最新刊期
-
SWSACNet： A change detection network for collapsed buildings extraction using multi-source images AI导读
“在遥感影像变化检测领域，研究者提出了一种新的网络模型SWSACNet，通过改进FTN模型，有效提高了倒塌建筑物的识别精度，为地震后建筑物损害评估提供了新方法。”
摘要：Objective Change detection networks based on deep learning are widely used in water monitor, urban change, etc.. However, collapsed buildings, as one of the change objectives, are rarely targeted for change detection networks. This study proposes an end-to-end collapsed building extraction model based on a change detection network including the sliding-window feature enhancement and convolution attention mix mechanism, which called SWSACNet (Sliding-Window-Shift Attention Convolution mix Network).MethodSWSACNet is an improvement of Fully Transformer Network (FTN). FTN is a network completely composed of Swin Transformer. Besides, it has a unique frame, which involves four parts: SFE (Siamese Feature Extraction), DFE (Deep Feature Enhancement), PCP (Progressive Change Prediction), DS (Deep Supervision). By encoding and decoding the feature of change objects deeply, FTN is able to learn what the collapsed buildings has changed in two temporal images and suppress irrelevant information. ACmix, a blend of convolution and attention mechanism, has been proved better performance than Swin Transformer in mainstream datasets. However, due to the different sensors, platforms, etc., the spatial heterogeneity of target features in different source remote sensing images will affect the accuracy of change detection. Concerning this problem, we designed a similarity sliding window to match the feature maps of two temporal images. Hence, we replace Swin Transformer with ACmix to extract and restore earthquake-damaged features efficiently in the phase of SFE and PCP, and using similarity sliding window to reduce misidentifications of collapsed buildings in different source image pairs before the phase of DFE. Result Taking the earthquake with 7.8 magnitude on February 6th, 2023, in Turkey as an example, establish a building seismic damage change detection dataset which consists of pre-earthquake Gaofen-2, Google images and post-earthquake Beijing-3 images, and collapsed buildings were extracted based on the SWSACNet, FTN, STANet based on the Siamese self-attention mechanism, DASNet based on a dual-attention fully-convolutional neural network, and the conventional fully-convolutional early fusion FC-EF network. The experimental results show that SWSACNet achieves the highest accuracy with F1 score of 80.8% and mIoU of 67.8%. The ablation experiments of the improved model indicates that SWSACNet obtains highest precision among three structure combinations. Beyond that, by smoothing the BJ-3 image that has higher spatial resolution to make the gradient change rate of image pairs closer, we acquire a new dataset and use it to retrain the five models. We found that the precision of five retrained models increases 1% at least, which also illustrates that appropriately narrowing the gap of gradient change rate of image pairs is an effective preprocessing for models to recognize the collapsed buildings. Finally, applying SWSACNet to three different data combinations covering Fevaipasa, Nurdagi and Islahiye area, the results show that it achieves 60.84% of average F1 score. Conclusion The application of SWSACNet indicates that the model needs richer pre- and post-earthquake training dataset and structural improvement to enhance its generalization.关键词：multi-source images;deep learning;change detection;collapsed building extraction- 21
- |
- 17
- |
- 0
更新时间：2024-07-16 - “在气象资料同化领域取得新突破，研究人员基于神经网络技术，成功构建了风云四号A星可见光反射率资料同化的快速观测算子NNFO，计算效率显著提升，为提高气象预报精度提供有力支持。”
摘要：Satellite all-sky visible reflectance contain critical information on cloud and precipitation. Therefore, data assimilation (DA) of these data has great potential to improve the forecasting skills of numerical weather prediction (NWP) models. Conventional forward operators for the DA of visible reflectance are based on numerical methods to simulate the radiative transferring processes and suffer from high computational burden. Therefore, conventional forward operators cannot meet the needs for operational DA.Objective The study is designed to construct a fast and accurate forward operator for the DA of visible reflectance data provided by the Advanced Geostationary Radiation Imager (AGRI) onboard the Fengyun-4A satellite. The forward operator should be comparable to conventional forward operators in accuracy and overwhelms the latter in computational efficiency.Method A feed-forward neural network was utilized to construct the forward operator. The input parameters of the neural network-based forward operator (NNFO) include cloud water path converted into logarithmic space, the mixing ratio of ice cloud water path and the total cloud water path, the effective radius of cloud liquid droplets, underlying surface albedo, solar zenith angle, satellite zenith angle, and relative azimuth angle between the sun and the satellite. The output of NNFO is the top-of-atmosphere reflectance. A series of sensitivity studies were performed to find the optimal (or at least, the sub-optimal) neural network settings, which include 5 hidden layers, 57 nodes in each hidden layer, the “swish” activation function for the hidden layers, and 512 batch size. In addition, the neural network is trained with an adaptive learning rate depending on the training epoch and the loss for the validation dataset, which was defined by the root mean square error (RMSE).ResultNNFO is compared with RTTOV-DOM, a typical forward operator based on discrete ordinate method to simulate the radiative transfer processes. The results indicate that NNFO is 15 or 6 times faster than RTTOV-DOM for serial or parallel modes. The mean difference, RMSE, and mean absolute error of the difference of reflectance simulated by RTTOV-DOM and NNFO (RTTOV-DOM minus NNFO) is 0.001, 0.048, 0.029, implying that the simulation accuracy between the two forward operators are comparable to each other. In addition, NNFO is validated by FY-4A/AGRI one-week reflectance observations. The results revealed that the probability density function of the simulation errors conforms to a Gaussian function, with the mean bias and standard deviation of -0.016 and 0.052, respectively.ConclusionNNFO is comparable to traditional forward operators in accuracy with a distinguished advantage in computational efficiency. Nevertheless, it is noteworthy that the current version of NNFO only supports the Ensemble Kalman Filter methods (including its variants). When it comes to the four-dimensional variational methods, NNFO should be developed by including its adjoint. In addition, NNFO could be further improved by including the aerosol effects. Improving NNFO in the aforementioned aspects and extending NNFO to DA applications are ongoing.关键词：Fengyun-4A satellite;Advanced Geostationary Radiation Imager (AGRI);visible reflectance;RTTOV;data assimilation;forward operator;neural network;adaptive learning rate- 10
- |
- 6
- |
- 0
更新时间：2024-07-16 - “遥感地表温度研究领域取得显著进展，中美两国贡献突出。中国学者发挥关键作用，推动研究国际化。载文期刊Remote Sensing和Remote Sensing of Environment表现突出。AVHRR、MODIS、Landsat等数据广泛应用，研究热点转向机器学习、局部气候带等新领域。”
摘要：Land surface temperature (LST), which determines the radiation and energy exchange at the surface, is one of the key parameters in the physical process of the ground surface. It has been widely used in numerical forecasting, agricultural situation estimation, disaster monitoring, ecological environment assessment, and many other aspects. With the development of remote sensing technology, the study of LST based on remote sensing has received extensive attention from scholars at home and abroad.In this paper, based on the core database of Web of Science, we used VOSviewer and CiteSpace software to conduct a bibliometric study on 5336 papers on remote sensed LST from 1985 to 2023. We analyzed the number of papers, research institutions, countries, authors, issuing journals and keywords, and looked forward to the future trend by combining with the hot spots of current research.Results show that (1) the field of remote sensed LST has experienced rapid growth, especially after 2012, showing exponential expansion; China and the United States have made particularly outstanding contributions to this field, with 2,169 and 1,422 papers respectively; Chinese scholars have played a pivotal role in the field of remote sensed LST, and the Chinese Academy of Sciences (CAS) has achieved remarkable results; the organizations have shown a close cooperation between institutions, and presenting an internationalized research pattern, which promotes the sharing of scientific research resources and dissemination of knowledge worldwide. (2) Over time, the research center in the field of remote sensed LST still focuses on the basic disciplines, and gradually shifts to the applied disciplines. In terms of the journals, Remote Sensing has become the main journal in the field of remote sensed LST with its all-open-source feature, and Remote Sensing of Environment leads in popularity with a high citation count of 38,884, underscoring its academic influence. (3) By clustering the keywords, it can be divided into four mainstream clusters. Land surface temperature data from sources like AVHRR, MODIS and Landsat are extensively utilized, marking a shift in research focus from traditional topics like vegetation index and soil to emerging domains such as machine learning and local climate zones. This evolution is characterized by ongoing innovation and development in both technological methods and research content. In the past decade, in addition to the traditional research directions of inversion, validation and normalization of remotely sensed surface temperature, reconstruction, downscaling and spatial-temporal fusion have become emerging research hotspots in the field.关键词：remote sensing;land surface temperature;bibliometrics;research hotspots;Development trends;fields of application;satellite data- 167
- |
- 8
- |
- 0
更新时间：2024-07-16 - “在月球表面撞击坑识别领域，本文提出了一种基于迁移学习的无锚深度卷积神经网络方法，有效提高了小尺度撞击坑的自动识别精度，为月球表面年代研究提供了新方案。”
摘要：Objective Impact crater are the most typical and common geomorphic units on the lunar surface. Their morphological WU characteristics and spatial distribution record the evolution history, climate formation and surface age of the moon. Due to the impact crater’ fuzzy edges and nested impact craters, the automatic recognition of impact craters (especially small-scale one) still faces problems such as low accuracy, and difficulty in identifying small-scale within large-scale impact crater. Therefore, to address the mentioned issues, this paper proposes an automatic anchor-free convolutional neural network method based on transfer learning strategy to automatically locate and identify small-scale lunar impact craters.Methodology Instead of using post-processing operations such as non-maximum suppression, the proposed method directly locates the centers of impact craters and regresses their sizes on the high-resolution feature map generated by the stacked hourglass network, thus realizing the automatic recognition of different types of impact craters. At the same time, the idea of transfer learning is used for training, not from scratch, so that the model has higher reliability and robustness.Results This paper selects the Orthophoto image captured by LRO WAC camera on the Lunar Reconnaissance Orbiter and the Robbins impact crater database in the Sinus Iridum and Oceanus Procellarum region to verify the reliability and robustness of the developed method. The method in this paper achieves a recall rate of up to 74.71% and an accuracy of up to 75.97%. Compared with other existing methods, it shows remarkable advantages in recognition accuracy, and has a high adaptability in extracting and identifying impact crater in different areas of the lunar surface. When comparing the performance and accuracy by drawing roc curves of different models, it can be concluded that the idea of Transfer learning can effectively help the models converge better and improve the performance of model classification. The number of craters identified by the model was calculated and compared with the number of craters in Robbins database. The result shows that the model proposed in this paper has the ability to identify small-scale craters, and the number of craters identified in a certain diameter range is greater than that in Robbins database. Therefore, the model proposed in this paper can provide a new tool for improving the lunar surface impact crater database.Conclusions The automatic recognition model in this paper can effectively achieve the extraction of impact craters and can to a certain extent solve the identification of small and medium-sized impact craters within nested impact craters. The key insight behind the developed method is to improve the model recognition accuracy based on transfer learning. Meanwhile, it does not need the non-maximum suppression operation, which would effectively realize the identification of small and medium-sized ones within nested impact craters. Although the developed method shows a superior recognition performance for the lunar impact craters, there are still the following shortcomings. The generalization of the model still needs to be further improved. Therefore, in the future work, we will also make full use of the similarity of impact crater in different regions to reduce the rate of missed detection of impact crater on the lunar surface.关键词：Impact crater extraction;Intelligent recognition;deep learning;target recognition;Sinus Iridum region;Oceanus Procellarum region- 11
- |
- 3
- |
- 0
更新时间：2024-07-15 - “最新研究揭示了非零闭合相位对InSAR形变测量的影响规律，为常规SBAS方法的基线选择和相位偏差改正提供了技术参考。”
摘要：Time-series InSAR (Interferometric Synthetic Aperture Radar), as an advanced radar remote sensing technique, achieves high-precision measurements of subtle surface deformations, widely applied in fields of geological parameter inversion and infrastructure monitoring. The recent emergence of NCP (Nonzero Closure Phase), also referred to as phase bias, challenges the conventional assumption of InSAR phase consistencies, disrupting the phase consistency assumption in InSAR data processing and leading to systematic biases in deformation results, particularly in SBAS-InSAR (Small Baseline Subset InSAR) method. This study aims to analyze the mechanisms and sources of NCP from a mathematical and physical perspective, analyze the relationship between phase bias and land cover types under different multi-look ratios, and explore its impact on conventional SBAS-InSAR methods with various time baseline combinations and the numbers of interferograms.The study first systematically explains the mechanisms and sources of NCP to clarify potential error sources in InSAR. The relationship between phase bias and different land cover types under varying multi-look ratios is analyzed based on existing research. Finally, the study investigates the impact of NCP on conventional SBAS deformation measurements. Specifically, the deformation results are analyzed under short and long time baseline combinations to determine the impact of phase bias, and the effect of introducing additional interferograms is evaluatedThe study finds that phase bias varies across different land cover types. Vegetation-covered areas are more susceptible to phase bias, while built-up areas are less affected. Apart from special multi-look ratio scenarios such as 1:1, there is no significant difference in phase bias across different multi-look ratios. Deformation results show considerable variance under different time baseline combinations, with short time baseline combinations exhibiting greater deformation bias than long time baseline combinations. Introducing long time baseline interferograms effectively mitigates the impact of phase bias, while increasing the number of short time baseline interferograms does not significantly affect the results. Overall, increasing the average time baseline helps reduce phase bias until the effect stabilizes.This study provides detailed technical insights into the selection of baselines and correction of phase deviation for conventional SBAS methods. It highlights the varying impact of phase bias on different land cover types and the effectiveness of using long time baseline interferograms to mitigate phase bias effects. These findings serve as a valuable reference for improving the accuracy of deformation measurements in InSAR applications.ObjectiveMethod Result 关键词：InSAR;NCP;phase bias;baseline combination- 12
- |
- 9
- |
- 0
更新时间：2024-07-15 - “在耕地非农化行为监测领域，本研究提出了一种新的高分辨率遥感图像分割方法。通过构建建筑物样本数据集和设计DHRformer模型，实现了对潜在非农化区域内建筑物的精准提取。实验结果表明，该方法在F1-score、mAcc及MIoU精度指标上均优于现有方法和传统方法，具有较高的实用价值。”
摘要：The arable land is the basis for ensuring the sustainable development of agriculture, and rapid and precise monitoring of the non-farming behavior of arable land is of great significance to China's food production and security.To accurately monitor the indiscriminate occupation of arable land for non-farming construction, this paper proposed a new method for accurate segmentation of the non-farming behavior of arable land in high-resolution remote sensing images. Firstly, a sub-meter-level sample dataset of buildings in potential non-agriculturalized areas was constructed based on multi-temporal remote sensing data; then, the extraction of buildings in potential non-agriculturalized areas was completed using the Deep Learning Model for Monitoring Non-Agriculturalization Behavior of Cultivated Land (DHRformer) designed in this paper. The DHRformer model consisted of a high-resolution network and a two-branch decoding structure, which enhanced the feature information of non-agricultural buildings through multi-scale fusion and expansion strategies so as to obtain richer information about the details of non-agricultural buildings.The Hecheng District in Huaihua City was selected as the study area, where the DHRformer model, along with several popular semantic segmentation methods and traditional methods, was employed for comparison. The experimental results show that this paper's method has better performance in segmentation and edge characterization of potential non-agriculturalized buildings, reaching 89.81%, 89.37%, and 80.35% in F1-score, mAcc and MIoU accuracy metrics, respectively, and the segmentation accuracy is better than that of existing methods and Conventional methods.Thus, the DHRformer model proposed in this paper has high practical value in the task of monitoring the non-farming behavior of arable land.关键词：non-agriculturalization of cultivated land;building segmentation;high resolution network;two-branch decoding structure;encoder-decoder;high resolution image- 7
- |
- 7
- |
- 0
更新时间：2024-07-15 - “在遥感信息提取领域，极化SAR地物分类技术取得显著进展。本文综述了基于极化分解、经典机器学习、深度学习等方法的分类技术，以及在海面溢油探测、舰船检测等任务上的应用，为相关研究提供了新方向。”
摘要：Remote sensing technology enables us to monitor the earth from space, sense the rhythm of rivers, lakes, and seas and the pulse of social and economic development in real-time, and effective early warning, prevention, and evaluation of natural disasters, in which SAR technology plays an increasingly important role. Remote sensing image classification is an important step of remote sensing image analysis, and it has always been one of the hot spots in related research fields. Due to the complexity of ground target characteristics and the diversity of remote sensing imaging techniques, the accurate interpretation of remote sensing images requires a deep understanding of the characteristics of the image and making full use of the prior knowledge of ground objects. In recent years, with the development of synthetic aperture radar (SAR), especially polarimetric SAR technology, the research of remote sensing object classification has developed rapidly. In this paper, the research progress of polarimetric SAR remote sensing image classification is reviewed. This paper firstly introduces the basic theory of SAR remote sensing and the main data sources of spaceborne SAR, then introduces the decomposition of polarimetric SAR data, the classical machine learning algorithms for polarimetric SAR, the deep learning-based algorithms and the methods of fusing optical and SAR images as well as the classification algorithms based on compact polarimetric SAR. Then, the paper introduces the research progress of polarimetric SAR image classification for marine oil spill detection, ship detection, coastline extraction, land use classification, and sea ice/ ice cap classification. Finally, the development trend of polarimetric SAR image classification is prospected. From the perspective of the authors, the development of polarimetric SAR classification has the following trends: 1) From single polarimetric to multi- and compact polarimetric SAR modes; 2) From medium/low resolution, small range to high resolution, large range remote sensing applications; 3) From single temporal to multiple temporal sequence images analysis applications; 4) From manually design of feature extraction methods to automatic feature extraction using deep learning models; 5) From single-source SAR image classification to SAR, optical, LiDAR and other multi-source image fusion classification. To make full use of the information provided by polarimetric SAR data sources, it is necessary to master the key technologies of radar signal processing, image analysis, pattern recognition, multi-source information fusion, big data analysis, and other aspects. With the rapid development of technology, talents with interdisciplinary backgrounds such as electronic engineering, remote sensing, and artificial intelligence are urgently needed in this field. The authors hope that through the introduction of this article, readers can improve their understanding of the field of SAR remote sensing classification to a certain extent, to better grasp the development trends of this technology.关键词：polarimetric SAR;remote sensing;classification;multi-source information fusion- 17
- |
- 4
- |
- 0
更新时间：2024-07-15 - “在矿区地表裂缝监测领域，研究者设计了双任务卷积神经网络Goaf-DTNet，通过信息互补提高裂缝检测精度，为矿区监测提供有效数据。”
摘要：ObjectiveAutomatic detection of ground cracks in the goaf of coal mines plays an important role in production security and ecological environment management. Due to the complex background in the coal mine, the variable geometry and scale of cracks. Automatically detecting ground cracks in the goaf remains a challenging task. Cracks extraction could be treated as a segmentation task from a global view or a skeleton extraction task (boundary detection task) from a local view. There has been a lot of work view cracks extraction as a single task independently. However, those methods cannot produce enough data for subsequent measurement and quantitative evaluation about the extent of detriment of ground cracks. And they neglect the information interaction of different tasks which could potentially improve the accuracy and efficiency. In order to solve these problems, a dual-task convolutional neural network (CNN), named Goaf-DTNet (Dual-Task Convolutional Neural Network for Goaf Crack Recognition) is designed to automatically detect ground cracks using unmanned aerial vehicle (UAV) imagery with high spatial resolution.MethodIn Goaf-DTNet, an ASPP (atrous spatial pyramid pooling) module was introduced to extract multi-scale semantic information. Considering the characteristics of ground cracks in the goaf, MFFM (Multi-scale feature fusion module, MFFM) in the crack segmentation branch was designed to further integrate local and global contextual information. In order to improve the accuracy, in the crack skeleton extraction branch, SGFM (Segmentation-guided crack skeleton feature extraction module) was used to provide more spatial information through spatial attention mechanism. The proposed dual-task model can explicitly avoid calculating parameters in the shared layers, thus reduce the memory footprint and speed up each task. Meanwhile, through the communication between two tasks, the complementary information could improve the detection accuracy of each task.ResultIn the task of skeletion extraction, the F1-score and IoU (intersection over uinon) is 0.71 and 0.55 respectively, whiech is about 1% higer than the second best method, and the IoU is 16.35% higher than PSPNet. In the task of surface segmentaion, the proposed model perform best over all compared methods, the ODS(optimal dataset scale) and AP(averge precision) value is 0.56 and 0.54 respectively. In addition, the test results in a open-source dataset present that our method is better than the others, the F1, IoU value is 0.89 and 0.80 respectively. In terms of skeletion extraction, the OIS(optimal image scale), ODS, AP is 0.58, 0.58 and 0.55, respectively. In order to prove the effectiveness of the proposed MMFM and SGFM, ablation experiments were conducted. Results of ablation experiments show that, after adding MMFM in the crack segmentation branch, F1 and IoU increase 0.0827 and 0.0918 respectively, which reveals that it is helpful for crack detection through fusing the local and global information. In terms of crack skeleton extraction task, OIS, ODS and AP increased by 0.1246, 0.1630 and 0.2140 respectively after the SGFM was embedded into the branch.ConclusionThe experimental results show that the Goaf-DTNet is effective in ground cracks detection from the goaf of coal mines. MFFM proposed in this paper is conducive to the model to obtain more complete and continuous crack identification results by integrating multi-scale context semantic information. SGFM uses information from the surface segmetation branch to provide more spatial information for linear crack features, which effectively improve the detection accuracy. Furthermore, Taking advantage of the synergy between the two tasks , the accuracy of each task is imporved.关键词：Coal mines;goaf;unmanned aerial vehicle (UAV);crack detection;convolutional neural networks (CNNs);Multi-task learning;deep learning- 15
- |
- 6
- |
- 0
更新时间：2024-07-15 - “在农业资源调查领域，专家建立了多尺度时空全局注意力模型MSSTGAM，有效提高遥感影像时间序列农作物分类精度，为农业监管和规划提供解决方案。”
摘要：Automatic intelligent interpretation of the fine types of crops by utilizing remote sensing image time series plays an important role in the fields of agricultural resource investigation, supervision and planning. The existing deep learning methods extract local spatial or local temporal information by convolutional or recurrent neural networks which have inadequate utilization of spatial-temporal information, resulting in low classification accuracy.The self-attention mechanism is able to fully exploit data information by obtaining global attention. Thus, we propose a multi-scale spatial-temporal global attention model (MSSTGAM), which combines spatial self-attention mechanism and temporal self-attention mechanism to construct multi-scale spatial-temporal global attention, and fully mine the information of remote sensing image time series for the fine classification of crop types.The proposed method is evaluated and tested on the publicly available dataset PASTIS and custome Mississippi dataset. The results demonstrate that MSSTGAM is capable of identifying crop types from remote sensing image time series, and achieves the best quantitative result compared with other methods. Moreover, the visualization results of MSSTGAM have better inner-parcel spatial consistency.This paper’s findings show that multi-scale spatial-temporal global attention has significant theoretical and practical significance and is more effective for the fine classification of crop type from remote sensing image time series.关键词：remote sensing image time series;crop type classification;self-attention mechanism;global attention;multi-scale spatial-temporal feature- 191
- |
- 82
- |
- 0
更新时间：2024-06-26 - “最新研究揭示了华北平原近二十年地下水储量变化的时空特征，分析了降水、农业用水和居民用水对地下水储量的影响，为水资源管理提供了重要参考。”
摘要：”Objective” Groundwater constitutes a pivotal source of water supply for agricultural and residential purposes in the North China Plain. However, the excessive exploitation of groundwater in the North China Plain has given rise to numerous ecological and environmental challenges in recent years. Therefore, exploring the shifts in water resources in this region over the past two decades holds significant importance for effective water management. Notably, the satellite alternation during 2017-2018 resulted in a notable gap in data, often posing challenges for researchers to comprehensively cover this transition period in their studies, thus constraining the scope and depth of their analysis.”Method” To address this issue, our study employed the Singular Spectrum Analysis (SSA) method to interpolate and fill in the missing GRACE Mascon gravity satellite data. Furthermore, we leveraged the GRACE CSR Mascon RL06 gravity satellite data, GLDAS model data, and irrigation water use data to investigate the spatial and temporal patterns of water storage changes in the North China Plain from 2002 to 2022. Additionally, by integrating precipitation station data, wheat distribution data, and luminous radiation intensity distribution data, we delve into the impacts of precipitation, agricultural water use, and domestic water use on the spatiotemporal changes in water storage in the region. This comprehensive approach, considering both the main income items and expenditure items of groundwater resources, offers a robust framework for analyzing the intricate water dynamics in the North China Plain.”Result” The study reveals several key findings:(1) Groundwater storage anomaly in the North China Plain exhibited a decreasing trend, with rates of -0.19 cm/a and -1.69 cm/a during 2002-2011 and 2012-2019, respectively, followed by an increase of 4.78 cm/a in 2020-2022. (2) Spatially, there is an upward trend in the northeastern part and a downward trend in the southwestern part of the North China Plain between 2002 and 2022, with a more prominent decrease in the north compared to the south. (3) The average monthly water consumption for farmland irrigation in spring and summer stands at 2.42 cm/month, peaking at 5.00 cm/month. (4) The uneven spatial distribution of annual precipitation from 2002 to 2022 led to a more pronounced decline in groundwater reserves specifically in the northern section of the North China Plain. Spatial variations in agricultural and residential water usage significantly influenced the trends of groundwater reserve changes across the region. Notably, these spatial differences were mirrored in the varying degrees of agricultural and residential water consumption. The escalating agricultural and residential water demands exacerbated the already decreasing trend of groundwater storage in the northern part of the North China Plain. (5) An increase in monthly precipitation positively impacts groundwater, and the direction of this relationship aligns with the variation in groundwater storage anomaly. These findings provide crucial insights into the complex dynamics of water resources in the North China Plain and inform effective water management strategies.”Conclusion” This paper delves into the spatial and temporal patterns of groundwater storage anomaly in the North China Plain over recent years, unveiling the intricate influence of multifaceted factors, encompassing precipitation, agricultural water consumption, and domestic water utilization, on the dynamic anomaly in groundwater storage in this region.关键词：North China Plain;GRACE Mascon;SSA;groundwater storage anomaly;irrigation water use;precipitation;agricultural water use;residential water use;impact analysis- 145
- |
- 65
- |
- 0
更新时间：2024-06-18 - “最新研究揭示了格陵兰北部冰前河网的空间分布与形态特征，为理解冰盖融水汇流过程提供了重要数据支撑。”
摘要：ObjectiveEach summer, proglacial river networks develop on the northern Greenland and can route large volumes of surface meltwater into the ocean, acting as important meltwater connections between the ice sheet and the ocean. However, the spatial distribution and geomorphology of the proglacial river networks on the northern Greenland remain unclear. Based on 10 m resolution Sentinel-2 satellite images and 30 m resolution Copernicus DEM, this study maps proglacial water on the northern Greenland (with an area of 100,132 km2) in 2020 using an automatic water remote sensing information extraction algorithm constrained by routing process.MethodFirst, river features are enhanced from the image background by using modified normalized difference water index (MNDWI), Gabor filter, and path opening operator. Second, the area of interest (AOI) for rivers is constructed to reduce the interference of bare ground and shadow features by combining height above the nearest drainage (HAND) AOI and topographic depressions. Third, the derived water mask is interested with DEM-modeled drainage networks to delete pseudo drainage channels, to generate continuous, realistic drainage networks and to classify proglacial river networks and isolated lakes based on their morphometric characteristics. Finally, proglacial river networks are connected by using continuous DEM-modeled drainage networks to produce the dataset of 10 m resolution continuous proglacial river networks and isolated lakes (CPRs&ILs).ResultOur mapping results show the spatial distribution of proglacial river networks, compare with four water remote sensing datasets (Dynamic World V1, CALC-2020, Esri Land Cover and ESA WorldCover) and quantitatively analyze the length, width, area, drainage density and order of river networks. Our results indicate that: (1) this study accurately extracts and divides remote sensing information of the proglacial river networks and isolated lakes, and the overall accuracy of river network remote sensing information extraction is 93%±2%, which is better than the four comparison datasets (overall accuracy of 83%-89%). Our results can accurately reflect the spatial distribution of the proglacial river networks in the study area, especially small rivers during the melting period; (2) in 2020, a total of 995 proglacial river networks, covering a total water area of 1,832.6 km², develop on the northern Greenland, and are able to route 90.5% of total surface meltwater runoff into the ocean; (3) proglacial river networks have significant network order difference ranging from 1 to 5. Order 1-2 river networks account for over 84% river networks, whereas the limited number of 40 (<5%) order 4-5 high-order river networks dominate the length of river networks (52.9%), water area (63.9%), catchment area (54.1%), and the routing of surface meltwater runoff (69.3%).ConclusionViewing collectively, this study produces a high-resolution proglacial water dataset with large spatial coverage, making up for the lack of precision of the existing datasets, and shows the overall distribution of the large-scale proglacial river networks. Our findings reveal that the widely-distributed proglacial river networks on the northern Greenland are dominated by high-order river networks, and substantially route surface meltwater, thereby improve our understanding of meltwater routing process from the supraglacial to proglacial regions on the northern Greenland.关键词：proglacial rivers;proglacial lakes;river networks;river network order;Greenland ice sheet;Remote sensing dataset;DEM drainage network;river and lake classification;accuracy verification- 54
- |
- 37
- |
- 0
更新时间：2024-06-18 - “最新研究进展显示，遥感技术已成为监测河流水文情势的重要手段，尤其在青藏高原等水文监测资料缺乏区域。本文综述了基于光学影像、雷达影像、卫星测高等数据的河流水文要素遥感监测技术，为理解河流水文变化过程提供了新视角。”
摘要：Rivers are integral to the water cycle, underpinning human development, ecological health, and regional climate stability. Recently, global warming, glacial melt, and recurring hydrological disasters have intensified disturbances in river systems, necessitating broad-scale monitoring of complex hydrological changes. While traditional field measurements are valuable, limitations in their spatial and temporal coverage call for alternative approaches. With the advancement of sensor technology and the proliferation of satellite platforms, (satellite) remote sensing has emerged as a pivotal method for contemporary river hydrology monitoring. Compared to hydrological field measurements, it offers significant advantages in terms of real-time data acquisition, vast spatial coverage, and reduced economic costs. Various remote sensing monitoring techniques have been extensively applied to monitor river characteristics such as area/width, water level fluctuations, runoff estimation, and forming diverse-scale remote sensing products of hydrological elements. This paper reviews various monitoring techniques for river hydrological variables using optical or radar imaging and satellite altimetry. It analyzes the latest research progress in the hydrologic variables, encompassing river width, water area, water level, runoff, and their changes. Additionally, the spatial scale and feasibility of previous literature are thoroughly discussed. The Tibetan Plateau, known as the "Roof of the World", is one of the regions with a serious shortage of in-situ hydrological monitoring data, despite being the source of major rivers in Asia. The application of remote sensing technology for river hydrological monitoring on the Tibetan Plateau encounters challenges in data sharing, pronounced spatial and temporal heterogeneity of hydrological processes, and intricate response characteristics to a warming and humidification climate.This study begins by examining the main satellite remote sensing data sources and methods used to monitor various hydrological elements of rivers. It summarizes the current research progress in river hydrology monitoring using remote sensing technologies and explores future development opportunities. The review also addresses the advancements and challenges of hydrological remote sensing techniques specifically applied to river monitoring on the Tibetan Plateau. Several persistent issues in river hydrological remote sensing development have been identified: (1) The accuracy of extracting river area and width in regions with complex topography is severely affected by mixed pixels and spectral similarities. (2) In areas with sparse or no hydrological stations, assessing remote sensing data's quality and potential applications remains challenging. (3) There is a notable lack of comprehensive monitoring and studies on the spatial and temporal patterns of hydrological changes in the inland flow areas of the Tibetan Plateau. Future research directions for remote sensing of river hydrology are outlined as follows: (1) Integrate multi-source remote sensing data and enhance the technologies and their applications for hydrological monitoring. (2) Optimize and innovate more universally applicable remote sensing algorithms for river hydrology. These priorities aim to address the critical challenges in hydrological remote sensing and enhance the capability and accuracy of monitoring systems, particularly in complex and underserved regions like the Tibetan Plateau. This paper aims to promote the deepening of river hydrology research on the Tibetan Plateau region, providing more accurate and scientific-technical support for practical water resources management and policy-making.关键词：River;Hydrology;remote sensing;Tibetan Plateau;water extent;water level;runoff- 70
- |
- 49
- |
- 0
更新时间：2024-06-18 -
A LiDAR Biomass Index – based method for Aboveground Biomass Calculation of Larch Sub-compartments AI导读
“在森林生物量估算领域，本研究利用激光雷达生物量指数（LiDAR Biomass Index，LBI）计算了落叶松单木生物量并验证了其精度，为大范围森林生物量估算提供了新方法。”
摘要：Objective The LiDAR biomass index can calculate the above ground biomass of individual trees based on airborne LiDAR data, and it has been verified to have high accuracy for biomass calculation at tree level and plot level. However, its ability to complete large-scale forest biomass mapping has not been fully explored. The aim of this research is to verify the accuracy of LBI for aboveground biomass estimation at sub-compartment scale taking the widely planted Larix olgensis tree species in north China as an example and laying a theoretical foundation for the widespread application of this index.Method First, the existing tree species classification results based on hyperspectral data was used to select the point clouds of Larix olgensis species in Mengjiagang forest farm. Second, the NSC algorithm was used to complete the individual tree segmentation of the selected point clouds. And then, the LBI was used to calculate the forest biomass of each individual tree. Combining with the AGB_LBI biomass model of Larix olgensis species that has been constructed based on 35 individual sample trees, the biomass of each individual tree was calculated, and the biomass of the each sub-compartment was obtained through accumulating the biomass of individual trees within the sub-compartment. In this research, the calculation accuracy was verified through the silviculture survey data obtained from the local forestry department, including over 70000 individual trees. Meanwhile, the universality of LBI in estimating the biomass for the same tree species of different regions at the sub-compartment level was evaluated based on the existing AGB_LBI models of other forest farms, and the results were compared with the commonly used LiDAR metrics-based regression (LMR) methods.Result The results indicated that LBI can achieve forest biomass estimation at the sub-compartment level with high accuracy. When using individual trees samples selected from different regions to calibrate the AGB_ LBI model, the obtained biomass values were comparable with the measured data, with R2 ranging from 0.86 to 0.87 and rRMSD (relative Root Mean Square Difference) ranging from 34.20% to 40.23%. The biomass results calculated from each model did not have significant differences. However, the increase in the number of sample trees used for model calibration still has a certain impact on the robustness and accuracy of biomass calculation. Overall, the accuracy of LBI-based method is comparable to the LMR method although the sample trees used to calibrate the AGB_LBI model is only accounts for 1% that used to calibrate the LMR model. Meanwhile, the LBI method has stronger universality among the same tree species in different forest farms. Finally, the AGB_ LBI model was used to calculate the biomass of each individual tree in the western region of Mengjiagang forest farm and complete the biomass mapping. The obtained biomass distribution has a similar trend to the existing biomass map and is consistent with the forest sub-compartment map, which achieved high consistency at the scale of 20 m×20 m (R2=0.75, RMSD=1.55 t).Conclusion The high-precision estimation of biomass by LBI at the sub-compartment scale demonstrates its potential for conducting large-scale estimation of forest AGB. However, due to the difficulty in obtaining validation data, this research only verified its accuracy on the species of Larix olgensis, and did not conduct experiments on other tree species. But previous studies have shown that this method can theoretically be applied to more tree species and forest situations, which is worth further exploration. Overall, this research provides a theoretical basis for more precise, large-scale, and high-precision forest biomass estimation.关键词：LBI;airborne LiDAR;individual tree;sub-compartment level;biomass- 37
- |
- 30
- |
- 0
更新时间：2024-06-18 - “长江中游区域水文干旱监测取得新进展，基于卫星遥感技术构建的动态水体指数DWI与标准化降水指数SPI显著正相关，为区域水文干旱监测与评估提供重要参考。”
摘要：Under the influence of global climate change and human activities, the frequency and intensity of drought have increased, posing challenges to food, ecology, and water security. Using satellite remote sensing technology for rapid acquisition, comprehensive coverage, and high-precision monitoring of surface water is crucial for understanding the mechanisms and evolution of hydrological drought. With the continuous expansion of global dynamic surface water coverage products, there's a valuable opportunity to explore the response relationship between surface water dynamics and drought. This study focused on the middle reaches of the Yangtze River and its 4683 sub-basin units, utilizing the Global Surface Water Dynamic (GLAD-GSWD) dataset to quantify abnormal changes in seasonal surface water areas from 1999 to 2020 through the Dynamic Surface Water Index (DWI). It aimed to investigate the response relationship between hydrological drought and meteorological drought from the perspective of dynamic surface water area, conducting an analysis of its uncertainties. The results showed that: (1) DWI effectively depicted changes in the wet/dry status of the middle reaches of the Yangtze River and its sub-basin units. From 1999 to 2020, the fluctuations of DWI time series and SPI time series in the middle reaches of the Yangtze River are basically the same, which can capture the impacts of most of the extreme hydrological and climatic events on the dynamic ranges of surface water in the region. The fluctuations of the DWI time series show significant positive correlations with the SPI time series at the four time scales (1, 3, 6, and 12 months), with the highest correlations with the SPI6 (r=0.484, p<0.01). Approximately 88.2% of the 4683 sub-basin units showed a significant positive correlation between DWI and SPI, with response times predominantly between 6 and 12 months; (2) The study also uncovered uncertainties in the correlations between DWI and SPI, primarily stemming from two key factors. The quality of raw data posed a challenge, marked by insufficient seasonal surface water area coverage and limited observed data, potentially leading to the degradation of correlations. The HANTS algorithm combines smoothing and filtering techniques to effectively identify, remove, and fill in outliers in time series data. The data reconstructed based on HANTS effectively improved the monitoring effect of DWI, but it couldn't completely eliminate the impact of data quality. Secondly, the complex interplay of meteorological droughts at both local and upstream basins significantly influenced surface water in the sub-basins. Consequently, as the statistical analysis expanded to larger basin scales, the direct impact of meteorological factors on basin water bodies strengthened, resulting in a more pronounced correlation. The methodology employed in constructing hydrological drought monitoring indicators, with sub-basins as monitoring units and dynamic water bodies as monitoring data, not only offers invaluable insights for regional and global hydrological drought monitoring and assessment but also underscores the pivotal role of remote sensing dynamic water products in advancing these endeavors. Additionally, efforts to enhance the quality and resolution of satellite remote sensing data promise in refining the accuracy and reliability of hydrological drought monitoring, thereby bolstering resilience against water-related crises on both regional and global scales.关键词：Dynamic Surface Water;Hydrological Drought;remote sensing;drought index- 37
- |
- 24
- |
- 0
更新时间：2024-06-18 - “最新研究揭示，中法海洋卫星CFOSAT搭载的海浪波谱仪SWIM在识别降雨事件时存在低估。通过改进MP算法，提高了降雨标识的准确度，为雷达观测结果的精度提升提供了有效解决方案。”
摘要：Objective Rain flag is necessary for Ku-band altimeters, because the presence of rain in the sub-satellite track will cause attenuation of the backscatter signal, which can lead to errors of the altimeter products. The SWIM (Surface Wave Investigation and Monitoring) instrument payload on the CFOSAT (China–France Oceanography Satellite) is a Ku-band (13.575 GHz) real aperture radar which illuminates the surface sequentially with six incidence angles. The nadir beam of the SWIM can be used as an altimeter except for measuring the SSH (Sea Surface Height). The rain events identified by rain flag in SWIM L2 nadir products offered by CNES (Centre National d’Etudes Spatiales) has been underestimated compared to these in Jason-3. Thus, the rain flag in SWIM L2 nadir products need to be improved.Methods Apparently, the dual-frequency rain flag algorithm used in Jason-3 products cannot be applied in SWIM which only works on Ku-band. To address this issue, a rain flag based on MP (Matching Pursuit) algorithm is introduced and modified to make it applicable to SWIM in this article, which is extremely versatile and can be easily adapted to any altimeter data. The along track waveform of mispointing angles can easily be decomposed by MP algorithm based on wavelet packet decomposition. Then the intervals where the mispointing angles presents short-scale coherent variations can be detected. Except for rain events, the- 11
- |
- 13
- |
- 0
更新时间：2024-06-18 - “在森林辐射传输建模领域，专家基于随机辐射传输理论开发了三维解析模型ESRT，通过实测数据验证，为复杂森林场景监测提供理论基础。”
摘要：Objective The complicated spatial heterogeneity in forest scenes may have an effect on canopy reflectance. How to address heterogeneity is a challenge in the field of radiative transfer modeling. To overcome the accuracy limitations of classical analytical models caused by simplified scenarios and the efficiency limitations of computer simulation models in large-scale applications, we developed a three-dimensional analytical radiative transfer model called ESRT based on stochastic radiative transfer theory. At present, the application of ESRT in different complex forest scenes still needs to be explored, and more field data is essential for model verification.Method This paper introduces the basic principles and input-output parameters of the ESRT model, in which two key input parameters were proposed to express different heterogeneous canopy structures: (1) the intercanopy heterogeneity index (yr), representing the ratio of tree crowns with different optical properties to the total number of trees; (2) the intracrown heterogeneity index (yi), denoting the ratio of elements with different optical properties within a single tree crown to the total number of elements. To evaluate the model performance in simulating different kinds of heterogeneous forest canopy spectra, 21 30m×30m sample plots in mixed forests and 50 10m×10m quadrats in pest-damaged forests were set up with individual tree measurements and remote sensing data acquirements. Control experiments based on the original SRT model and the three-dimensional model LESS were conducted for the two cases to compare simulation results with the extended ESRT model. Based on the framework of ESRT, sensitivity analyses were conducted to reveal the effect of mixing and pest levels on forest canopy spectra.Result The results showed that compared to the original SRT model simulations, the canopy spectra simulated by the extended ESRT model have better consistency with the measured spectra from the sample plots for the cases of mixed forests (R2 = 0.77，RMSE = 0.075) and pest-damaged forests (R2 = 0.64, RMSE = 0.039), and the simulation accuracy is closer to that of a 3D computer model. The conifer-broadleaf ratio and the vertical distribution of damaged foliage can both affect the canopy spectral signals. In mixed forests, the canopy NDVI decreases with the decrease of forest coverage and with the increase of coniferous tree species proportion. Canopy coverage is the main factor affecting NDVI when the coverage is low, but the impact of mixing on NDVI becomes more apparent when the coverage is high. In pest-damaged forests, the sensitivity of BRF to yi varies significantly with damage types. BRF of the bottom damaged forest exhibits slight change at lower yi and then shows a sharp change toward the maximum of yi. On the contrary, BRF of the top damaged forest changes dramatically at lower yi but levels off at higher yi.Conclusion The three-dimensional analytical model ESRT balances the simulation accuracy of three-dimensional structures with the simulation efficiency of classical analytical models, resolving the difficulty in accurately and efficiently simulating radiation transfer in the presence of heterogeneity. The computation time is suitable for large-scale heterogeneous forest canopies. The extended ESRT can simulate forests with mixed canopy types and heterogeneous leaf distribution structures within the canopy, potentially aiding forest managers in more accurately and effectively monitoring forest dynamic changes.关键词：radiative transfer;three-dimensional analytical model;heterogeneous forest canopy;Mixed Forest;forest pest- 17
- |
- 18
- |
- 0
更新时间：2024-06-18 - “在三维点云数据处理领域，基于深度学习的点云刚性配准技术取得显著进展。本文系统综述了相关技术研究，为后续探索提供参考，为自动驾驶、机器人等领域的应用提供解决方案。”
摘要：Point cloud registration is the process of spatial alignment of two or more point clouds through geometric transformations. As a fundamental task in 3D point cloud data processing, it is an important preprocessing operation for tasks such as 3D modeling, object recognition and scene understanding. Due to the non-structural, sparse, and uneven characteristics of point cloud data, point cloud registration remains one of the hotspots and challenges in computer vision, mapping and remote sensing, although there have been a lot of studies. With the emergence and rapid development of neural networks, deep learning has shown great potential in applications such as point cloud classification, recognition, detection, and reconstruction. In recent years, many researchers have also attempted to apply deep learning techniques to point cloud registration. Deep learning-based point cloud registration methods can automatically learn highly discriminative and robust point cloud features that contain geometric structural information and semantic information, which are crucial for achieving high registration accuracy. In this paper, the research on pairwise point cloud rigid registration technology based on deep learning is systematically reviewed and analyzed. Firstly, the feature extraction network based on deep learning is introduced. Then, the research progress of the registration methods of correspondence estimation, pose regression and scene flow estimation is reviewed, and the characteristics, advantages and disadvantages of these three methods are summarized. Then, this paper systematically summarizes and categorizes existing publicly available datasets that can be used for rigid point cloud registration. Finally, this paper summarizes the current research status of point cloud registration, explains the advantages and limitations of existing methods in feature learning, registration accuracy, registration efficiency and other aspects, and prospects for future research directions, proposing three exploration directions: (1) Existing deep learning-based point cloud registration methods require a large amount of annotated data, while data annotation time and manpower cost are high. Methods such as few-shot learning, self-supervised learning, weakly supervised learning and unsupervised learning can alleviate the need in deep learning technology for large amounts of labeled data to some extent. (2) The matching primitives in existing deep learning-based point cloud registration methods still primarily focus on 3D points. Compared to geometric elements such as 3D line segments and planes, the ambiguity of 3D points is large and the mismatching rate is high. In the future, it is worth considering using deep learning techniques to automatically extract the geometric structural elements of scenes from 3D point cloud data, and then further solve the registration problem of 3D point clouds in large scenes based on these geometric structural elements. (3) Existing deep learning-based registration methods mainly utilize spatial geometric features of scenes, with less consideration of semantic information. In recent years, there have been rapid advances in deep learning-based point cloud semantic understanding techniques. Therefore, in the future, it is worth considering combining spatial geometric structural information with semantic information to solve the registration problem of complex scenes in 3D point clouds.关键词：3D point cloud;rigid registration;deep learning;point cloud feature;correspondences estimation;pose regression;scene flow estimation;registration datasets- 45
- |
- 23
- |
- 0
更新时间：2024-06-18 - “最新研究揭示了作物病害遥感监测的光谱响应机制及其特异性，为精准防控和粮食安全提供了重要依据。”
摘要：Remote sensing monitoring of crop diseases plays a crucial role in food security in terms of the precision management of chemical fungicides and the efficient assessment of crop losses. Spectroscopic detection of disease infection has been investigated for numerous crop diseases individually. However, it remains unclear how biochemical and spectral variations differ in response to divergent diseases given the distinct symptoms caused by different pathogens.Objective This study aimed to determine the pathological mechanism and specificity of the spectral responses of two types of fungal diseases by comparing their specific spectral signatures and disease monitoring performance.Methods The biotrophic wheat powdery mildew (WPM) and the semi-biotrophic rice leaf blast (RLB) diseases were used as examples for the comparison. With the reflectance measurements of infected leaves and radiative transfer modeling, a comparative analysis for these two diseases was conducted in terms of spectral responses, leaf biochemical and structural parameters. Additionally, we assessed the specificity of various disease-related spectral features, which were proposed in previous studies for the monitoring of WPM or RLB, by accuracy comparison in the detection of diseased leaves and the estimation of leaf lesion proportion (LLP).Results The results showed significant differences in the intensity of spectral responses to the two diseases despite the similarity observed in the general trend in spectral variations. In addition, distinct variations appeared in the spectral shape at the green peak and near-infrared plateau between WPM and RLB. Moreover, the pigment variations in response to two infections were generally similar, whereas the response was more pronounced for RLB. Notably, the leaf water content and structural parameter displayed significant changes only in relation to the severity of RLB. In disease detection, the spectral features developed for WPM or RLB generated higher accuracy in detection of the target disease than the other disease. Wavelet features of WF3,820 and WF5,866 displayed the highest accuracy and specificity for WPM and RLB, respectively. Regarding the severity quantification, most spectral features exhibited higher sensitivity to the LLP of RLB than to that of WPM. Specifically, a variat of rice blast index (RIBIred) and the photochemical reflectance index (PRI) demonstrated the highest accuracy and specificity in the LLP estimation of WPM and RLB, respectively. Among the WPM- or RLB-related spectral features, RIBIred showed the optimal monitoring performance and specificity in both disease detection and severity estimation (Overall accuracy = 0.74, R2 = 0.58).Conclusions Our findings provide solid evidence and new insights into disease-specific spectroscopic monitoring by associating spectral responses with pathogenesis of two types of fungal diseases. This study offers significant contributions to the understanding of disease monitoring mechanisms and the identification of multiple diseases with hyperspectral remote sensing.关键词：Disease detection;severity quantification;PROSPECT;biochemical parameters;leaf structure;spectral index;continuous wavelet transform- 30
- |
- 35
- |
- 0
更新时间：2024-06-18 - “在水质监测领域，广东省某城中村河流的非光学特性水质参数反演研究取得新进展。专家建立了基于LBFGS加速多层感知网络模型（LBFGS-MLP），验证了其在总磷、总氮和氨氮浓度反演中的有效性和可行性，为更全面地评估城市河流水体状况提供理论依据和参考。”
摘要：Water is the source of life, the foundation of survival, the necessity of production, and the basis of ecology. However, under the dual pressures of human activities and climate change, aquatic ecosystems are facing increasingly severe challenges, especially the serious problem of water pollution, which directly threatens the physical and mental health of residents. Water quality monitoring plays a crucial role in water pollution control, which precisely evaluates the health of water bodies and promptly adjusts control strategies, ensuring the stability and health of water environmental quality. Hyperspectral remote sensing holds significant potential in water quality monitoring, and with the rapid development of unmanned aerial vehicles (UAV) and hyperspectral technology, UAVs equipped with hyperspectral sensors have greatly improved in spectral resolution and spatial resolution. This makes water quality parameter inversion using hyperspectral remote sensing gradually become a research hotspot. However, current research predominantly focuses on optical water quality parameters, with relatively less emphasis on non-optical parameters that also reflect human activities' impact on water bodies. In this study, an urban river in a certain village in Guangdong Province is chosen as the study area, and an experiment is conducted involving UAV for hyperspectral remote sensing image acquisition and simultaneous water sample collection. Then, we propose a multilayer perceptron (MLP) network model accelerated by limited-memory Broyden-Fletcher-Goldfarb-Shanno (LBFGS), named LBFGS-MLP for the inversion of non-optical water quality parameters. The parameters include total phosphorus (TP), total nitrogen (TN), and ammonia nitrogen (NH3-N), which are important indicators for measuring the nutritional status of water bodies. Specifically, through Pearson correlation analysis, spectral bands related to the three non-optical water quality parameters of TP, TN, and NH3-N, are selected as model inputs. Subsequently, based on exploring the impact of different network depths and optimization algorithms on model performance, the LBFGS optimization algorithm is adopted to accelerate the multi-layer perceptron network, and the loss function is mean squared error (MSE). Finally, the LBFGS-MLP model is applied to spatially analyze the concentrations of TP, TN, and NH3-N in the study area. Overall, the LBFGS-MLP model demonstrates significantly better accuracy on both training and testing datasets for the concentrations of TP, TN, and NH3-N compared to the Random Forest, CatBoost, and XGBoost models, particularly in the inversion of TN and NH3-N concentrations. The model's coefficients of determination (R2) are 0.71, 0.82, and 0.72, and the mean absolute errors (MAE) are 0.0118, 0.0394, and 0.0601 mg/L, respectively. The concentrations of TP, TN, and NH3-N in the study area are mainly distributed between 0.1~0.3 mg/L, 2~5 mg/L, and 0.1~0.4 mg/L, respectively, consistent with the survey results. Through this study, the effectiveness of the MLP algorithm in the inversion of non-optical water quality parameters is verified, providing a theoretical basis and reference for a more comprehensive assessment of the urban river water body condition.关键词：Non-Optical Water Quality Parameters;machine learning;hyperspectral remote sensing;concentration inversion- 60
- |
- 23
- |
- 0
更新时间：2024-06-18 -
The development and application of satellite-based limb scattering for atmospheric compositions AI导读
“临边散射卫星遥感技术为平流层臭氧监测提供数据支持。20余年来，OSIRIS、SCIAMACHY、OMPS和OMS等载荷搭载于多颗卫星，通过正向模型和反演算法，可获取O3、NO2、BrO廓线及平流层气溶胶和云信息。本文综述了该技术进展，为我国临边探测技术发展提供参考。”
摘要：ObjectiveSatellite-based limb scattering measurement technique had provided valuable data sets for long-term dynamic monitoring of stratospheric ozone and ozone-related atmospheric components, such as NO2 and BrO. Since 2001, the development of this field has gone through more than 20 years, we summaries the principle and application of this technique, meanwhile analyses the problem to provide reference for the development of domestic limb scattering detection technologyMethodSince the OSIRIS onbord the Odin satellite platform, followed by SCIAMACHY onboard the ENVISAT platform, OMPS onboard the Suomi NPP and NOAA-21 platform and OMS-L onboard the FY-3F platform, all include limb scatter detection capability. Based on the design specifications of the payloads, such as wavelength coverage, spectral resolution, signal-to-noise ratio and instrument response function et al, with a radiative transfer model (RTM) capable of simulating the observed limb radiance at a series of tangent heights, O3, NO2, BrO profile, stratospheric aerosol and cloud information can be retrieved from limb scattering spectra. In this paper, we reviewed the development of satellite-based limb scattering technique, including instruments characteristics, RTM, inversion algorithms, products and applications.ResultIn terms of forward models, the simulation of limb scattering needs to consider atmospheric scattering (single and multiple scattering), refraction, aerosol parameterization schemes and instrument characteristics under full spherical atmospheric conditions. In the aspect of retrieval algorithms, the wavelength shift correction, pointing information correction and stray light correction are needed to construct observation vectors for atmospheric parameters. The retrieved parameters had played an important role in analyzing stratospheric ozone dynamics and its related nitrogen oxides (e.g.NO2) and halogen (e.g.BrO), as well as monitoring stratospheric clouds and aerosols.ConclusionOverall, limb scattering satellite remote sensing technology can provide 2-3km vertical resolution and almost global coverage detection capability due to its advantages in sampling frequency and observation geometry. However, there are still unresolved problems in limb scattering technology. For the forward model, a fast radiative transfer model specific for limb scattering sensors is critical to meet the needs of operational application. In addition, the limb scattering sensors all have the problem of pointing information error and the field of view is subject to the pollution of stray light. Accurate laboratory calibration and further analysis of the source of pointing error are effective ways to correct the influence of stray light and the registration error of tangent height. Meanwhile, accurately characterizing the aerosol characteristics and cloud top height on the limb path are also key steps to reduce the inversion uncertainty. This paper can facilitate the development and application of domestic limb scattering detection technology.关键词：limb scattering;stratospheric ozone;RTM;Inversion algorithm;application progress- 16
- |
- 20
- |
- 0
更新时间：2024-06-18
- Postal code：100190
- Tel：010-58887052 Email：nrsb@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备20021838号-9
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0