最新刊期
卷 24 ， 期 8 ， 2020
-
摘要：An ocean color satellite of China (Haiyang-1C, HY-1C) was launched on September 7, 2018, equipped with several optical sensors, such as the Chinese Ocean Color and Temperature Scanner (COCTS), Coastal Zone Imager (CZI), and ultraviolet imager. These instruments were tested in orbit for 6 months and used in ocean and coastal zone environmental monitoring in June 2019. The optical remote sensing of oil spills is a key research direction in marine environmental monitoring. Significant progress has been made in recent years, demonstrating its ability to detect, classify, and estimate the volumes of various oil spills. In this paper, the marine oil spill incident near Dongsha Island in the South China Sea on February 20, 2019 was used as a case study. COCTS and CZI captured the oil spill, and the oil spill area was scanned using the Visible Infrared Imager Radiometer Suite (VIIRS) and Moderate Resolution Imaging Spectroradiometer (MODIS Terra & Aqua) on the same day. The angles (between the viewing direction and the direction of mirror reflection, θm) of COCTS, CZI, VIIRS, and MODIS Terra and Aqua images were extremely small, indicating that sunglint reflectance in these images can be ignored. Therefore, the light absorption and backscattering characteristics of oil spills and surrounding oil-free seawater account for the differences among the images. In other words, these light signals can only weakly detect marine oil spills compared with strong sunglint reflection. However, if backscattering can be distinguished, it can be useful in the identification of various types of weathered oil. COCTS, CZI, and VIIRS capture marine oil spills with high radiometric resolutions and signal-to-noise ratios. Moreover, oil spills form oil emulsions and non-emulsified oil slicks, which can be distinguished by CZI due to its high spatial resolution (~50 m). Uncertainty analysis of COCTS and CZI shows that the difference between an oil spill and the oil-free seawater of COCTS or CZI is obvious, implying that CZI can be used in estimating the volumes of oil spills. In the near future, the same sensors in the HY-1D satellite will be available, and a network observation system of the HY-1C/D satellite will provide global images daily. This resource will likely play an important role in future ocean color remote sensing.关键词：remote sensing;HY-1C;COCTS;CZI;oil spills;optical remote sensing- 1327
- |
- 261
- |
- 21
发布时间：2020-08-28
Research Progress
-
摘要：Sun-Induced chlorophyll Fluorescence (SIF) has been recently used as a novel indicator of photosynthesis of vegetationdue toits direct relation to vegetation photosynthesis. The mechanism behind the SIF signal is rather complicated. Thus, the physical process and the interaction between SIF and vegetation structureshould be understoodfor better interpretation of SIF data. In this respect, the development of relevant SIF models is the key to improve the understanding and use of SIF signals. In this review, we introduce various SIF models in different scales by clarifying the mechanism behind each model andpropose the prospects in the future work.(1) At the leaf level, fluorescence models are generally based on leaf optical properties models and focus on the simulation of leaf reflectance and transmittance. Several theories canexplain the propagation of light in a turbid medium or a simplified blade. The simplest one assumes that light decays exponentially within the blade according to Beer’s law.The Kubelka–Munk differential equations arealsoused to solve radiation propagating in a turbid medium, which isfollowed by the Plate model fordiving the blade into several homogenous layers. The most popular leaf fluorescence model called Fluspect is based on the PROSPECT model, which follows Plate’s theory. The key objective is to simulate the re-absorption effect accurately due to the band overlap between SIF emission and chlorophyll absorption.(2) At the canopy level, fluorescence models incorporate the canopy radiative transfer and leaf fluorescence models, which can be characterized as 1D and 3D models. 1D models, such as FLSAIL, FluorSAIL,andSCOPE, incorporate SAIL model with a leaf fluorescence model and assume the canopy as several horizontally homogenous layers. 3D models, such as DART, FluorWPS, andFluorFLIGHT, simulate the canopy fluorescence in a realistic scene using ray tracing method. They are suitable to heterogeneous vegetation canopies.(3) At the ecosystem level, these fluorescence models help reduce the uncertainties in simulating carbon cycle and predicting ecological system response to global change by incorporating them with land surface models, such as NCAR CLM4, BEPS, and BETHY.Despite the advances in the SIF models across multiple scales, further studies are still needed with respect to model development, validation, and inversion. For example, accurate tree positions and canopy structure parameters can be derived from light detection and ranging data, which enables thereconstruction of 3D scenes based on the real landscape. With the development of insitu SIF measurement techniques, SIF models can be validated with the measurements except cross-validation through various models. SIF model inversion is a prospective research area to derive vegetation structure and biochemical parameters through spectral data. The novel machine learning approaches may also provide new opportunities to be incorporated with SIF models to solve inversion problems.关键词：sun-induced chlorophyll fluorescence;SIF model;radiative transfer model;vegetation structure;multiple scattering- 996
- |
- 481
- |
- 5
发布时间：2020-08-28 -
摘要：Steep terrain produces serious topographic effects on remote sensing satellite imageries. Serious topographic effects cause difficulty in classifying vegetation species and retrieving key essential climate variables (such as albedo, Leaf Area Index, and fraction of absorbed photo synthetically active radiation). These effects also bring complexity in distinguishing the unhealthy change in land covers over rugged terrains. Topographic correction is necessary for remote sensing applications over mountainous areas. Researchers have attempted to remove or at least reduce topographic effects in remote sensing imageries by using various standard methodological algorithms. The background and the topographic correction model have been reviewed in former studies. However, several effective topographic correction models, which have high quality and have been newly developed during these years, have not been mentioned and recommended comprehensively. Therefore, topographic correction models and evaluation methods for optical remote sensing imageries from the presented research chain were reviewed comprehensively in this paper. The aim was to determine the potential effective solutions for topographic effect correction over rugged terrains. This study is important for quantitative remote sensing applications in mountainous areas.Topographic correction models have been explored for more than 30 years (Fig. 1). These models can be divided into three categories, namely, the regression model, the Lambertian-based model, and the non-Lambertian-based model. The regression model generally has obvious advantages of simplified formulation and easy operation. However, this model has the shortage of lack of physical meaning for empirical parameters. The Lambertian-based model has rigorous mathematical formulas, which have clear physical meaning for parameters and are easy to operate for reducing complex topographic effects. Meanwhile, the Lambertian models are built on the basis of several assumptions and have the obvious shortage of ignoring the non-Lambertian surface reflectance. This negligence may result in overcorrection, especially over shady surfaces. The non-Lambertian model can improve the performance of the Lambertian-based model, especially over heterogeneous land surfaces.On the basis of this study, we summarized the problems in existing topographic correction model and provided several possible suggestions for the development of a topographic correction model in optical remote sensing. These suggestions can provide important guidance to the research on topographic correction and its practical application.关键词：remote sensing;image procession;Lambertian based model;Non-Lambertian model;land surface reflectance;topographic correction assessment- 1025
- |
- 442
- |
- 6
发布时间：2020-08-28
Review
-
摘要：Evapotranspiration (ET) is an important component in the soil-vegetation-atmosphere continuum. Remotely Sensed ET (RS-ET) provides multi-scale and spatiotemporally continuous information over the land surface, and has become an effective approach to obtain ET.Due to the heterogeneity of land surface and complexity of meteorological conditions at the near-surface layer, there exist various uncertainties derived from the model mechanism, parameterization scheme, input data and time scale conversion, which hindered correct estimations of ET, and further effect on its application. Therefore, it is essential to validate RS-ET to optimize the models and improve the associated products. This paper evaluated a group of validation methods for RS-ET (including evaporation and transpiration), which usually consists of direct validation and indirect validation. An overview of the principles, applicability, advantages, and disadvantages for all the validation methods were summarized. Direct validation is based on in situ measurements (including (micro-) lysimeter, stem sap flow, bowen ratio energy balance system, eddy covariance, and scintillator) to get the ground truth value, which can be used as the primary and reliable method to validate RS-ET and usually employed at the pixel and regional (or basin) scales. In the absence of ground truth ET, indirect validation becomes feasible, which can be classified into (1) cross-validation, (2) Multi-scale validation based on high spatial resolution remote sensing data, and (3) spatiotemporal variation analysis that combines multiple ET impact factors. Nevertheless, there are still a series of theoretical and methodological challenges in the validation of RS-ET, such as the scale mismatch between in situ measurement and remote sensing pixels due to the land surface heterogeneity. It is well-acknowledged that how to get the ground truth value at pixel and regional scales is the core issue of validation.This study demonstrated that validations of RS-ET products can be not only applied over homogeneous land surface but also heterogeneous surface with further development, which may at least but not limited to quantification of the spatial heterogeneity of land surface hydrothermal conditions, optimization of the experimental sites for validation over heterogeneous land surface, multi-scale measurements of ET on heterogeneous surface, acquisition of ground truth ET at pixel and regional scales, validation demonstration and uncertainty analyses of the validation process. Moreover, this study also proposed a generalized validation framework to validate RS_ET products at different scales (pixel scale and regional scale), which included direct validation (as the priority method) and indirect validation methods (as the auxiliary method), multiple validation data (i.e., ground truth ET at the pixel and regional scale, ET reanalysis data, various ET products, estimated ET from models and ET impact factors). The current framework focused on evaluating the accuracy and the spatiotemporal variations, identifying the error sources of the RS-ET products and analyzing the uncertainties during the validation process. This work is expected to improve the land surface remote sensing products and promote the development of quantitative remote sensing science.关键词：remotely sensed evapotranspiration;heterogeneous surface;validation;direct validation;indirect validation- 875
- |
- 522
- |
- 21
发布时间：2020-08-28 -
摘要：The classification of hyperspectral image remains a challenging task because of the complexity of spectral and spatial structures, high dimensionality, and strong correlation between adjacent bands. The combination of spatial and spectral information can provide significant advantage in terms of reducing the uncertainty of the samples because the same object has different spectrums and objects with the same spectrum in a hyperspectral image. The Local Binary Pattern (LBP) has also been introduced for spatial-domain feature extraction and classification of hyperspectral images as a simple but powerful texture descriptor. More recently, deep learning has been proven to be a preferable way to extract nonlinear high-level features because of its hierarchical learning framework. The combination of LBP features and the CNNs can lessen the workload of CNNs because of the discrimination capacity of LBP features. In this paper, a novel classification method combining DC–CNN and LBP features, called LBP Dual-Channel CNN (LBP–DC–CNN), is proposed.In LBP–DC–CNN, original hyperspectral data and LBP features are processed in a DC–CNN framework. On the one hand, original hyperspectral data is fed into a 1D–CNN model to extract original spectral features. On the other hand, LBP features are fed into an identical1D–CNN model to extract hierarchical spatial features further. Next, the fully connected layers of the two 1D–CNN models in the DC–CNN framework is concatenated into a fused layer, thus completing the fusion of spectral features and spatial features. Finally, the fused layer is fed into a softmax layer to conduct classification.(1) The OAs of LBP–DC–CNN are better than those of LBP–CNN and DC–CNN, which validate the feature extraction capacity of the CNNs and the advantage of LBP features. LBP–DC–CNN provides better accuracy than that of DC–CNN, which is an advantage of LBP features compared with the spatial features extracted by 2D–CNN model. In addition, the accuracy of LBP–DC–CNN is better than that of LBP–CNN, which validates the reasonability and discriminative power of the dual-channel CNN framework.(2) The OA of LBP–DC–CNN is apparently superior to those of compared methods, which makes DC–CNN and LBP features advantageous. For the Indian Pines data, LBP–DC–CNN (i.e., 98.54 %) yields approximately 2% higher accuracy than the DC–CNN (i.e., 96.68%)and approximately 4% higher accuracy than the LBP–CNN (i.e., 94.74 %). For the University of Pavia data, LBP–DC–CNN (i.e., 99.73 %) yields approximately 1 % higher accuracy than the DC–CNN (i.e., 98.74 %) and approximately 4 % higher accuracy than the LBP-CNN (i.e., 95.92 %). For the Salinas data, LBP–DC–CNN (i.e., 99.56 %) yields approximately 2 % higher accuracy than the DC–CNN (i.e., 97.33 %) and approximately 5 % higher accuracy than the LBP–CNN (i.e., 94.52 %).(3) LBP–DC–CNN can improve the class-specific accuracy of some ground materials, such as Corn-notill and Soybean-mintill in the Indian Pines data, Asphalt and Bricks in the University of Pavia data, and Grapes_untrained and Vinyard_untrained in the Salinas data. LBP features aremore discriminative than spatial features extracted by 2D–CNN.ResultExperiments were conducted on the Indian Pines dataset, Pavia University dataset, and Salinas dataset to verify the performance of LBP–DC–CNN compared with conventional methods. The results are as follows:关键词：remote sensing;hyperspectral image;classification;CNN;deep learning;LBP- 629
- |
- 191
- |
- 8
发布时间：2020-08-28
Doctor's Voice
-
摘要：ASTER GDEM V2 is one of the most widely used DEM data for Antarctic Ice Sheet analysis. However, it cannot be directly used due to its large anomalies in the snow-covered plateaus of the continent, where the surface reflectance is very high and no obvious surface features are available for the generation of high-quality DEM.This study rectifies the large anomalies in ASTER GDEM V2 over Byrd Glacier using a contour correction method. Then the accuracy of the anomaly-corrected ASTER GDEM V2 is quantitatively evaluated against.Results show that the RMSE of ASTER GDEM V2 decreases from 26.56 m to 18.77 m （-6.79 m）, which is lower than the RMSE of ICESat-1 DEM （121.24 m）. In addition, the accuracy of the corrected ASTER GDEM V2 is barely influenced by slope and no significant systematic errors can be observed. In contrast, the accuracy of ICESat-1 DEM is very sensitive to slope. Through further topographic profile analysis, the noises on previously non-corrected ASTER GDEM V2 are found to be mainly distributed in the low flat areas. Those noises can be effectively removed by using the contour correction method.This study indicates that the corrected ASTER GDEM V2 is well qualified for analyzing Byrd Glacier, Antarctica.关键词：remote sensing;Digital Elevation Model （DEM）;ASTER GDEM V2;ICESat/GLAS;Antarctic;Glacier- 675
- |
- 178
- |
- 3
发布时间：2020-08-28 -
摘要：Mesoscale Convective System (MCS) is the main reason of formation strong convective weather. Therefore, it is quite necessary for us to study the evolution and characteristic of MCS. Adjacent convective cells cannot be well distinguished is the main problem in convective cell detection. To address this issue, a new method is proposed based on H-maxima transform using infrared and water vapor channel data from FY-2F meteorological satellite.Firstly, Pixels with bright temperature greater than 241K (Baseline threshold of convection system) were removed and normalized to [0, 1]. Secondly, H-maxima transform technology was used to extract seed points of convective cells, and a new criterion for connected domain was designed to cluster adjacent seed points and mark them in order. Finally, a new merging method was developed to make the seed points grow or merge with adjacent seed points.Experimental results on satellite images from infrared channel and water vapor channel show that the proposed HTC method efficient and accurate, including initial, mature and dissipation stages of Mesoscale Convective Systems lifecycle. This study showed that the proposed method has a considerable application prospect for detecting convective cells in the field of meteorology. In addition, the proposed method is not only suitable for detecting single convective cells, but also capable of multiple convective cells detection.关键词：remote sensing;convective cell;convective core;H-maxima transform;clustering;FY-2;seed point- 563
- |
- 132
- |
- 1
发布时间：2020-08-28
Technology and Methodology
-
摘要：Microwave radiometers have been widely applied in polar region research because of their all-weather and all-time capabilities. Microwave Radiation Imager (MWRI) on FY-3B is the microwave radiometer of China’s own research and development and has aroused widespread concern. Long time series of earth observation data records play an important role in the research of earth environment changes and trends. The Arctic region is used as the study area and the data of Advanced Microwave Scanning Radiometer-2 (AMSR-2) on Global Change Observation Mission 1st-Water (GCOM-W1) are considered the standard data in providing the intercalibration result and the basis of retrieving remote sensing parameters in Arctic region in the future. Ascending and descending brightness temperatures at 10 channels in 2015 from FY-3B/MWRI are calibrated against those from GCOM-W1/AMSR-2.Before brightness temperature data analysis and intercalibration, the data are processed in five steps. The first step is reading data, in which the DN value of remote sensing is transferred to brightness temperature value in the research region. The second step is data quality control. If the standard deviation of values in a grid and eight surrounding grids is more than 3 K, then the values in the nine grids should be eliminated. The value that is more than 300 K or less than 10 K should also be eliminated. In the third step, stereographic projection is used to project the brightness temperature value, time, longitude, and latitude into 896×608 grids. In the fourth step, data at the land-sea boundary and Marginal Ice Zone (MIZ) should be eliminated because of the mixed pixel. First, the grid data of 7×7 around the land data are marked as land, and the data marked as land are removed. Then, the ratio of V187 to V365 from AMSR-2 is used to calculate the MIZ. The ratio, which is equal to 0.92, is viewed as the threshold to divide the sea ice and the open water. Thereafter, the 3×3 grid is set as a template. If the template includes the grids that represent sea ice and open water, then the nine grids are eliminated. The fifth and last step is to set the time window as 30 min and the space window as 12.5 km ×12.5 km and convert 2D matched data to 1D data for intercalibration.The intercalibration results of MWRI and AMSR-2 are as follows. First, the brightness temperature data of each channel of MWRI are smaller than those of AMSR-2, and the absolute values of monthly bias of vertical polarization channels are greater than those of horizontal polarization channels at the same frequency. The difference in monthly bias between ascending and descending orbits is small in each channel, which is less than 1 K. Second, the difference in the monthly bias in each channel between the ascending and descending orbits in the sea ice area is less than 1 K, while that in the open water is between 0 and 1.5 K. Third, linear regression analysis shows that most of the correlation coefficients of MWRI and AMSR-2 in each channel are above 0.99, which indicates a good correlation. The slope and intercept of the intercalibration of each channel in the ascending and descending orbits are obtained. Fourth, the brightness temperature of MWRI after calibration is consistent with that of AMSR-2. This consistency indicates that the intercalibration is effective.关键词：remote sensing;MWRI;AMSR-2;Arctic area;brightness temperature;intercalibration;FY-3- 693
- |
- 350
- |
- 2
发布时间：2020-08-28 -
摘要：Understanding of the spatiotemporal changes in permafrost Active Layer Thickness (ALT) in the Pan-Arctic region is important for global carbon flux simulation, climate change prediction, and freeze–thaw risk assessment. Many studies have been conducted on this subject. However, most previous works have relied on limited sites or regional simulation with a spatial resolution of 25 km or coarser. The spatiotemporal characteristics of ALT change at a landscape level needs to be explored, especially in the crucial infrastructure concentrated region. This study simulated the permafrost ALT in the Pan-Arctic region from 2001 to 2017 at kilometer level using Stefan method and permafrost site records, MOD11B3, and MCD12C1 data. Results showed that approximately 78.4% of the study area had an increase trend in ALT with a rate of 0.22 cm/a (p<0.05). Furthermore, the change in ALT spatially varied. The significantly increased areas were mainly distributed in Rocky Mountains, Laurentian Plateau of Canada, Central Siberian Plateau, and Central Siberian Plateau of Russia, with an increase rate between 0.5 and 1 cm/a. Meanwhile, the decreased areas were mainly concentrated in Hudson Bay Coastal Plain, Labrador Plateau of Canada, north East Siberian Mountains, north Central Siberian Plateau, east of Lake Baikal, and Taylor Peninsula of Russia. During this period, the ALT in 80% of the oil and gas areas had increased, with an increase rate between 0.1 and 0.7 cm/a. The variation in ALT was consistent with the temperature change. The ALT also varied with vegetation types in the order of ALT in forests > ALT in grasslands > ALT in savannahs > ALT in shrublands. However, the relationship between ALT and the thickness of snow cover was highly complicated. The results will deepen our understanding of the permafrost freeze–thaw pattern in the northern high latitudes and provide insights into the identification and prevention of freeze–thaw risk in the Pan-Arctic permafrost region.关键词：Pan-Arctic region;climate change;permafrost;active layer thickness;remote sensing- 648
- |
- 169
- |
- 4
发布时间：2020-08-28
Remote Sensing Applications
- 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