最新刊期

  • GAO Weiqiang,HAO Xiaohua,HE Dongcai,SUN Xingliang,LI Hongyi,REN Hongrui,ZHAO Qin

    Corrected Proof
    DOI:10.11834/jrs.20242483
    摘要:The abstract of this study contains four sections objective, method,result and conclusion High Mountain Asia (HMA) is the richest high altitude region in the world except for the poles in terms of glacier and snow resources, The accurate monitoring of HMA snowpack distribution is important for HMA snowmelt runoff simulation, climate change prediction and ecosystem evolution. Fractional Snow Cover (FSC) can quantitatively describe the extent of snow cover at the sub-image scale, and is more suitable for reflecting the distribution of snow in complex mountainous areas than binary snow. The objective of this study is to develop a new HMA snow area ratio inversion algorithm and integrate the algorithm into Google Earth Engine to prepare a set of long time series HMA snow area ratio products.Method Considering the influence of HMA topography and sub-bedding type on the accuracy of snow accumulation information extraction, this paper proposes a Multivariate Adaptive Regression Splines (MARS) model LC-MARS to invert the proportion of snow accumulation area in Asia by integrating topography correction and subland class feature extraction. The FSC extracted by Landsat-8 is used as the true value, and the LC-MARS model is tested for inversion FSC accuracy using binary and error validation methods, and the performance of linear regression models trained with the same training samples and the LC-MARS model for inversion HMAFSC accuracy is compared, and the accuracy of the FSC inversion of the LC-MARS model with SnowCCI and MOD10A1 is also compared.Result (1) The overall accuracy of FSC binary validation of LC-MARS model inversion showed that Accuracy and Recall were 93.4% and 97.1%, respectively, and the overall accuracy of error validation showed that RMSE was 0.148 and MAE was 0.093, both binary validation and error validation indicated that the FSC accuracy of LC-MARS model inversion was higher. (2) The LC-MARS model trained based on the same training samples has higher FSC accuracy than the linear regression model in forest area, vegetation and bare land inversions, indicating that the LC-MARS model is more suitable for FSC inversions in mountain and forest areas. (3) The overall RMSE of MOD10A1 is 0.178 and MAE is 0.096; the overall RMSE of SnowCCI is 0.247 and MAE is 0.131. The accuracy of FSC prepared by LC-MARS is higher than that of MOD10A1 and SnowCCI, indicating that FSC inversion by LC-MARS has some application value.Conclusion The LC-MARS model can fit high-dimensional nonlinear relationships and significantly improve the inversion accuracy of FSC in mountain and forest areas. The computational efficiency of the LC-MARS model based on Google Earth Engine is high, and it is suitable for preparing FSC products of large scale long time series. In this study, the day-by-day MODIS FSC products of HMA from 2000 to 2021 were prepared based on the LC-MARS model, which provides important data support for the study of climate change, hydrological and water resources in HMA.  
    关键词:High Mountain Asia;Fractional snow cover;MODIS;MARS;Terrain correction   
    6
    |
    2
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 55033313 false
    发布时间:2024-04-10
  • Fan Zhanling,Chen Chongcheng

    Corrected Proof
    DOI:10.11834/jrs.20232592
    摘要:The cultural tourism industry emphasizes creativity and scene experience, which is highly compatible with the metaverse. The research of the cultural tourism metaverse is in the infancy stage. The concept, key technologies, and application scenarios are still being explored.Firstly, the paper reviews the evolutionary process of the metaverse. The basic definition, conceptual model, and main features of the cultural tourism metaverse are proposed. The paper considers the cultural tourism metaverse as a subsystem of the metaverse. It is a cultural tourism Internet form formed by existing information technology in the cultural tourism industry. It is a reconfiguration and virtual symbiosis of cultural tourism activities in a three-dimensional digital world. It provides an immersive experience of cultural tourism scenes based on extended reality technology. Based on the digital twin technology, it generates a mirror image of the real-world cultural tourism scene. And relying on the political, economic, and cultural systems under the unified architecture of the metaverse, it realizes the all-around integration of the virtual world and the real world of the cultural tourism industry. The conceptual model of the cultural tourism metaverse consists of key technologies, guarantee systems, participating subjects, and product services. The development of the cultural tourism metaverse cannot be achieved without the support of key technologies. As an industry metaverse application, it does not exist in isolation. Its smooth and reliable operation is closely related to the social, political, economic, cultural, legal and moral system established by the future holistic metaverse. The participating subjects of the cultural tourism metaverse include suppliers, consumers, developers, governors and researchers. It is immersive, interactive, customized, cultural, educational, connected, and interdisciplinary, etc.Secondly, the paper summarizes the research progress and application of key technologies in the field of cultural tourism. These key technologies include basic support technologies (such as intelligent communication, internet of things perception, artificial intelligence, positioning and navigation, big data computing and storage), virtual-real connection technologies (such as virtual geographic environment, digital twin, virtual digital human, decentralization) and virtual-real interaction technologies (such as extended reality, brain-computer interface and video games). The cultural tourism metaverse involves far more information technologies than what is mentioned in the paper. Moreover, all information technologies developed so far in the real world will be reflected in future metaverse scenarios or continuously upgraded and improved.Lastly, we look forward to the application scenarios of the cultural tourism metaverse. Such as cultural heritage digitization and protection, scenic spot (hotel) development and management, guided tour service, cultural tourism marketing, industry supervision, etc. The paper argues that the Lfuture research directions of the cultural tourism metaverse include personalized construction of cultural tourism virtual scenes, rapid migration of existing virtual reality cultural tourism scenes, integration of virtual and real cultural tourism scenes, seamless positioning and navigation of the cultural tourism metaverse, visitor experience and interaction, business models and monetization, and interoperability standards for virtual and real cultural tourism scenes. Metaverse development also faces issues such as ethics, security and privacy, technology, and the challenge of realistic national sovereignty. While technology is always hovering, these technologies will eventually reshape the future shape of society and human life.  
    关键词:Metaverse;Cultural tourism;virtual geographic environment;Virtual and reality integration;Virtual world;Extended reality   
    317
    |
    158
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 37558168 false
    发布时间:2024-04-09
  • Zhang Dongyun,Luo Jiancheng,Wu Tianjun,Dong Wen,Sun Yingwei,Yang Yingpin,Hai Yunrui,Meng Bowen,Liu Wei

    Corrected Proof
    DOI:10.11834/jrs.20241349
    摘要:Agriculture is an important component of the national economy. Obtaining the spatial distribution information of crops accurately is the basis of precision agricultural application. In this paper, we explore the complementary advantages of remote sensing data from different sources in terms of spatial and temporal characteristics, and design a remote sensing mapping method of agricultural planting structure based on spatial and temporal collaboration. The cropland-parcel extracted from high spatial resolution remote sensing images was taken as the basic unit, and the spectral time sequence information of multi-temporal remote sensing images was combined. With the support of deep learning, crop classification and identification at cropland-parcel scale and precise mapping of planting structure can be realized, and the spatial distribution characteristics of main crops can be analyzed. The experimental results in Yellow River Irrigation Area of Ningxia (YRIA-NX) show that : (1) A total of 1.49 million cropland-parcel were obtained in the study area, with a total area of about 540,000 hm2, and the overall classification accuracy was 0.80; (2) compared with the traditional mapping unit and machine learning method, the crop planting structure information obtained by the Bi-LSTM network based on the cropland-parcel scale is more consistent with the actual agricultural tillage management unit, and the classification accuracy can be guaranteed higher; (3) the maize, rice, wheat and vegetables are the main crop of the study area. Maize is the most dominant crop with the largest planting area and the most extensive spatial distribution. The vegetable fields are mainly concentrated distribution in Yongning Town and Qingtongxia Town. Rice is concentrated in areas with convenient irrigation while wheat planting area on a large scale is less. The planting of other crops after wheat harvest in summer is mainly concentrated in the Qingtongxia and the multiple cropping index decreased gradually from south to north.  
    关键词:crop planting structure;spatial-temporal collaboration;deep learning;cropland-parcel;NDVI time series;Yellow River Irrigation Area of Ningxia (YRIA-NX)   
    19
    |
    2
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 55013471 false
    发布时间:2024-04-09
  • WANG Baozhen,REN Huazhong,MU Xiaodong,ZHU Jinshun,LIU Rongyuan,ZHAO Cong

    Corrected Proof
    DOI:10.11834/jrs.20243236
    摘要:Objective The observation of geostationary orbit satellite, with the characteristics of wide range, high frequency and fixed point observation, provides an important way to obtain land surface and atmospheric parameters and can monitor the change of land surface temperature over long time series. As the second generation of China’s geostationary meteorological satellites, Fengyun-4 A and B satellites constellation observation greatly expand the scope of meteorological observation and improve data utilization efficiency. Before the joint use of the two satellites’ data, the radiometric consistency between the same bands observation of the two satellites needs to be explored.Method Taking the thermal infrared band data as an example, the radiometric consistency between three thermal infrared bands (centered at 8.5µm, 10.8µm and 12.0µm) of Fengyun-4 A/B satellites was studied in four experimental areas: Dunhuang calibration field, Hulunbuir Grassland, Chaohu Lake, and South China Sea. The heterogeneity of the Dunhuang calibration field and the Hulunbuir Grassland study area was evaluated firstly. A method was then proposed to correct the angular effects and spectral response function differences in observation from geostationary orbit satellites. This method corrected the angular effects by establishing empirical relationships between simulated radiance data at different viewing angles and eliminated spectral response function differences by correcting the brightness temperature and radiance on basis of the lookup tables. Finally, this method was used to correct the thermal infrared data of the two satellites at the same observation time, and the brightness temperature after correction were compared to analyze the radiometric consistency of the corresponding thermal infrared bands.Results Based on the data analysis, after removing the difference of space-time, observation angle and spectral response function of two satellites, the results show a strong positive correlation between the brightness temperatures of the three thermal infrared bands on Fengyun-4A and Fengyun-4B satellites. The brightness temperature errors are small, indicating good radiometric consistency. However, there is still a slight variation in brightness temperature among those different thermal infrared bands. The consistency of brightness temperature of the second thermal infrared band is better than that of the third band, and the third band performs better than the first band. The root mean square errors of brightness temperatures for the three bands range from 0.28 to 1.51 K, with deviations between -1.13 and 0.85 K. The brightness temperature deviations in the second and third bands exhibit a noticeable positive skewness, while the deviations in the first band shows a negative skewness.Conclusion By comparing the brightness temperature data before and after correction, it is not difficult to find that the method proposed in this paper has good applicability for the study of radiometric consistency in thermal infrared band of geostationary orbit satellite. The results show that the radiometric consistency of thermal infrared radiation between the two satellites is generally good although it might be influenced by land cover types. The findings in this paper provides important guidance for the joint utilization of thermal infrared data from Fengyun-4A and Fengyun-4B satellites.  
    关键词:remote sensing;geostationary orbit satellite;Fengyun-4;AGRI;thermal infrared radiation;LST;brightness temperature;consistency comparison   
    15
    |
    0
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54977809 false
    发布时间:2024-04-09
  • Zhao Cong,Xu Wei,Zhang Zhaoxu,Wu Zihua,Han Guhuai,Qin Qiming

    Corrected Proof
    DOI:10.11834/jrs.20243286
    摘要:Objective Coalbed methane (CBM) is a self-generated, self-storage, unconventional clean energy source which exists in coal seams and surrounding rock formations. Studies have shown that hydrocarbon can leak to the surface along the fractures, joints, microfractures, and pores of the upper overburden in oil or gas-rich areas, known as hydrocarbon micro-seepage. In CBM-enriched areas, hydrocarbon micro-seepage changes the chemical composition and chemical environment, and blocks the respiration of plant roots, so that the chloroplast synthesis process will be hindered, inhibiting photosynthesis and vegetation growth, eventually leading to abnormal changes in vegetation solar-induced chlorophyll fluorescence (SIF). Therefore, the abnormality of vegetation SIF can be an essential clue for remote sensing surveys of potential CBM enrichment areas.Method In this paper, convolutional neural networks (CNNs) are applied to the downscaling of SIF. 1-km SIF data of the study area between 2000 and 2020 are obtained, taking the southern part of Qinshui Basin (Qinshui County) in Shanxi Province as the study area. The vegetation-covered region in the study area is classified into farmland, grassland, and woodland. The spatial and temporal variation characteristics of vegetation SIF in CBM-enriched areas are categorized and discussed. At the same time, the possible reasons for the abnormal variation of vegetation SIF in CBM-enriched areas are analyzed and elucidated through the comparison between the study area and the control area.Result The results show that among the three vegetation types, the average SIF of woodland is the highest, followed by grassland, and the average SIF of farmland is the lowest; each vegetation type in the study area shows a noticeable trend of SIF growth from 2000 to 2020, and the SIF growth rates of vegetation in the study area from June to September are 0.0058, 0.0052, 0.0036, 0.0021 W·m-2·μm-1·sr-1·a-1.Conclusion In this paper, it is found that the SIF values of woodland in the CBM-enriched area that are less affected by artificial factors are significantly lower than those of the control area, and the average growth rate from 2000 to 2020 is also lower than that of the control area. This indicates that forest land is more likely to be subjected to the stress of hydrocarbon micro leakage in coalbed methane enriched areas compared to other influencing factors. This empirically shows that hydrocarbon micro-seepage in CBM-enriched areas may have a significant impact on vegetation SIF.  
    关键词:coalbed methane;solar-induced chlorophyll fluorescence;spatiotemporal change;vegetation growth stress;hydrocarbon micro-seepage;downscaling;convolutional neural network;Qinshui basin   
    8
    |
    1
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54977769 false
    发布时间:2024-04-09
  • JIAO Miao,QUAN Xingwen,HE Binbin,YAO Jinsong

    Corrected Proof
    DOI:10.11834/jrs.20243082
    摘要:Objective Forest wildfires have occurred frequently in Sichuan province in recent years, posing a significant threat to local ecological security as well as the lives and property of people and rescue workers. The purpose of this study was to investigate the temporal and spatial characteristics of forest grassland fires in Sichuan Province from 2001 to 2021, and to provide useful information for fire prevention and control decisions.Method Based on MCD64A1, Fire_CCI51 and MCD14ML multi-source remote sensing fire products, extracted effective fire points, acquired regional fire data, and explored the temporal trend and spatial distribution of forest and grassland fires by using geographic information system. The relationship between climatic, combustible, and topography environment factors and fire was examined using mathematical statistics and an adaptable fuzzy neural network.Result The results showed that the fire frequency and fire area increased from 2001 to 2014, and the fire occurred frequently from January to May. The spatial distribution of grassland fires is heterogeneous, mainly concentrated in the southwest of Sichuan Province, while the northeast of China has increased significantly recently. In the correlation analysis of various influencing factors, the correlation between forest fire and fuel water content is high, and environmental variables are the main driving factors of temporal and spatial characteristics of forest fire. The correlation between grassland fire and meteorological factors is high, but it is speculated that human factors have great influence on grassland fire characteristics.Conclusion Based on the analysis of the temporal and spatial characteristics of fire in Sichuan province, provides a decision basis for forest grassland fire prevention and control policy, early warning and monitoring in this region.  
    关键词:remote sensing;Sichuan province;MCD64A1;Fire_CCI51;MCD14ML;forest grassland fire;spatial distribution;time trend;spatio-temporal characteristics   
    72
    |
    24
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54702672 false
    发布时间:2024-04-03
  • Zhao hewen,Chen tao

    Corrected Proof
    DOI:10.11834/jrs.20243530
    摘要:Predicting land subsidence is crucial for conducting in-depth analyses and providing early warnings about urban land subsidence patterns. However, traditional numerical prediction models often struggle to accurately capture the intricate characteristics of land subsidence data, leading to less precise predictions. This study focuses on Xinmi City and endeavors to improve the accuracy of land subsidence prediction by combining time series feature extraction methods with time series prediction techniques.In this study, 232 interference images provided by HyP3 were utilized to acquire land subsidence information in Xinmi City spanning from January 2018 to December 2022, employing SBAS-InSAR technology. Recognizing the challenge of achieving high accuracy in directly predicting land subsidence data, this study developed a land subsidence prediction model that integrates trend and seasonal characteristics using Long Short-Term Memory networks (TSC-LSTM, Trend Seasonal Characteristics-LSTM). The TSC-LSTM model capitalizes on the strengths of weighted regression seasonal trend decomposition (STL) for extracting time series features from settlement data and the long short-term memory model (LSTM) for addressing the vanishing gradient problem in time series prediction. This fusion of techniques allows for a more precise analysis of land subsidence data and enables highly accurate predictions. Distinguishing itself from the conventional LSTM model, the TSC-LSTM model refrains from directly inputting ground subsidence data. Instead, it employs STL to meticulously extract both trend and seasonal characteristics from the land subsidence data. This approach maximizes the utilization of characteristic information inherent in the land subsidence data. Subsequently, these features are fed into the LSTM model for prediction. This unique methodology reduces noise interference and significantly enhances the accuracy of model predictions.This research leverages time-series InSAR data from 2018 to 2022 for Xinmi City, employing the TSC-LSTM model, deep learning architectures (RNN and LSTM), and conventional machine learning algorithms (MLP and SVR) to forecast the cumulative subsidence data for five subsidence centers using SBAS-InSAR. It identifies the two most optimal models and validates their efficacy in single-point prediction scenarios, utilizing domain-specific terminology. Research findings indicate the following:(1) Between 2018 and 2022, Xinmi City experienced a land subsidence rate ranging from -60.3 to 51.96 mm per annum, resulting in the identification of five distinct land subsidence center areas. Among these, the highest cumulative settlement and uplift reached 304.9 mm and 197.68 mm, respectively. The universality of the TSC-LSTM model across diverse datasets has been corroborated, demonstrating its high precision, exceptional generalization capability, and stable high performance in the prediction of land subsidence, employing specialized terminology. (2) The TSC-LSTM model exhibited exceptional performance in predicting the five subsidence center areas. The R² values for the TSC-LSTM model range from 0.9985 to 0.9992, significantly surpassing the second-best model, LSTM, which has an R² range of 0.9662 to 0.9872. Moreover, the RMSE values for the prediction accuracy of the TSC-LSTM model were less than 2 mm, achieving a range of 1.2426 to 1.7403 mm. (3) Single-point prediction results demonstrate the TSC-LSTM model's superior ability to accurately capture local changes in the cumulative settlement data. The TSC-LSTM model proposed in this study outperforms the traditional LSTM model in terms of prediction accuracy and model stability, thereby providing robust support for in-depth research on urban land subsidence.  
    关键词:Ground settlement prediction;TSC-LSTM;SBAS-InSAR;Cumulative settlement data decomposition;Xinmi City;LSTM   
    34
    |
    9
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54662966 false
    发布时间:2024-04-02
  • HE Jiayue,SU Nan,XU Congan,YIN Lu,LIAO Yanping,YAN Yiming

    Corrected Proof
    DOI:10.11834/jrs.20243249
    摘要:In recent years, there has been significant interest and focus on Synthetic Aperture Radar (SAR) ship detection. Its distinctive strengths position it as a pivotal player in numerous realms of research. However, the inherent characteristics of SAR images have also introduced a range of challenges. For instance, in contrast to optical images, the characteristics of SAR imaging lead to its feature representation are not intuitive. Additionally, due to the constrained number of SAR image data, achieving satisfactory results with existing methods that depend on a substantial number of annotated SAR images might be challenging.The exploration of how to effectively utilize a limited quantity of SAR images to learn a high-performance SAR ship detection network is a question worth delving into. It is clear that there are inherent limitations in single-modality SAR detection algorithms. Therefore, we can choose to leverage other effective modalities to assist the SAR modality in completing tasks. For instance, in the task of SAR image target detection, optical images can serve as a supplementary data source. Utilizing a large volume of optical data to assist in training SAR data contributes to the development of a more knowledge-rich model. Hence, exploring the most reasonable training approach to effectively utilize images from both SAR and optical modalities is a question worth investigating.To tackle these challenges, a SAR ship detection algorithm called MCMA-Net, which is based on multi-level cross-modality alignment is proposed in this paper. MCMA-Net enriches the SAR feature representation by incorporating valuable knowledge from optical modality. Firstly, we propose a neighborhood-global attention-based feature interaction network (NGAN), which employs a neighborhood attention mechanism to enable local interaction of low-level features and a global self-attention mechanism to capture global context from high-level features. While taking into account the ability of global context modeling, the encoding ability of local features is improved, NGAN makes the network pay more attention to the corresponding information at different levels and can promote the subsequent multi-level modality alignment. Secondly, we propose a multi-level modality alignment module (MLMA), which aligns the features in different hidden spaces of the two modalities from three levels. MLMA facilitates the model to acquire modality-invariant features, bridging the modality gap and realizing the optical knowledge transmission. The valuable information from the optical modality can compensate for certain deficiencies in SAR images. With the aid of these two modules, we have incorporated optical superiority information upon the foundation of SAR's inherent advantages, achieving an enhancement in the performance of SAR detection tasks.Plenty of experiments show that our algorithm is superior to the current detection algorithm. It is gratifying to note that, whether on publicly or our own SAR image dataset, MCMA-Net has consistently achieved optimal detection results. This further proves that MCMA-Net is a model with stable performance and robustness. The visualized results indicate that MCMA-Net achieves superior detection capabilities in complex scenarios. The ablation experiments demonstrate that compared to the baseline model, our algorithm achieved a 2.7% increase in mAP on the SSDD dataset. Various experimental results have consistently validated the rationality of MCMA-Net.  
    关键词:SAR;target detection;cross-modality;feature alignment;attention mechanism   
    21
    |
    10
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54654173 false
    发布时间:2024-04-02
  • LIU Jingwen,LIU Jianzhong,ZHU Kai,ZHANG Jingyii,ZHANG Ke,WU Congzhei,LEI Danhong

    Corrected Proof
    DOI:10.11834/jrs.20243512
    摘要:Based on the Chang'E lunar exploration data and other lunar geological data, the team led by the Institute of Geochemistry of the Chinese Academy of Sciences has compiled a series of lunar global geological maps. These maps include a 1:2,500,000-scale global geologic map, a digital global lithologic map, and a global tectonic map. At the same time, 30 lunar quadrangle geological maps were also compiled, including the LQ-20 Nectaris Quadrangle map. The LQ-20 map is eye-catching because it depicts the impact event of the Mare Nectaris Basin, which divided the Nectarian and Aitkenian periods, during the Nectarium classic formation event.This article provides a detailed overview of the development history of lunar geological mapping. It discusses the various geological features observed in the LQ-20, including rock types, impact basins, crater materials, and structural characteristics:12 rock types, basin formations on 8 impact basins, crater materials from 260 impact craters, and 16 types of structural characteristics, which covering geological information in the sixth periods.Finally, we discussed the regional geological evolution history of the LQ-20 map from the Magmatic Oceanic Period to the Copernican Period. We also compared the expression of geological features in the LQ-20 map to that of the USGS lunar geological map, highlighting the differences and advantages: the fundamental distinction in conveying geological information lies in the comprehensive establishment and systematic representation of the overall lunar geological processes as an integrated geological system; this involves providing a clear explanation of the geological causes for each unit, ensuring that each geological feature conveys its origin; furthermore, each region is characterized by a distinct hierarchy of geological units arranged in chronological order. This project would provide lunar geological background information for lunar exploration, scientific research, and related endeavors, benefiting not only China but also the global communities.  
    关键词:lunar geological map;LQ-20;Mare Nectaris;Remote Sensing Interpretation;Geological Evolution History   
    16
    |
    4
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54654131 false
    发布时间:2024-04-02
  • WEN Qian,AN Yuan,CHEN Guangjian,WU Ruijiao,LIN Jiawen

    Corrected Proof
    DOI:10.11834/jrs.20243227
    摘要:Band selection is a crucial task in the dimensionality reduction of hyperspectral remote sensing imagery. Its objective is to choose a subset of bands that contain less redundant information, higher information content, and exhibit class discriminability. To address the issues of existing band selection methods based on nearest neighbor subspace partitioning, which do not consider the spatial distribution of objects and neglect the impact of noisy bands when computing cluster centers, this paper proposes a hyperspectral image band selection method that integrates spatial spectral structure with improved local density, referred to as ISSS-ELD.This method first performs image segmentation on the hyperspectral image using entropy-based approach to acquire homogeneous regions. The composite region-level neighboring band correlation coefficient vector is obtained by integrating the correlation coefficient matrix of these homogeneous regions. Subsequently, a Gaussian kernel is applied to globally smooth the neighboring band correlation coefficient vector, reducing the influence of noisy bands. Bands are grouped based on extremum points in the smoothed vector. Finally, the product of the maximized improved local density and band information entropy serves as the criterion for selecting representative bands.This paper conducted experiments on hyperspectral image datasets including Indian Pines, Botswana, and Salinas. Different band selection methods were evaluated by calculating metrics such as classification accuracy, average correlation coefficient, and noise robustness of the selected bands. The results indicate that: (1) Compared to pixel-level correlation-based partitioning methods, the utilization of region-level correlation coefficients results in more reasonable grouping of neighboring bands, reducing band redundancy while retaining some potential characteristic bands. The classification performance on the three datasets improved by 2.63%, 0.68%, and 0.16% respectively. (2) In contrast to methods solely using information entropy for band assessment, the proposed approach of maximizing the product of improved local density and information entropy proves effective. On the three datasets, Overall Accuracy (OA) increased by 4.13%, 0.5%, and 0.21% respectively. (3) Compared to six other advanced band selection methods, the proposed method achieved significant performance improvements: OA increased from 62.34% to 75.03%, from 86.74% to 88.28%, and from 86.04% to 92.36% on the three datasets. Furthermore, the selected subset of bands by our method is dispersed, concentrating in regions with higher information entropy, and effectively avoiding the inclusion of noisy bands.In summary, the band subset selected by the proposed band selection method exhibits low redundancy, high information content, strong class separability, and robustness against noise, effectively addressing the challenges in hyperspectral image band selection.  
    关键词:remote sensing;hyperspectral images;band selection;subspace division;peak density;local density;information entropy;classification   
    20
    |
    16
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54647591 false
    发布时间:2024-04-01
  • DUAN Dexin,LU yao,HUANG Liwei,LIU Peilin,WEN Fei

    Corrected Proof
    DOI:10.11834/jrs.20243269
    摘要:(Objective)In the realm of intelligent processing of remote sensing images, the third-generation brain-inspired spiking neural network (SNN) surpasses its predecessor, the second-generation artificial neural network (ANN), with remarkable advantages in terms of high energy efficiency, high precision, and high interpretability. These advantages stem from the SNN's characteristic features, including superior energy efficiency, elevated sparsity, and remarkable bio-plausibility. The integration of these features in the SNN presents an enthralling solution to the challenges faced in remote sensing image processing and holds immense potential for advancing this field further.(Method)The proposed algorithm introduces a novel approach by initially establishing a target detection neural network, which serves as the source artificial neural network for pre-training. This source network utilizes a dynamic clipping threshold activation function to optimize its performance. Subsequently, the algorithm transforms the source network into a brain-inspired spiking neural network by leveraging the mapping relationship between activated neurons and spiking neurons. This conversion process effectively incorporates the clipping thresholds obtained during training. By seamlessly transitioning the source network into an SNN, the algorithm ensures the preservation and enhancement of key characteristics essential for remote sensing image processing.(Result)To evaluate the efficacy of the proposed method, extensive experiments were conducted on two widely recognized open remote sensing datasets, namely SAR-Ship-Detection-Datasets (SSDD) and RSOD. The experimental results highlighted the exceptional capabilities of the proposed method in transforming the source network into a brain-inspired spiking neural network, with negligible loss in performance. Furthermore, the transformed SNN exhibited remarkable accuracy in detecting and recognizing remote sensing targets within significantly reduced time steps. The performance achieved by the transformed SNN was comparable to that of the source artificial neural network while tremendously reducing power consumption by over two magnitudes. This outcome highlights the immense potential of the proposed method in revolutionizing the field of remote sensing image processing by delivering high precision and interpretability, paired with significantly reduced energy consumption.(Conclusion)The integration of the third-generation brain-inspired spiking neural network (SNN) into the domain of remote sensing image processing holds tremendous potential, primarily due to its remarkable advantages in high precision and low energy consumption. The proposed algorithm highlights the distinctive attributes of the SNN, such as its low delay, high bionics, and the ability to inherit the high precision observed in the source network. These characteristics are promising indicators of the SNN's capability to significantly enhance the intelligent processing of remote sensing images.By leveraging the SNN's high precision and embracing bio-plausible principles, the proposed algorithm lays a robust foundation for future advancements in the field of remote sensing. The SNN's inclusion in the image processing pipeline brings forth a paradigm shift, challenging traditional assumptions and unlocking new possibilities. The application of the SNN in remote sensing target detection offers high accuracy in identifying and classifying targets with remarkable precision. In conclusion, the proposed algorithm exhibits characteristics of low delay and high bionics while inheriting the high precision of the source network. This indicates its potential to significantly improve the intelligent processing of remote sensing images, offering high accuracy, precision, interpretability, low energy consumption and high bio-plausibility.  
    关键词:SAR;optical remote sensing;target detection;deep learning   
    24
    |
    6
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54637555 false
    发布时间:2024-04-01
  • Xie Xinyao,Li Ainong,Tian Jie,Wu Changlin

    Corrected Proof
    DOI:10.11834/jrs.20243038
    摘要:Objective Mountain ecosystems with coverage of approximately 24% of the terrestrial surface,are the key component of earth’s carbon cycle in terrestrial ecosystems. Vegetation in mountain ecosystems can regulate the energy budget via mediating the exchange of energy and substance, and thus has been regarded as an essential bio-indicator for the global climate change over the past decades. Accurate estimation of mountain vegetation gross primary productivity (GPP) plays a vital role in understanding the function of mountain ecosystems and characterizes the ecosystem responses to climate change. Due to the effect of complex mountainous conditions and the limitations from the spatial resolutions, there are obvious topographic errors and spatial scaling errors in mountain vegetation GPP estimates. Thus, it is crucial to evaluate the error sources in estimating mountain vegetation GPP across multiple spatial scales.Method In this paper, we selected the Wanglang National Nature Reserve - a typical mountainous ecosystem of southwest China as the study area. This study used an eco-hydrological model called Boreal Ecosystem Productivity Simulator (BEPS)-TerrainLab to obtain the vegetation GPP and analyze the topographic errors and spatial scaling errors at the fine, medium, and coarse spatial scales (i.e., 30 m, 480 m, and 960 m). At the fine, medium, and coarse spatial scales, the topographic errors in estimating vegetation GPP were evaluated across four scenarios that characterize the effects of different topographic features. Spatial scaling errors were illustrated at the scales of 480 m and 960 m, respectively. Finally, the agreement index (d), determination coefficient (R2), root mean square error (RMSE), and mean bias error (MBE) were used to evaluate the topographic errors and spatial scaling errors in modeling mountain vegetation GPP at the fine, medium, and coarse spatial scales.Result Results showed that the multi-scale vegetation GPP estimates across different simulation conditions presented obvious spatial differences (the difference among regional mean value upped to 198 gC m-2 yr-1). The topographic errors of vegetation GPP estimates showed a decreasing trend with the decrease of spatial resolution, suggestting that more attention should be paid to high spatial resolution (the MBE value is 200 gC m-2 yr-1). Specifically, the error caused by ignoring the redistribution of soil water was observed to be the largest source of topographic errors. As for the spatial scaling errors, an increasing trend with the decrease of spatial resolution was found, highlight the necessity of reducing the spatial scaling errors in middle and coarse spatial resolution GPP estimates (161 and 210 gC m-2 yr-1).Conclusion During the process of generating the multi-scale mountain vegetation GPP products, it was necessary to remove the topographic effects on high spatial resolution GPP estimation. Simultaneously, attention should also be given to the spatial scaling errors of GPP products at middle and coarse spatial resolutions. Considering the obvious topographic errors caused by ignoring the water redistribution, accurate estimation of soil moisture would improve the quality of GPP products over mountainous areas, especially these products at high spatial resolution.  
    关键词:Gross primary productivity;mountain ecosystems;multiple spatial scales;topographic errors;spatial scaling errors   
    14
    |
    7
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 54411976 false
    发布时间:2024-04-01
  • TIAN Fuyou,CAO Yupei,ZHAO Hang,WU Bingfang,ZENG Hongwei,LIU Yazhou,QIN Xingli,ZHANG Miao,ZHU Liang,ZHU Weiwei

    Corrected Proof
    DOI:10.11834/jrs.20243191
    摘要:Objective Accurate identification of agricultural fields, the smallest units of agricultural farming, is crucial for monitoring land resources and arable land utilization. Manual mapping methods are time-consuming and expensive, incapable of real-time or near-real-time updates. To enhance agricultural field delineation efficiency, we introduce Field-Net, a field segmentation model leveraging spatial attention mechanisms and multi-task learning in this research.Method Field-Net is based on the UNet architecture, integrating spatial attention mechanisms and a multi-task learning approach. In addition to the segmentation task, we incorporate boundary identification and distance to the field boundary as two additional tasks, enabling the model to learn more representative features related to fields. The model's performance was evaluated using GF-1 and ZY-3 satellite images with a spatial resolution of 2 meters in Lijin County, Dongying City, Shandong Province. We labeled 3,480 tiles measuring 256x256 pixels in the YanWo district, with 3,000 used for training, 360 for validation and 120 for spatial generalisation performance test.Results Initially, we analyzed the loss weights for the three tasks—mask, boundary, and pixel-to-boundary distance—in multi-task learning using a gradient test. We found that for multi-task learning, the loss weights should prioritize the mask segmentation task as the primary task and the others as secondary. Across the entire test set, Field-Net achieved an overall accuracy of 92.23% and an IOU of 87.05%. We compared Field-Net with four state-of-the-art architectures: DeepLabv3+, HRNet, LinkNet, and D-LinkNet. Field-Net outperformed them all in semantic segmentation tasks, with an IOU 0.26% higher than Link-Net, the most accurate among the four selected models, and 7.59% higher than DeepLab v3+. In the spatial generalisation performance test, the average IOU of Field-Net model is 3.51% higher than that of Link-Net model, and the spatial generalisation performance is significantly improved. Ablation tests demonstrated that the spatial attention mechanism and multi-task learning strategy improved F1-Score by 1.01% and IOU by 1.6% compared to the ResUNet model. The multi-task learning strategy led to a 0.18% F1-Score improvement for Field-Net and a 0.21% improvement in IOU.Conclusion While challenges remain in identifying contiguous fields due to unclear boundaries, future enhancements could incorporate multi-temporal and higher-resolution remote sensing images to improve field feature discrimination. Feature visualization analysis revealed that the spatial attention mechanism and multi-task learning strategy enabled the model to learn clustered features at field boundaries and within plots, enhancing feature representativeness. Overall, the Field-Net model supports field-level monitoring of cropland use, including non-agricultural applications, such as grain production, enhancing the efficiency and timeliness of land resource monitoring. In the process of generating the field dataset in China, the complex and fragmented cropland bring considerable challenges to this task. In the future, the problem of lack of samples for model training can be solved by accumulating field segmentation datasets from different regions by borrowing the paradigm of Image-Net, while a more general model for channels, regions, and sensors should be constructed subsequently. In the future, with the arrival of the "large model" era of deep learning, for the task of parcel segmentation, it is also necessary to construct a model to segment every field from the perspective of both the model and the dataset.  
    关键词:Filed segmentation;Field-Net model;Spatial attention mechanism;Multi-task learning;GF satellite data   
    111
    |
    41
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50733504 false
    发布时间:2024-03-25
  • FENG Quanlong,JIANG Zihang,NIU Bowen,GAO Bingbo,YANG Jiangyu,YANG Ke

    Corrected Proof
    DOI:10.11834/jrs.20243139
    摘要:Black soil is a valuable and essential soil resource, particularly in the northeastern region of China, where it serves as the primary grain-producing area. However, the quality of local agriculture is significantly impacted by soil erosion, with erosion gullies representing a prominent manifestation of this issue. Erosion gullies, formed due to soil erosion, often interconnect within a hydrological network, creating a tree-like distribution of erosion gully systems that inflict severe damage upon cultivated land. Therefore, the accurate identification and detection of erosion gullies are pivotal for safeguarding arable land.Objective This study explores the feasibility of utilizing remote sensing imagery for erosion gully detection and identification, leveraging its vast coverage and multiple capture instances. Additionally, we introduce a novel deep learning model tailored for erosion gully recognition based on a multi-scale dense dilated convolutional neural network. Our model incorporates dense connections of multi-scale dilated convolutional residual modules, optimized to aggregate the multi-level spatial features of erosion gullies.Method The research was conducted within Hailun City, located in Heilongjiang Province, serving as the study area. Our approach involved cropping remote sensing images into predefined patches, which were then annotated to construct training datasets comprising two categories: erosion gullies and non-gullies. Subsequently, the model was trained on the training dataset and evaluated on the test dataset, with weight selection based on the highest test dataset accuracy. Utilizing these selected weights, we performed sliding window identification across the entire Hailun City area, thereby generating spatial distribution data for erosion gullies. Furthermore, we accomplished erosion gully area localization based on scene-level labels and class activation maps, offering guidance for boundary extraction.Results Our findings demonstrate the efficacy of the proposed model, achieving an impressive overall accuracy of 95.80% and a kappa coefficient of 0.9152. This outperforms traditional deep learning models such as GoogLeNet, ResNet, DenseNet, and Swin-Transformer. Notably, during the sliding window recognition phase, the overall accuracy dipped slightly compared to the test phase due to the increased complexity of remote sensing imagery in practical applications. To address this challenge, we recommend incorporating a fusion of remote satellite images and street view imagery in future research to enhance recognition capabilities in complex scenarios.Conclusion This study underscores the effectiveness of erosion gully identification through the application of a multi-scale dense dilated convolutional neural network. It serves as a valuable tool for providing precise spatial distribution data concerning erosion gullies, thereby contributing to integrated land management in the black soil region of Northeast China.  
    关键词:erosion gullies;Black soil region in Northeast China;deep learning;Feature extraction;Scene recognition;Dilated convolutional neural network;remote sensing monitoring;Cropland protection   
    119
    |
    60
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50143219 false
    发布时间:2024-03-12
  • KE LiWei,YANG JiuChun,WANG JiaQi,LI Ying

    Corrected Proof
    DOI:10.11834/jrs.20243429
    摘要:This study aims to conduct a spatiotemporal analysis of the three-dimensional morphology of Erdao gully in Jiutai District, Changchun City, Jilin Province, to comprehend the erosion process and provide guidance for precise deployment of water and soil conservation measures in farmland channels in the Northeast Black Soil Region. Utilizing multi-year orthophoto images and high-resolution DEM data derived from drone remote sensing, two-dimensional parameters of the gully were extracted. Combining visual interpretation with auxiliary data, the overall development of Erdao gully over several years was determined, and a set of quantitative gully morphology indicators was established. The study analyzed the three-dimensional morphological changes in the most active eastern head of Erdao gully from 2017 to 2021. Simultaneously, rainfall-related indicators were collected and analyzed from the Jiutai Meteorological Station over five years to discuss the relationship between rainfall and gully erosion. The findings reveal that Erdao gully in the Black Soil Region area exhibited stable head development over five years, with gradual expansion and the most severe erosion occurring on the sunny slope of the head. The number and intensity of rainfall events were identified as the main factors influencing gully head expansion in this region. The process of gully three-dimensional morphological evolution based on drone remote sensing effectively portrays the erosion process in the Black Soil Region and offers valuable insights for the precise deployment of water and soil conservation measures in the future, contributing to land quality protection and ensuring food security in agricultural production areas.  
    关键词:UAV remote sensing;Thin Black Soil Region;Gully;Three-Dimensional Morphology;Gully Shape Index System;Soil Conservation;Soil erosion;Spatiotemporal Change Analysis.   
    63
    |
    38
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50119761 false
    发布时间:2024-03-12
  • SU Boyang,ZHAN Wenfeng,DU Huilin,JIANG Sida,WANG Chenguang,DONG Pan,WANG Chunli,LIU Zihan

    Corrected Proof
    DOI:10.11834/jrs.20243362
    摘要:Investigations into the diurnal evolution differences between surface urban heat islands and canopy urban heat islands (termed Is and Ic, respectively) hold significant values in enhancing our comprehension of the vertical structure of urban climates at a fine time-scale. However, both the hourly surface air temperature (Ta) from densely distributed weather stations within cities and the hourly land surface temperature (Ts) that possesses a relatively high spatial resolution and that can be employed for monitoring thermal conditions of urban surfaces are largely absent. Previous studies comparing hourly Is and Ic have mostly focused on individual cities. In this study, we utilize hourly Ta measurements from high-density meteorological stations (1544 stations) and Ts observations derived from a diurnal temperature cycle (DTC) model to examine the hourly Is, Ic, and the associated hourly differences (quantified as ΔUHI, calculated by subtracting Is from Ic) over 27 Chinese megalopolises. Furthermore, we analyze the hourly patterns of ΔUHI (e.g., maximum ΔUHI, minimum ΔUHI, and duration of ΔUHI > 0) across cities with different climate backgrounds and city sizes. We obtain the following findings: (1) At the national scale, the annual mean ΔUHI remains positive throughout the diurnal cycle. The hourly ΔUHI pattern generally exhibits a peak shape, with the ΔUHI increasing from morning and reaching its maximum (1.7 °C) at around 4:00 PM. Subsequently, it gradually decreases and reaches its daily minimum (0.1 °C) at around 2:00 AM, with the most rapid decline occurring around sunset. (2) Across different climate zones, from subtropical to temperate cities, both the maximum and minimum ΔUHIs follow a decreasing trend, the times at which they occur are gradually delayed, and the duration of ΔUHI greater than 0 °C gradually decreases. (3) For cities with different sizes, the variation magnitude of ΔUHI curve generally decreases and the time of minimum ΔUHI advances as city size increases. The duration of ΔUHI greater than 0 °C also increases with city size. We consider this study can promote the understanding of the contrasting patterns between hourly differences in surface urban heat islands and canopy urban heat islands across cities with diverse background climates. The research results contribute to a deeper understanding of the vertical spatial characteristics of urban heat islands at a fine time scale.  
    关键词:Surface urban heat island;Canopy urban heat island;Diurnal variation;Climate zone;City size;thermal infrared remote sensing   
    72
    |
    29
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50119722 false
    发布时间:2024-03-12
  • ZHANG Zhihao,WANG Qunming,DING Xinyu

    Corrected Proof
    DOI:10.11834/jrs.20243112
    摘要:Objective Fractional vegetation cover (FVC) is an important indicator to characterize the spatial distribution of vegetation on the land surface. Remote Sensing Satellites (such as Landsat and Sentinel-2) can acquire fine spatial resolution FVC data at 10 m level, which are crucial source for researches on global ecosystem. However, due to cloud contamination and limited temporal resolution of the satellites, a large amount of fine spatial resolution FVC data are not available in the temporal domain. This paper considers the collaboration of 30 m Landsat-8 and 10 m Sentinel-2 to increase the temporal frequency of the observations.Method To deal with the difference in the spatial resolution, a deep learning-based method named FVC-Net is proposed in this paper. FVC-Net fuses 30 m Landsat FVC with 10 m Sentinel-2 normalized difference vegetation index (NDVI) directly, producing 10 m Landsat FVC. Specifically, a two-branch network based on multi-scale attention mechanism is designed, in which the channel enhancement blocks are used in both FVC and NDVI branches for feature extraction and fusion. Then, the spatial attention blocks are used to increase the spatial details of the fused FVC features. The scheme designed in FVC-Net can help to characterize the non-linear relationship between 10 m NDVI and 30 m FVC effectively.Result In the experiments, the proposed FVC-Net was compared with four typical non-deep learning-based and four deep learning-based fusion methods. It was found that FVC-Net is consistently more accurate than the eight benchmark methods. The 10 m FVC results can display more spatial details than original 30 m FVC.Conclusion The proposed FVC-Net is an effective solution to downscale 30 m Landsat FVC to 10 m by fusion with 10 m Sentinel-2 NDVI, which can effectively overcome the differences between Sentinel-2 and Landsat data at different time points. FVC-Net has the potential to be applied to downscale the current 30 m Landsat FVC products at the global scale, of which the predictions can support the researches in the related fields greatly.  
    关键词:Fractional Vegetation Cover (FVC);normalized difference vegetation index (NDVI);deep learning;downscaling;data fusion   
    68
    |
    126
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50119651 false
    发布时间:2024-03-12
  • XU Yufei,SUN Rui,HUANG Xinyu

    Corrected Proof
    DOI:10.11834/jrs.20243323
    摘要:In recent years, the frequent occurrence of forest fires has brought great impact to people's normal work and life as well as the natural ecosystem. The assessment of fire hazard is of great significance to the prevention of forest fire and the allocation of fire resources. This paper collects the historical forest fire events in China from 2002 to 2020, which are distributed in 5 climate regions in China: plateau mountain climate, temperate continental climate, temperate monsoon climate, subtropical monsoon climate, and tropical monsoon climate., integrated meteorological factors, vegetation index, topographic factors in different regions, and used the random forest method to establish a comprehensive forest fire hazard assessment model. Fire influencing factors were calculated from different data products, fire events were selected using FIRMS images, meteorological factors were calculated using ERA5-land data, topographic factors were calculated using SRTM DEM products, and vegetation indices were calculated using MODIS reflectance product MCD43A4. The fire hazard assessment model can predict the time series of fire hazard and evaluate the spatial distribution of fire hazard. The fire occurrence location of the test data is different from that of the training data. From the test case results, the accuracy of the established fire hazard assessment model is relatively high, and the area under the ROC curve reaches 0.84, which achieves good results in the time series prediction of forest fire hazard and the spatial distribution assessment of forest fire hazard. Also, the predicted fire hazard value is close to the pre-calibrated fire hazard value. The results of time series prediction of fire hazard and evaluation of fire hazard spatial distribution are good, which are close to the actual situation. At the same time, the model established in this paper ranked the importance of the factors affecting the occurrence of fires. The most important factor was the annual diurnal sequence factor, reflecting the seasonal factor, followed by the moisture factor and the growth of vegetation, and the importance of topographic factors was lower. Importance ranking can help to understand the driving effect of different factors on the occurrence of fires, and understand which factors have a greater impact on the occurrence of forest fires. Although the area of forest fire occurrence is different and the factors affecting the fire occurrence are different, the change rule of fire hazard value was similar, the fire hazard value is higher in the week before the fire, and the fire hazard value is lower in other times. The spatial distribution of fire hazard is reasonable, and the fire hazard value in the fire area gradually increases from two months before the fire to the day of the fire. Also, the fire hazard in the same area 1 year before was significantly lower than the fire hazard value on the day of the fire, which can accurately assess the fire hazard situation. The forest fire hazard assessment model established in this paper involves comprehensive indicators, which can accurately assess the fire hazard situation. At the same time, it can be applied to different regions in China to partially solve the problem of regional restrictions.  
    关键词:forest fire;fire hazard;remote sensing;Random Forest;hazard monitoring   
    160
    |
    57
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50115691 false
    发布时间:2024-03-12
  • CHEN Man,HUANG Yongjie,XU Lei,PAN Zhisong

    Corrected Proof
    DOI:10.11834/jrs.20243407
    摘要:Objective Remote sensing image interpretation has essential application values in various fields, such as urban management, maritime monitoring, and resource planning. As an important and challenging task in remote sensing image interpretation, instance segmentation of remote sensing images can achieve target-level localization and pixel-level classification of objects of interest with fine granularity, making it a current research hotspot. However, most existing remote sensing image instance segmentation methods adopt the fully supervised paradigm and require expensive pixel-level labels. Moreover, remote sensing images often have issues such as mixed foreground and background and complex target contours, making segmentation challenging.Method To overcome these challenges, we propose a prior information-driven system suitable for weakly supervised instance segmentation tasks in remote sensing images and a multi-prior driven weakly supervised instance segmentation network (MPD-WSIS-Net) to address these challenges. The prior information of weakly supervised instance segmentation can be divided into task prior and image prior. The task prior mainly comes from the bounding box detection task, which is closely related to instance segmentation. Specifically, the paper obtains prior information from three components: box-mask projection consistency constraint, pixel discrimination difficulty representation function, and center position prior constraint. The image prior comes from summarizing or excavating information about the image itself. In this study, we focus on the relationships between neighboring pixels in the image and the gradient information of targets. By integrating these constraints and the pixel discrimination difficulty representation function, we establish a complete prior information driving system to effectively enable MPD-WSIS-Net to perform instance segmentation tasks under weakly supervised conditions in remote sensing images.ResultMPD-WSIS-Net was compared with weakly supervised methods, hybrid supervised methods, and fully supervised methods on optical and SAR remote sensing image datasets. Compared to weakly supervised methods, MPD-WSIS-Net achieved better segmentation results. Compared to hybrid supervised methods, the segmentation performance of MPD-WSIS-Net on both optical and SAR remote sensing image datasets significantly surpasses that of Mask R-CNN and CondInst of 50% pixel-level annotations. It is also competitive compared to Mask R-CNN and CondInst under 75% pixel-level annotation conditions. Compared to fully supervised methods trained with pixel-level labels, MPD-WSIS-Net can achieve 89.3% of fully supervised Mask R-CNN's AP value on optical and 84.3% on SAR remote sensing image datasets. Furthermore, we have demonstrated the positive impact of each prior information component in MPD-WSIS-Net on instance segmentation performance in optical remote sensing images through ablation experiments.Conclusion This study constructs a prior information-driven system consisting of task and image priors through a detailed analysis of the prior information in weakly supervised instance segmentation tasks. The specification of prior information is achieved through the box-mask projection consistency constraint, pixel discrimination difficulty representation function, center position prior constraint, neighborhood visual consistency constraint, and gradient consistency constraint. This research can enable MPD-WSIS-Net to complete instance segmentation tasks without pixel-level annotations and provide high-performance and low-cost prescription for fine-grained interpretation of optical and SAR remote sensing images.  
    关键词:remote sensing image;instance segmentation;fine-grained interpretation;weakly supervised learning;priori information;driven system;target contour;annotation cost   
    56
    |
    31
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50115652 false
    发布时间:2024-03-12
  • SHI Xin,XING Mengdao,ZHANG Jinsong,LIU Huitao,WANG Hongxian

    Corrected Proof
    DOI:10.11834/jrs.20243258
    摘要:Large-scale regional observations by remote sensing satellites play an important role in mapping, disaster relief and other fields. The efficiency of single satellite observation is low, and multi-satellite collaborative observation is the main means to solve the problem of rapid observation in large areas. At present, multi-satellite collaborative observation is mostly studied on optical remote sensing satellites, but there is less research on collaborative observation of SAR satellites. Moreover, SAR satellites have different imaging mechanisms and imaging modes from optical satellites, so the optical satellite collaborative planning method cannot be fully applied to SAR satellites. In order to give full play to the performance of SAR multi-satellite collaborative observation, it is urgent to study SAR multi-satellite collaborative regional observation technology.First of all, this study analyzes the coverage calculation of large-scale complex areas, and proposes a complex area coverage calculation method that combines Gaussian projection, grid division and geometric operations, which can realize coverage calculation of any complex area. Then, an accurate coverage analysis was performed on the SAR strip imaging mode, and a candidate area decomposition method combining angle restriction and two-dimensional decomposition was proposed. The optimization space is reduced through angle restriction to improve optimization efficiency, and complex continuous optimization problems are discretized through two-dimensional decomposition, making it possible to apply genetic algorithms for optimization. Finally, an improved genetic algorithm combining greedy algorithm initialization, elite retention strategy and cubic fitness function is proposed for regional coverage optimization. Chromosome encoding, crossover, and mutation operations are designed for optimization. The elite retention strategy is used to improve the optimization speed and stability, and the cubic fitness function is used to improve the optimization effect.This study selected four on-orbit SAR satellites: Gaofen-301, Gaofen-302, Gaofen-303, and Haisi-1, and three regional targets: Beijing, Tianjin, and Shanghai for simulation experiments. The simulation time is 5 days, the orbit data uses real TLE data, the SGP4 orbit propagation model is used, and the beam parameters of the SAR satellite are reasonably simulated. Experimental results show that the coverage rate of the proposed method on three regional targets in Beijing, Tianjin, and Shanghai is increased by 3.17%, 2.94%, and 9.02% respectively compared with the greedy algorithm. In the case of finer grids, the coverage result of the proposed method in the Shanghai area is 7.3% higher than that of the greedy algorithm.This study analyzes the SAR multi-satellite collaborative complex area coverage planning technology, constructs a feasible SAR multi-satellite collaborative complex area observation planning process, and proposes a complex area coverage planning method suitable for the SAR multi-satellite strip imaging mode. This algorithm can provide a technical basis for the establishment of a SAR multi-satellite collaborative regional planning system. However, the proposed method simplifies the constraints at the imaging signal processing level of SAR satellites, and subsequent research will conduct in-depth research based on the characteristics of SAR satellite imaging signal processing.  
    关键词:spaceborne SAR;multi-satellite collaboration;mission planning;regional observation;coverage calculation;regional decomposition;greedy algorithm;genetic algorithm   
    43
    |
    13
    |
    0
    citations on Dimensions.
    citations on Dimensions.
    <HTML>
    <L-PDF><Enhanced-PDF><Meta-XML>
    <引用本文> <批量引用> 50096825 false
    发布时间:2024-03-11
0