特征可以显著提升LULC分割的性能。

【doi】

https://doi.org/10.1016/j.jag.2024.104093

【作者信息】

Junfu Fan, 山东科技大学土木工程与测绘学院，中国山东省淄博市255000；中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室，北京市100101

Zongwen Shi, 山东科技大学土木工程与测绘学院，中国山东省淄博市255000

Zhoupeng Ren, 中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室，北京市100101

Yuke Zhou, 中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室，北京市100101

Min Ji，山东科技大学测绘与地理信息学院，中国青岛市266510

A local enhanced mamba network for hyperspectral image classification

一种用于高光谱图像分类的局部增强的曼巴网络

【摘要】

Deep learning has significantly advanced hyperspectral image (HSI) classification, primarily due to its robust nonlinear feature extraction capabilities. The vision transformer has achieved notable performance but is limited by the quadratic computational burden of its self-attention mechanism. Recently, a network based on state space model named Mamba, has attracted considerable attention for its linear complexity and commendable performance. Nevertheless, Mamba was originally designed for one-dimensional causal sequence modeling, and its effectiveness in inherent non-causal HSI classification remains to be fully validated. To address this issue, we propose a novel Local Enhanced Mamba (LE-Mamba) network for hyperspectral image classification, which mainly comprises a Local Enhanced Spatial SSM (LES-S6), a Central Region Spectral SSM (CRS-S6), and a Multi-Scale Convolutional Gated Unit (MSCGU). The LES-S6 improves non-causal local feature extraction by incorporating a multi-directional local spatial scanning mechanism. Additionally, the CRS-S6 employs a bidirectional scanning mechanism in the spectral dimension to capture fine spectral details and integrate them with spatial information. The MSCGU utilizes a convolutional gating mechanism to aggregate features from diverse scanning routes and extract high-level semantic information. The overall accuracies of LE-Mamba on Indian Pines, WHU-Hi-HanChuan, WHU-Hi-LongKou, and Pavia University datasets are 99.16 %, 98.16 %, 99.57 %, and 99.63 %, respectively. Extensive experimental results on these four public datasets demonstrate that the LE-Mamba outperforms eight mainstream deep learning models in classification performance.

【摘要翻译】

深度学习在高光谱图像（HSI）分类方面取得了显著进展，主要得益于其强大的非线性特征提取能力。视觉变换器在此领域表现出色，但其自注意力机制的二次计算负担限制了其应用。最近，基于状态空间模型的网络Mamba因其线性复杂性和优异性能引起了广泛关注。然而，Mamba最初是为一维因果序列建模设计的，其在固有非因果高光谱图像分类中的有效性尚待充分验证。为了解决这一问题，我们提出了一种新颖的局部增强Mamba（LE-Mamba）网络用于高光谱图像分类，主要由局部增强空间SSM（LES-S6）、中心区域光谱SSM（CRS-S6）和多尺度卷积门控单元（MSCGU）组成。LES-S6通过引入多方向局部空间扫描机制来改善非因果局部特征提取。此外，CRS-S6在光谱维度上采用双向扫描机制，以捕捉精细的光谱细节并将其与空间信息结合。MSCGU利用卷积门控机制聚合来自不同扫描路径的特征，并提取高级语义信息。LE-Mamba在印度松树、WHU-Hi-HanChuan、WHU-Hi-LongKou和帕维亚大学数据集上的整体准确率分别为99.16%、98.16%、99.57%和99.63%。在这四个公共数据集上的大量实验结果表明，LE-Mamba在分类性能上优于八种主流深度学习模型。

【doi】

https://doi.org/10.1016/j.jag.2024.104092

【作者信息】

Chuanzhi Wang, 航空科学与工程学院，北京航空航天大学，中国，北京 100191

Jun Huang, 航空科学与工程学院，北京航空航天大学，中国，北京 100191

Mingyun Lv, 航空科学与工程学院，北京航空航天大学，中国，北京 100191

Huafei Du, 航空科学与工程学院，北京航空航天大学，中国，北京 100191

Yongmei Wu, 航空科学与工程学院，北京航空航天大学，中国，北京 100191

Ruiru Qin，航空科学与工程学院，北京航空航天大学，中国，北京 100191

MFI: A mudflat index based on hyperspectral satellite images for mapping coastal mudflats

MFI：基于高光谱卫星图像的泥滩指数，用于制图沿海泥滩

【摘要】

China’s coastal mudflats, threatened by artificial reclamation and climate change, are undergoing drastic changes and their accurate mapping is important for their conservation and restoration. Traditional classification methods, which require large samples and complex classifiers, tend to have low computational efficiency and poor generalization ability; thus, they are unsuitable for the rapid and accurate extraction of coastal mudflats. This study proposes a Mudflat Index (MFI) based on hyperspectral images. MFI amplifies the difference in spectral characteristics between mudflats and other land cover types in intertidal environments, effectively improving the discrimination between coastal mudflats, salt marshes, mangroves, and muddy waters. Four typical coastal mudflat areas (i.e., the Yellow River Delta in Shandong, the Radial Sand Ridges of the South Yellow Sea in Jiangsu, Hangzhou Bay in Zhejiang, and the Qinzhou Bay-Nanliu River Estuary in Guangxi) based on ZY1-02D were selected as the study areas. The extraction accuracies in the four study areas are 97.60%, 96.88%, 97.16% and 96.97%, respectively. The further extraction experiments were calculated based on hyperspectral data from GF-5, PRISMA, and Hyperion. Sample datasets were produced using field surveys and Google Earth high-resolution imagery. Compared to the Hyperspectral Bare Soil Index (HBSI), Normalized Difference Bare Soil Index (NDBSI) and Microphytobenthos Index (MPBI), MFI demonstrates superior performance with average SDI value improvements of 0.82, 0.71 and 1.17, respectively, in distinguishing mudflats from other typical land cover types in the intertidal zone. The extraction results were also compared with those derived from Support Vector Machine (SVM) and Random Forest (RF) classifications, showing that MFI outperformed SVM and RF by an average of 1.52% and 0.58%. The results show that MFI can be applied to different hyperspectral remote sensing images and different areas of mudflat extraction. The MFI-based method is simple, fast and accurate at extracting the mudflat in the intertidal environment.

【摘要翻译】

中国的沿海泥滩正受到人工填海和气候变化的威胁，正在经历剧烈变化，因此准确绘制它们的分布图对于保护和恢复至关重要。传统的分类方法需要大量样本和复杂的分类器，通常计算效率低，泛化能力差，因此不适合快速准确地提取沿海泥滩。本研究基于高光谱图像提出了一种泥滩指数（MFI）。MFI放大了泥滩与潮间带环境中其他土地覆盖类型之间的光谱特征差异，有效改善了沿海泥滩、盐沼、红树林和泥水之间的区分能力。选择了四个典型的沿海泥滩区域（即山东的黄河三角洲、江苏南黄海的放射沙脊、浙江的杭州湾和广西的钦州湾-南流江口）作为研究区域，基于ZY1-02D进行研究。这四个研究区域的提取准确率分别为97.60%、96.88%、97.16%和96.97%。进一步的提取实验是基于GF-5、PRISMA和Hyperion的高光谱数据计算的。样本数据集是通过实地调查和谷歌地球高分辨率影像生成的。与高光谱裸土指数（HBSI）、归一化差异裸土指数（NDBSI）和微藻底栖指数（MPBI）相比，MFI在区分泥滩与潮间带中其他典型土地覆盖类型方面表现优越，平均SDI值改善分别为0.82、0.71和1.17。提取结果还与支持向量机（SVM）和随机森林（RF）分类得到的结果进行了比较，显示MFI的平均性能优于SVM和RF，分别提升了1.52%和0.58%。结果表明，MFI可以应用于不同的高光谱遥感图像和不同区域的泥滩提取。基于MFI的方法在潮间带环境中提取泥滩简单、快速且准确。

【doi】

https://doi.org/10.1016/j.jag.2024.104140

【作者信息】

Gang Yang, 宁波大学地理与空间信息技术系，中国宁波 315211；中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室，中国北京 100101；宁波大学东海研究所，中国宁波 3152112；宁波大学土地与海洋空间利用与治理研究协同创新中心，中国宁波 315211

Chunchen Shao, 宁波大学地理与空间信息技术系，中国宁波 315211

Yangyan Zuo, 宁波大学地理与空间信息技术系，中国宁波 315211

Weiwei Sun, 宁波大学地理与空间信息技术系，中国宁波 315211；宁波大学东海研究所，中国宁波 3152112

Ke Huang, 宁波大学电气工程与计算机科学学院，中国宁波 315211

Lihua Wang, 宁波大学地理与空间信息技术系，中国宁波 315211

Binjie Chen, 宁波大学地理与空间信息技术系，中国宁波 315211

Xiangchao Meng, 宁波大学电气工程与计算机科学学院，中国宁波 315211

Yong Ge，中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室，中国北京 100101

NR-IQA for UAV hyperspectral image based on distortion constructing, feature screening, and machine learning

基于失真构建、特征筛选和机器学习的UAV高光谱图像NR-IQA

【摘要】

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.

【摘要翻译】

评估无人机高光谱图像（UAV-HSI）的质量对于评估传感器性能、识别失真类型和测量数据反演精度至关重要。由于缺乏参考图像，UAV-HSI的质量评估倾向于无参考图像质量评估（NR-IQA），提供了多种应用。采用机器学习技术的遥感图像NR-IQA方法已经出现，然而，针对包含多种类型和多种失真的UAV-HSI的NR-IQA方法尚未开发。本文介绍了一种针对UAV-HSI的NR-IQA方法，采用机器学习技术。我们总结并模拟了UAV-HSI中的失真类型，构建了一个质量评估数据集，该数据集基于23幅原始高质量图像和806幅模拟降级的UAV-HSI。提取了129个特征，包括纹理、颜色、变换域、结构和统计方面，通过随机和过滤特征选择算法形成了七个特征集。利用该数据集和特征集训练了十个机器学习质量评估模型。结果表明，评估精度最高的模型是极端树（ET）（R² = 0.928，RMSE = 0.326，RPD = 3.601），使用特征集1，该特征集融合了Tamura纹理、颜色、小波变换和均值减去对比度归一化（MSCN）系数，共有11个特征，其预测和真实质量分数的PLCC和SROCC分别达到了0.963和0.925。此外，随机森林（RF）、梯度提升决策树（GBDT）、广义回归神经网络（GRNN）和极限学习机（ELM）也具有较高的评估精度（R² > 0.9，RPD > 2.5）。这些发现突显了我们提出的基于机器学习的NR-IQA方法在评估包含噪声、模糊、条纹噪声和多重失真的UAV-HSI质量中的适用性。此外，这项研究为其他高光谱图像质量评估的特征和模型选择提供了参考。

【doi】

https://doi.org/10.1016/j.jag.2024.104130

【作者信息】

Wenzhong Tian, 石河子大学机械与电气工程学院，中国新疆石河子；阿尔伯塔大学地球观测科学中心，加拿大阿尔伯塔省埃德蒙顿；石河子空间信息工程技术研究中心，中国新疆石河子；国家遥感中心石河子分中心，中国新疆石河子

Arturo Sanchez-Azofeifa, 阿尔伯塔大学地球观测科学中心，加拿大阿尔伯塔省埃德蒙顿

Za Kan, 石河子大学机械与电气工程学院，中国新疆石河子

Qingzhan Zhao, 石河子大学信息科学与技术学院，中国新疆石河子；石河子空间信息工程技术研究中心，中国新疆石河子

Guoshun Zhang, 石河子大学信息科学与技术学院，中国新疆石河子；石河子空间信息工程技术研究中心，中国新疆石河子；国家遥感中心石河子分中心，中国新疆石河子

Yuzhen Wu, 石河子大学信息科学与技术学院，中国新疆石河子；国家遥感中心石河子分中心，中国新疆石河子

Kai Jiang，石河子大学信息科学与技术学院，中国新疆石河子；国家遥感中心石河子分中心，中国新疆石河子

Corrigendum to “Identification of peat-fire-burnt areas among other wildfires using the peat fire index” [Int. J. Appl. Earth Observ. Geoinf. 132 (2024) 103973]

“利用泥炭火指数识别泥炭火烧毁区域”的更正 [《应用地球观测与地理信息国际期刊》132 (2024) 103973]

【摘要】

The authors regret <Page 1: ‘Andrey Sirin’ put in a frame>.

The authors regret <Page 1: ’v@gubkin.ru’ replace with ‘itkin.v@gubkin.ru’>.

The authors regret <Page 4: ‘Fig. 3. Values of peat and non-peat fire parameters for the 2010 fire season in the Moscow Region: Min temperature, °C (a), Area, ha (b), and Average FRP, MW (c). Values of peat and non-peat fire parameters for the 2010 fire season in the Moscow Region: Average temperature, °C (d), Duration, day (e), and Forest area, % (f). Values of peat and non-peat fire parameters for the 2010 fire season in the Moscow Region: Max FRP, MW (g), Max temperature, °C (h), and Min FRP, MW (i).’

replace with

‘Fig. 3. Values of peat and non-peat fire parameters for the 2010 fire season in the Moscow Region: Min temperature, °C (a), Area, ha (b), Average FRP, MW (c), Average temperature, °C (d), Duration, day (e), Forest area, % (f), Max FRP, MW (g), Max temperature, °C (h), and Min FRP, MW (i).>.

The authors regret <Page 7: ‘ifpi’ replace with ‘if pi’>.

The authors regret <Page 7, formula 2: ‘forsoilfire’ replace with ‘for soil fire’>.

The authors regret <Page 7, formula 2: ‘forsurfacefire’ replace with ‘for surface fire’>.

The authors regret < Page 7, formula 3: ‘≈ NA,B /NB (3)’replace with ‘ ≈ NA,B /NB, (3)’>.

The authors regret <Page 7: ‘X–’, ‘x–’ ‘

–’, ‘x–’, ‘Ip–’ replace with ‘X – ’, ‘x –’ ‘

–’, ‘x –’, ‘Ip –’>.

The authors regret <Page 8: ‘Fig. 4. Survival functions of Area, ha (a), Duration, day (b), Max FRP, MW (c) of fire (in logarithmic scale) for surface fires (green line) and peat fires (brown line). Survival functions of fire Max temperature, °C (d), Forest area, % (e), and Average temperature, °C (f) of fire for surface fires (green line) and peat fires (brown line). Survival functions of Min FRP, MW (g), Average FRP, MW (h), and Min temperature, °C (i) of fire for surface fires (green line) and peat fires (brown line). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)’

replace with

‘Fig. 4. Survival functions of Area, ha (a), Duration, day (b), Max FRP, MW (c) of fire (in logarithmic scale) for surface fires (green line) and peat fires (brown line). Survival functions of fire Max temperature, °C (d), Forest area, % (e), Average temperature, °C (f), Min FRP, MW (g), Average FRP, MW (h), and Min temperature, °C (i) of fire for surface fires (green line) and peat fires (brown line). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)’>.

The authors regret <Page 9: move article text to page 10>.

The authors regret <Page 11: put the table 1 before next paragraph >.

The authors regret <‘3.3. Comparison of the quality of fire identification using various indicatorsThe quality of identification by indicators of the first category (a higher value for soil fires, a lower value for surface fires) can be described by four conditional probabilities using formulas (2) and (3)’

replace with

‘3.3. Comparison of the quality of fire identification using various indicators

The quality of identification by indicators of the first category (a higher value for soil fires, a lower value for surface fires) can be described by four conditional probabilities using formulas (2) and (3):’>.

The authors regret <Page 12–13, Tables 3–7: Vertical center of text in cell>.

The authors would like to apologise for any inconvenience caused.

【摘要翻译】

作者对此致歉：<第1页：将‘Andrey Sirin’放入框中>。

作者对此致歉：<第1页：将’v@gubkin.ru’替换为‘itkin.v@gubkin.ru’>。

作者对此致歉：<第4页：将‘图3. 2010年莫斯科地区火灾季节的泥炭火和非泥炭火参数值：最小温度，°C (a)，面积，公顷 (b)，和平均FRP，MW (c)。2010年莫斯科地区火灾季节的泥炭火和非泥炭火参数值：平均温度，°C (d)，持续时间，天 (e)，和森林面积，% (f)。2010年莫斯科地区火灾季节的泥炭火和非泥炭火参数值：最大FRP，MW (g)，最大温度，°C (h)，和最小FRP，MW (i)。’

替换为

‘图3. 2010年莫斯科地区火灾季节的泥炭火和非泥炭火参数值：最小温度，°C (a)，面积，公顷 (b)，平均FRP，MW (c)，平均温度，°C (d)，持续时间，天 (e)，森林面积，% (f)，最大FRP，MW (g)，最大温度，°C (h)，和最小FRP，MW (i)。>。

作者对此致歉：<第7页：将‘ifpi’替换为‘if pi’>。

作者对此致歉：<第7页，公式2：将‘forsoilfire’替换为‘for soil fire’>。

作者对此致歉：<第7页，公式2：将‘forsurfacefire’替换为‘for surface fire’>。

作者对此致歉：<第7页，公式3：将‘≈ NA,B /NB (3)’替换为‘≈ NA,B /NB, (3)’>。

作者对此致歉：<第7页：将‘X–’，‘x–’，‘–’，‘x–’，‘Ip–’替换为‘X – ’，‘x –’，‘ –’，‘x –’，‘Ip –’>。

作者对此致歉：<第8页：将‘图4. 面积，公顷 (a)，持续时间，天 (b)，最大FRP，MW (c)的生存函数（在对数刻度下），表面火灾（绿色线）和泥炭火（棕色线）的生存函数。火灾的最大温度，°C (d)，森林面积，% (e)，和平均温度，°C (f)的生存函数，表面火灾（绿色线）和泥炭火（棕色线）。生存函数的最小FRP，MW (g)，平均FRP，MW (h)，和最小温度，°C (i)的生存函数，表面火灾（绿色线）和泥炭火（棕色线）。 （有关此图例中颜色参考的解释，请读者参阅本文的网络版本。）’

替换为

‘图4. 面积，公顷 (a)，持续时间，天 (b)，最大FRP，MW (c)的生存函数（在对数刻度下），表面火灾（绿色线）和泥炭火（棕色线）的生存函数。火灾的最大温度，°C (d)，森林面积，% (e)，平均温度，°C (f)，最小FRP，MW (g)，平均FRP，MW (h)，和最小温度，°C (i)的生存函数，表面火灾（绿色线）和泥炭火（棕色线）。 （有关此图例中颜色参考的解释，请读者参阅本文的网络版本。）>。

作者对此致歉：<第9页：将文章文本移至第10页>。

作者对此致歉：<第11页：将表1放在下一段之前>。

作者对此致歉：<‘3.3. 使用各种指标比较火灾识别的质量识别的质量可以通过四个条件概率来描述，使用公式（2）和（3）’

替换为

‘3.3. 使用各种指标比较火灾识别的质量

识别的质量可以通过四个条件概率来描述，使用公式（2）和（3）：’>。

作者对此致歉：<第12-13页，表3-7：单元格中文本垂直居中>。

作者为由此造成的任何不便致歉。

【doi】

https://doi.org/10.1016/j.jag.2024.104088

【作者信息】

Maria Medvedeva, 俄罗斯科学院森林科学研究所，莫斯科州乌斯彭斯科耶143030，俄罗斯

Victor Itkin, 俄罗斯科学院森林科学研究所，莫斯科州乌斯彭斯科耶143030，俄罗斯；古布金俄罗斯石油和天然气国立大学，莫斯科列宁大街65号，119991，俄罗斯

Andrey Sirin，俄罗斯科学院森林科学研究所，莫斯科州乌斯彭斯科耶143030，俄罗斯

Corrigendum to “Adaptive multi-object tracking based on sensors fusion with confidence updating” [Inter. J. Appl. Earth Obs. Geoinform. 125 (2023) 103577]

“基于传感器融合与置信度更新的自适应多目标跟踪”的更正 [《应用地球观察与地理信息国际期刊》125 (2023) 103577]

【摘要】

The authors regret <

1.

In the ‘Declaration of Competing Interest’, we forgot add the fund number after the fund name. It should be changed to the following content.

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Deer Liu reports financial support was provided by National Natural Science Foundation of China (42271434). Deer Liu reports financial support was provided by Natural Science Foundation of Jiangxi Province (20202BAB202025).

2.

In the Figure. 10. Compares the false positive count of our method with DeepFusion-MOT at different perception ranges, we mistakenly used the color in reverse. So it should be changed to the following figure.

【摘要翻译】

作者遗憾地声明：

在“利益冲突声明”中，我们忘记在基金名称后添加基金编号。应更改为以下内容： 作者声明以下财务利益/个人关系可能被视为潜在的竞争利益：Deer Liu 报告获得了中国国家自然科学基金（42271434）的财务支持。Deer Liu 报告获得了江西省自然科学基金（20202BAB202025）的财务支持。

在图10中，比较了我们的方法与DeepFusion-MOT在不同感知范围内的误报计数，我们错误地使用了反向颜色。因此应更改为以下图。

【doi】

https://doi.org/10.1016/j.jag.2024.104099

【作者信息】

Junting Liu, 江西科技大学土木与测绘工程学院，中国赣州 341400

Deer Liu, 江西科技大学土木与测绘工程学院，中国赣州 341400

Weizhen Ji, 北京师范大学地理科学学院遥感科学国家重点实验室，中国北京 100875

Chengfeng Cai, 江西科技大学土木与测绘工程学院，中国赣州 341400

Zhen Liu，江西科技大学土木与测绘工程学院，中国赣州 341400

Corrigendum to “Future challenges of terrestrial water storage over the arid regions of Central Asia” [Int. J. Appl. Earth Observ. Geoinf. 132 (2024) 104026]

“中亚干旱地区地表水储存的未来挑战”的更正 [《应用地球观察与地理信息国际期刊》132 (2024) 104026]

【摘要】

The authors regret:

This work was supported by the Western Scholars of the Chinese Academy of Sciences (2020-XBQNXZ-010), the Alliance of International Science Organizations (

ANSO-CR-KP-2020-11), the National Natural Science Foundation of P.R. China (Grant No. 42230708, 42361144887), the Joint CAS-MPG Research Project (HZXM20225001MI), the Tianshan Talent Project of Xinjiang Uygur Autonomous Region, China (2022TSYCLJ0056), and the Outstanding Postdoctoral Scholarship, State Key Laboratory of Marine Environmental Science at Xiamen University.The authors would like to apologise for any inconvenience caused.

【摘要翻译】

作者遗憾地表示：

本研究得到了中国科学院西部学者项目（2020-XBQNXZ-010）、国际科学组织联盟（ANSO-CR-KP-2020-11）、中华人民共和国国家自然科学基金（资助编号：42230708、42361144887）、中科院-马克斯·普朗克研究项目（HZXM20225001MI）、新疆维吾尔自治区天山人才项目（2022TSYCLJ0056）以及厦门大学海洋环境科学国家重点实验室的优秀博士后奖学金的支持。作者对此造成的任何不便表示歉意。

【doi】

https://doi.org/10.1016/j.jag.2024.104150

【作者信息】

Yuzhuo Peng, 厦门大学海洋与地球科学学院，海洋环境科学国家重点实验室，海洋气象与气候变化中心，厦门361102，中国

Hao Zhang, 中国科学院新疆生态与地理研究所，沙漠与绿洲生态国家重点实验室，乌鲁木齐，中国；中国科学院中亚生态与环境研究中心，乌鲁木齐，中国；中国科学院大学，北京，中国

Zhuo Zhang, 中国科学院新疆生态与地理研究所，沙漠与绿洲生态国家重点实验室，乌鲁木齐，中国；中国科学院中亚生态与环境研究中心，乌鲁木齐，中国；中国科学院大学，北京，中国

Bin Tang, 中国科学院大学，北京，中国；中国科学院大气物理研究所，大气科学与地球物理流体动力学数值模拟国家重点实验室，北京，中国

Dongdong Shen, 中国科学院大学，北京，中国；中国科学院大气物理研究所，大气科学与地球物理流体动力学数值模拟国家重点实验室，北京，中国

Gang Yin, 新疆大学地理与遥感科学学院，乌鲁木齐，中国；新疆大学绿洲生态重点实验室，乌鲁木齐，中国

Yaoming Li, 中国科学院新疆生态与地理研究所，沙漠与绿洲生态国家重点实验室，乌鲁木齐，中国；中国科学院中亚生态与环境研究中心，乌鲁木齐，中国；中国科学院大学，北京，中国

-----------关注公众号----------