最新刊期
卷 27 ， 期 2 ， 2023
-
摘要：Unmanned Aerial Vehicles (UAVs) could flexibly and efficiently observe land surface and obtain surface information at very high frequency and resolutions. In recent years, UAV technology has been greatly improved and the UAVs are increasingly applied in large amount of civilian fields. The UAV Remote Sensing (UAVRS) is a comprehensive technology composed of UAV flying aircraft, light remote sensing payload, satellite positioning device, remote-control system, communication technology and application technology. An UAVRS system is mainly composed of UAV platform, load system, ground control and data transmission system and image processing system. With the popularity of civilian UAV products and the development of UAV remote sensing payload, UAVRS also shows an explosive growth trend in various industries. However, the short battery life and limited coverage of the civilian drones are one of the key factors restricting the UAVRS applications. Besides, although the application of UAVs in various industries and fields continues to deepen, most of them are single-type applications in a certain discipline or industry. The 5G technology and wireless communication technology advance and will give rise to the UAVRS applications in terms of “networking”, “IoT”, and “real-time”, as well as cross-domain and convergence application. The modern cities are highly complex and sociotechnical. They comprise people and communities interacting with one another and with objects (e.g., roads, buildings) within a range of urban settings or contexts. It is extremely difficult to monitor and manage such complex sociotechnical systems. The monitoring and mapping of pollutions, traffic, infrastructures are great challenges in rapidly changing cities, and especially gained increasing attentions of citizens and are putting great stresses on policy makers and urban planners. Theoretical and practical efforts to create better city monitoring and management systems have a long history. In the 21st century, we recognize and conceive “creative”, “smart” and “knowledge” cities, in which the multidisciplinary Information and Information and Communication Technology (ICT) has played a vital role. The Internet of Things (IoT) promotes the development of the smart cities. However, in various urban domains, the cities equipped with a sensor-based web, or a cyber-physical information infrastructure, are far from sufficient to help policy makers or citizens to get a full understanding of the real time or near real time city conditions. To this end, due to their mobility, autonomous operation, and communication/processing capabilities, UAVs are envisaged in many smart city application domains. In this article, we describe a technology of air-ground collaborative low-altitude UAV remote sensing network system. The system composes of intelligent UAV, the UAV base station (drone-in box) and operation system. The UAV remote sensing network can provide high-frequency and real-time observations of the land surface. Based on the UAV remote sensing network, we constructed a comprehensive application for urban governance. An UAVRS network composed of eight UAV base stations was deployed in Danzao Town, Foshan City. Connected through 5G network, the UAVRS network advances the UAV inspection frequency, spatial coverage, and response time. It provides revolutionary and intelligent services for local agents and departments of water affairs, environment protection, public security, urban management, and emergency. The UAVRS network system in Danzao is expected to develop a real-world smart city paradigm that could be copied and migrated to other towns across the country.关键词：UAV;UAV base station;UAV remote sensing network;air-ground collaborate- 2290
- |
- 774
- |
- 0
发布时间：2023-06-30 -
摘要：Earth observation systems are one of the cornerstones of Earth system science, and some milestone observational experiments have contributed greatly to the maturation of Earth system science and its research methodology. Among these observational experiments, remote sensing experiments have always played a key role. The Heihe remote sensing experiment is a large-scale and multidisciplinary satellite-airborne-ground integrated remote sensing experiment conducted from 2007 to 2017 in the Heihe River Basin, a typical inland river basin in China. The main scientific objectives are to observe the ecohydrological processes in the mountainous cryosphere, artificial oasis, and natural oasis in the Heihe River Basin. It was implemented in two stages: the Watershed Allied Telemetry Experimental Research (WATER) and the Heihe Watershed Allied Telemetry Experimental Research (HiWATER). More than 670 researchers participated in the Heihe remote sensing experiment, and more than 650 experimental datasets have been shared open and free. Characterized by capturing heterogeneities of complex land surfaces of the entire river basin, the Heihe remote sensing experiment has made breakthroughs in developing innovative multiscale observation methods, improving quantitative remote sensing models, and enhancing the applicability of remote sensing in ecohydrological studies. Overall, these progresses have led to a deeper harmonization of quantitative remote sensing and integrated ecohydrological research.This paper reviews Heihe remote sensing experiments and prospects for the future development of experimental remote sensing. Aiming to address the scientific challenges of developing scaling methods, measuring heterogeneity, and quantifying uncertainties, we have made the following advances in Heihe remote sensing experiments. (1) Innovative observation methods such as integrated satellite-airborne-ground observation, nested multiscale point-footprint-watershed-basin observation, wireless sensor network observation, and flux matrix observation methods have been invented or refined into maturation. (2) A variety of multisource remote sensing cooperative inversion methods, e.g. different spatial resolutions, polar and geostationary orbits, and active and passive sensors, have been developed. In particular, radiative transfer models for heterogeneous land surfaces have been established and validated. (3) Systematic advances in remote sensing data product validation technology, including optimal sampling, and upscaling of in situ observations to pixel-scale truth, have been achieved and verified. (4) More than 10 types of high-resolution ecohydrological remote sensing data products over heterogeneous land surfaces, such as precipitation, snow cover, evapotranspiration, soil moisture, and net primary productivity, have been produced at the river basin scale. Moreover, based on our integrated ecohydrological models, the hydrological cycles at different scales were closed, the oasis-desert interaction mechanism was revealed, and a diagnostic equation to close the energy balance of the eddy covariance system was proposed.Currently, the integrated observation system of the Heihe River basin is operating by taking the heritage of the Heihe remote sensing experiment. The Heihe River basin observation system will continue to support the development of new theories and methods of Earth observation technologies and serve the exploration and practice of watershed science and regional sustainable development.关键词：Heihe river basin;remote sensing experiment;watershed observing system;multi-scale observation;airborne remote sensing;quantitative remote sensing;scale transformation;remote sensing products;validation of remote sensing products;watershed science;ecohydrology- 1612
- |
- 677
- |
- 2
发布时间：2023-06-30 -
摘要：Compared with the current powerful acquisition capabilities of remote sensing data, its intelligent processing and knowledge service capabilities are relatively lagging. The contradiction between the accumulation of massive multisource remote sensing data and the limited information island is becoming increasingly prominent. Therefore, there is an urgent need for effective remote sensing domain knowledge modeling technology to assist in mining the useful information of remote sensing big data and form knowledge service capabilities. A Knowledge Graph (KG) describes the concepts and their relationships in the physical world in symbolic form. It has strong knowledge modeling and reasoning capabilities and has been successfully applied in search engines, e-commerce, social network analysis and other fields. Inspired by the general KGs, this paper conceives of establishing a remote sensing domain KG for the first time, which can provide support for knowledge modeling and knowledge services in the remote sensing field.First, this paper reviews the development history of general KGs. Second, it discusses the technologies of constructing remote sensing KGs. Compared with general KGs, remote sensing KGs are oriented to the field of remote sensing geosciences. They have significant disciplinary characteristics and spatiotemporal graph characteristics in terms of graph nodes, graph relationships and graph reasoning. Specific performances are as follows: (1) Images are an important part of remote sensing, which play an irreplaceable role and are ignored by general KGs. (2) Remote sensing knowledge is oriented to spatial entities. In addition to semantic relationships, the description of entity relationships also requires spatial and temporal relationships. (3) Traditional logical reasoning and natural language processing learning reasoning cannot effectively deal with image entities and spatial relationships. To solve the above problems, this paper draws on the construction scheme of the general KG and related domain KG and proposes the basic construction process of the remote sensing KG.Third, it introduces typical geoscience application cases driven by remote sensing KGs, which include three cases: (1) Marine oil spill monitoring. Marine oil spill KG is used for oil pollution identification, cause reasoning, and spill risk assessment, etc. (2) Land cover classification. Coupling remote sensing KG reasoning and deep learning for land cover classification. Numerous experiments have proven that KG can improve the classification results. (3) Evaluation of the carrying capacity of resources and the environment and suitability of land and space development. Ontology can not only express the knowledge system of evaluation in a standardized manner but also infer the evaluation results based on the constructed knowledge. Finally, it analyzes the application status and future research directions of remote sensing KGs. This paper points out four feasible and important research directions: (1) Exploring the theories and methods of creating multimodal remote sensing KGs; (2) Cooperative update and alignment fusion of remote sensing KGs; (3) Intelligent remote sensing image classification based on remote sensing KG representation learning; and (4) Scientific decision support analysis assisted by remote sensing KGs.Generally, the research of remote sensing KGs is conducive to better summarizing the conceptual knowledge of remote sensing, managing the new knowledge contained in remote sensing big data, and providing flexible and convenient remote sensing knowledge query and service capabilities to users in multiple fields, and it will help comprehensively improve the application capabilities of massive multisource remote sensing observation results and will play an important role in the study of global remote sensing land cover classification, climate change, international humanitarian assistance, and so on.关键词：remote sensing knowledge graph;knowledge graph application;artificial intelligence;knowledge service;representation learning;domain knowledge modeling- 4418
- |
- 1074
- |
- 1
发布时间：2023-06-30 -
摘要：Remote sensing image registration is the process of spatial alignment of two or more images through geometric transformation. It is an important preprocessing operation for image fusion, change detection, agricultural monitoring and other remote sensing applications. Considering that remote sensing images have the characteristics of large-scale changes, complex ground covers and imaging modalities, although a large number of registration methods have been developed, there is still a lack of methods that can be widely used in different scenarios. Therefore, research on registration algorithms with high efficiency, high robustness, high precision and wide applicability is of great significance. In recent years, deep learning, which has achieved great success in the field of natural image and medical image registration, has provided a new method for remote sensing image registration. First, we introduced two kinds of traditional registration methods and analyzed the advantages and disadvantages of area-based and feature-based registration methods in detail from the aspects of registration accuracy, efficiency and algorithm robustness. Generally, there are two main problems in traditional methods: poor applicability and insufficient utilization of the deep semantic information of the image. Second, we focused on the important progress of deep learning in area-based registration methods and feature-based registration methods. According to the specific application purpose of deep learning, we made a more detailed division of the above two methods and summarized the advantages and disadvantages of the existing research. In addition, considering the importance of datasets for deep learning, we sorted and shared some public datasets for remote sensing image registration. Due to the great progress of earth observation technology, an increasing number of remote sensing images are being applied. Image registration is the key step of remote sensing image preprocessing and the basic research content of quantitative remote sensing analysis. In recent years, research on remote sensing image registration algorithms based on deep learning has shown an increasing trend, but it is still in the early stage, and the framework is not mature. It mainly includes but is not limited to the following shortcomings: (1) lack of open source standard datasets; (2) difficult to apply to large-scale remote sensing images; (3) insufficient utilization of geospatial information and spectral information of remote sensing images; and (4) long training time and the large computing overhead. From the perspective of data and methods, we looked forward to the application of deep learning in the field of remote sensing image registration and put forward four main research directions: (1) remote sensing image registration datasets; (2) registration methods based on hybrid models; (3) registration methods based on different neural networks; and (4) training strategies based on small samples.关键词：deep learning;image registration;area-based;feature-based;registration datasets- 1832
- |
- 1067
- |
- 0
发布时间：2023-06-30 -
摘要：The Volume Scattering Function (VSF) is an important Inherent Optical Property (IOP) of seawater, which is a critical and fundamental parameter to describe the angular distribution of the scattering of incident light by water. In particular, the VSF has a significant importance in the field of ocean color remote sensing, atmosphere-sea interaction, eco-disaster alerting. Due to a strong directional distribution characterized by a large dynamic range in signal in the whole scatter directions and faint backscattering signal, the VSF measurement and calibration are complex and the research is still in the process of exploration and progress. In this paper, the research progress of calibration method and measurement technology of VSF has been summarized, and the development tendency of the VSF calibration method has been predicted.Since the birth of the first backscattering in situ measurement instrument in the 1930s, the calibration methods of the VSF measurement improved gradually with the development of measurement techniques. There are three main methods for calibrating the VSF measurement: the first is based on the estimation of the scattering volume and the scattering flux, the second is based on a diffusing target throughout the sample volume to obtain the sensor response weighting function, and the third is based on the Mie theory and the standard solution or particles to obtain the sensor response weighting function. For the first calibration method, the scattering volume is generally calculated based on the optical and mechanical structures with the plausible assumptions, and the scattering flux can be obtained by the response of each scattering sensor to a Lambertian target as a function of distance, this method is limited in the range of angles due to the large error in the estimation of scattering volume of the forward small angle and the backward near 180°. For the second one, the core of this method is the obtaining of the detector’s weighting function, and it is not modeled by the structure parameters of the instrument, but accurately measured through the moving a diffusing target throughout the sample volume, this method can effectively solve the problems of difficult and inaccurate scatterer estimation, and it is more suitable for large backscattering with wide Field Of View (FOV). In addition, for the last one, based on standard materials with known physical features and Mie scattering or Rayleigh scattering, the scattering properties of standard materials can be calculated theoretically. The detector’s weighting function or calibration coefficient is obtained by matching the measured original signals with the theoretical value. Using this method, the calibration process of the VSF measurement has been significantly simplified, and no longer constrained to the scattering angle range. It is clear that the precision of this method is positively related to the degree of perfection of the theoretical model and the parametric information of the standard substance.With the gradual improvement in the understanding of the IOPs of water, and the increasing demand for in situ measurements, the VSF measurement technology and the calibrating precision are also constantly innovated and optimized. It can be anticipated that all above three main types of the VSF calibration methods or a variant combination of them will continue for a long period of time, while the calibration method based on standard matter and sensor response weighting function will play an important role in the future.关键词：Volume Scattering Function (VSF);Calibration techniques;Mie scattering;Weighting function;Inherent Optical Property (IOP);ocean color remote sensing- 853
- |
- 275
- |
- 0
发布时间：2023-06-30
Review
-
摘要：Long-term, quantitative and dynamic monitoring of large-scale regional ecological environmental quality using remote sensing images can provide strong decision support for regional sustainable development. Based on the Remote Sensing based Ecological Index (RSEI), an enhanced Remote Sensing Ecological Index (ERSEI) is proposed from the perspective of the elements of the coupled ecosystem considering regional characteristics and application requirements of the ecological environment in arid areas. The ERSEI considers the factors of greenness (NDVI), wetness (Wet), dryness (NDBSI), and heat (LST) while introducing the Comprehensive Salinity Index (CSI) and estimation model of water network density (EMW). The salinity and Water Network Density (WND) are included in the ecological environment quality assessment. With the help of the Google Earth Engine (GEE) cloud computing platform, ERSEI is applied to the Hohhot-Baotou-Ordos-Yulin urban agglomeration in the arid area of northwestern China. The results show that the ERSEI can fully reflect the detailed characteristics of the surface in arid areas and effectively highligh the gradual information of the radiation impact of the water network on the surrounding environment. According to the spatial measurement and time series evolution analysis of ERSEI in the Hohhot-Baotou-Ordos-Yulin urban agglomeration from 2000 to 2020, it is found that areas with good ecological environment quality are mainly distributed in the Hetao Plain, Daqing Mountain and the side close to Luliang Mountain. Areas with poor ecological environment quality are mainly concentrated in the Mongolian Plateau, near the Hobq Desert and the Mu Us Sandy Land, and the quality of the ecological environment has shown a continuous decline. Therefore, these areas should be treated as ecological risk early warning zones to strengthen governance. The ERSEI provides a fast and effective new method for the normalized monitoring of ecological environment quality in arid areas.关键词：ERSEI;spatial measurement of ecological environment;time series evolution analysis;Google Earth Engine;Hohhot-Baotou-Ordos-Yulin urban agglomeration;ecological risk warning zone- 2721
- |
- 804
- |
- 6
发布时间：2023-06-30 -
摘要：The Antarctic ice sheet is an important indicator of climate change and a driver of sea level rise, with a volume in sea level equivalent terms of 58.3 m. Its tiny change could have a significant impact on the global sea mean level, which is considered to be one of the most serious consequences of future climate change. Therefore, understanding where and how the Antarctic ice mass changes is of societal importance. Considering the few assessments of the interannual change in the Antarctic mass balance, the objective of this paper is to demonstrate the estimation and uncertainty of the annual material balance of the Antarctic ice sheet from 2005 to 2016 based on the Input‒Output Method (IOM). The reasons for the change in the Antarctic ice sheet in various basins have been investigated, which will provide effective supporting data for further studies on the loss of the Antarctic ice sheet.Compared with other methods for estimating the Antarctic mass balance, the IOM quantifies the difference between mass gain through primarily snowfall and loss by sublimation, meltwater runoff and ice discharge. The advantage of this approach is that it separately calculates changes in each component at the scale of individual glacier drainage basins. Unquestionably, using different datasets could cause great variation. Here, we improve the method to finely evaluate the ice discharge over the grounding line, which can accurately calculate the flux at each outlet unit, ensure time series continuity, and define flux export widths. Finally, for the first time, year-to-year estimates of the ice flux from the islands around Antarctica have been achieved.Our results are within a reasonable range compared to the international estimates of the Antarctic ice sheet mass balance. During 2005—2016, Antarctica was basically in a loss state, with an average mass loss of 109.1 ± 34.9 Gt/a and a standard deviation of ± 84.1 Gt/a. West Antarctica dominated the mass loss and contributed 65.1% of the total loss, East Antarctica 26.4%, the Peninsula 4.5% and islands 4.0%. All of East Antarctica was in a positive mass balance and showed evident ice mass loss in some basins. The Peninsula fluctuated at zero equilibrium. The islands, accounting for 1.15% of the Antarctic ice sheet, were assessed individually for the first time and found to be in a persistent negative mass balance, with mass loss even exceeding the Peninsula in some years. On the whole, the change in the Antarctic ice sheet mass balance was dominated by the surface mass balance, which was mainly influenced by interannual variability in the climatological factors. From a small-scale perspective, the dynamic changes in ice flux at the grounding line due to ice shelves thinning and iceberg calving affected the mass balance in some regions, resulting in an increase in mass loss during the years of calving events.This study improves the IOM method for the detailed assessment of the Antarctic ice sheet mass balance during 2005—2016.关键词：Antarctic Ice Sheet;Mass balance;input-output method;surface mass balance;ice flux- 2875
- |
- 272
- |
- 0
发布时间：2023-06-30 -
摘要：Miyun Reservoir has produced huge benefits in flood control, agricultural irrigation, power generation, aquaculture, tourism, and urban water supply. Accurate water mapping is of great significance to the ecological environment monitoring of the Miyun Reservoir and the management of the South-to-North Water Diversion Project. The purpose of this research is to design a long-term dynamic water mapping method for the Miyun Reservoir by solving difficult problems such as cloud and snow interference, terrain shadows, and mixed pixels encountered in the mapping to realize the analysis and monitoring of water surface information changes in the Miyun Reservoir.The research successfully applied the tasselled cap transformation to the cloud detection of Landsat series data. DEM data are used to remove terrain shadows during processing. The water index WI, which is very suitable for long-term dynamic mapping, is introduced for water extraction, and the local unmixing method is used to make the extracted water contours and subpixel small targets more accurate. In addition, the research innovatively uses the landscape separation index to conduct a macroscopic analysis of the water surface morphology of the Miyun Reservoir.The algorithm in this paper has completed the dynamic water map of the Miyun Reservoir from 1984 to 2020, and the accuracy of the mapping result is high. The overall accuracy of direct verification is 98.2% under the condition of no clouds and snow, which is comparable to the existing water system map product (improved FROM-GLC). The overall consistency of cross-validation is as high as 99.4%. In addition, a long-term analysis of water surface information, such as the area, coverage, and morphological characteristics of the Miyun Reservoir, has been performed. (1) The area of Miyun Reservoir changed tremendously during the 37 years from 1984 to 2020, with the largest being 151.6 km² and the smallest being 57.3 km². The area changes are mainly concentrated in three areas, including the northern area, the area where the Chao and Bai Rivers enter the reservoir, and the island in the center of the reservoir. (2) Based on the changes in the area of Miyun Reservoir, it is divided into five periods: “growth” (1984—1993), “peak” (1994—1999), “decline” (2000—2003), “protection” (2004—2014), and “recovery” (2015—2020). (3) The changes in Miyun Reservoir during one year mainly occurred in 4 areas, including the northern area, the area where the Chao River enters the reservoir, the island in the center of the reservoir, and the West Stone Camel subdam area. Due to the release of water from the reservoir in May before the arrival of the rainy season each year, the area is the smallest. The area reached the largest in a year at the end of the rainy season in August or September because a large amount of rainwater was accumulated. (4) The Miyun Reservoir’s landscape separation changed greatly during the 37 years. When the water volume is small, it splits into two reservoir areas in the east and west. The east and west reservoirs underwent a process of three divisions and three consolidations from 1984 to 2020. The three split periods include 1984—1986, 2003—2005, and 2014—2015, and the other years are in the consolidation period.关键词：optical remote sensing;dynamic water mapping;time series analysis;Miyun Reservoir;water area;landscape index;Landsat- 1748
- |
- 458
- |
- 1
发布时间：2023-06-30 -
摘要：Soil saline-alkali stress is the key factor of low plant productivity and the bottleneck of sustainable development in global saline-alkali areas. Obtaining regional soil salinity information both efficiently and reliably is a necessary problem to be solved. The rapid development of global navigation satellite system reflectometry (GNSS-R) provides a new opportunity to use spaceborne GNSS-R to retrieve soil salinity. The Cyclone Global Navigation Satellite System (CYGNSS) is one of the important components of the spaceborne GNSS-R mission, and the L-band used by CYGNSS is very sensitive to the soil dielectric constant, which provides a theoretical basis for estimating soil salinity. In this paper, CYGNSS was taken as the main data source, and the Yellow River Delta region, a typical area with extreme salinization of soil, was selected as the research religion to discuss the feasibility of soil salinity estimation by CYGNSS for the first time. A set of soil salinity retrieval methods was established.We proposed a physical model that took CYGNSS as the main data with some other auxiliary data fused. First, the surface reflectance was obtained by calculating the CYGNSS data of coherent signals based on the bistatic radar equation, and then the surface roughness and vegetation attenuation effects of the surface reflectance were corrected to calculate the magnitude of the soil dielectric constant. Second, based on the improved Dobson-S soil dielectric constant model as the physical model and the Soil Moisture Active Passive Mission (SMAP) soil moisture product as the main auxiliary data, a set of soil salinity retrieval methods was constructed to complete the soil salinity estimation in the Yellow River Delta High-efficiency Ecological Economic Zone in May 2020. Finally, the result was verified by the ground-measured conductivity value.It was found that the soil salinity retrieved from the CYGNSS data correlated well with the ground-measured conductivity, with a coefficient of determination (R2) equal to 0.88 and a Root Mean Squared Error (RMSE) equal to 1.06 mS/cm. Therefore, a high-precision soil salinization map of the Yellow River Delta was made by kriging interpolation based on the estimation result, which showed an obvious trend that soil salinity decreased gradually from coastal to inland at the regional scale with a strong spatial heterogeneity itself.In this paper, a physical model for soil salinity estimation based on the bistatic radar equation and dielectric constant model was proposed using CYGNSS as the main data source. The results of this study indicated that it is feasible to use CYGNSS to estimate soil salinity and proved the sensitivity of the L-band to soil salinity, providing a new idea for soil salinity retrieval on a regional scale. In future studies, the method of multisource data fusion can be considered to transform the estimation results from point data to planar area for expression, and a new validation method of high precision is needed due to the strong spatial heterogeneity of soil salinity.关键词：remote sensing;soil salinity;CYGNSS;the bistatic radar equation;dielectric constant;Yellow River delta- 1315
- |
- 375
- |
- 0
发布时间：2023-06-30 -
摘要：Phytoplankton are a vital component of marine organisms and play a crucial role in material circulation, energy flow and information transmission. The study of phytoplankton communities is important to understand carbon cycling and biodiversity, and the study of the biomass for a certain phytoplankton group can help understand the structural changes of phytoplankton communities. Diatoms are the dominant taxonomic group with the most abundant concentration in coastal water, and diatoms provide solid photosynthetic capacity (approximately 40% of marine primary productivity). Meanwhile, diatom concentrations will indirectly affect the breeding progress of holoplankton. In this study, observation samples (n = 252) were collected in the marginal seas of the northwest Pacific Ocean, covering the Bohai Sea (BS), Yellow Sea (YS), and East China Sea (ECS), from six cruise surveys that covered different seasons. The concentration of diatoms was calculated by phytoplankton pigment concentrations collected with the High-Performance Liquid Chromatography (HPLC) method with the CHEMTAX method, and the spatial distribution of measured diatom concentrations was analyzed. Based on in situ data parameters collected by multiple cruises, including phytoplankton absorption coefficients and diatom concentrations, we optimized Gaussian central bands and half-wave widths with derivative spectrum analysis and the peaks of the specific absorption spectrum, which were mentioned in existing research. Then, the correlation coefficients between diatom concentrations and different Gaussian peaks, which were decomposed by Gaussian function, were analyzed. We established Gaussian peaks, which had higher correlation coefficients with diatom concentrations, to establish an inversion model of diatom concentrations. The inversion models between Gaussian peaks and diatom concentrations were studied with different mathematical functions (linear function, quadratic function, logarithmic function, exponential function and power function), and the accuracy for each model was evaluated by two precision indicators (determination coefficients, R2, and median percent error, ME) to propose the optimal inversion model form. The parameters (A, B) were confirmed by leave-one-out methods later. The results of the research show the following: (1) The spatial distribution of measured diatom concentrations presented characteristics that were higher nearshore and lower offshore. The higher areas for BS appeared in the neighborhood, and the trend to the central area was decreasing progressively. Similar to BS, the spatial distribution of measured diatom concentrations was higher nearshore (Yangtze Estuary) and lower offshore (central area) in both the YS and ECS. (2) The evaluation indicators for our diatom concentration inversion model were relatively high. Comparing the diatom concentrations modeled in this study and calculated by the CHEMTAX method and using leave-one-out methods to verify the accuracy between them, the determination coefficient (R2) was 0.79, the Median Error (ME) was 49%, and the p value was under 0.001. It should be noted that some deviation appears in the lower concentration (<0.02 mg m-3); that is why the range of application should be concluded in later applications. In summary, the relatively high accuracy shows the applicability of the inversion model of diatom concentrations in this study, and the model lays the foundation for conducting research on satellite remote sensing inversion of diatom concentrations.关键词：remote sensing;Diatoms concentration;Gaussian model;CHEMTAX;Chinese marginal seas- 860
- |
- 282
- |
- 0
发布时间：2023-06-30 -
摘要：Vegetation phenology refers to the specific timing of periodic events in plants and how these timings are adapted by periodic variations in climate and environmental factors such as air temperature and soil moisture content. Vegetation phenology change trends are closely related to global climate change; therefore, studies on vegetation phenology can help us better understand global climate change and how vegetation reacts to climate change. Remote sensing technology has been the main means for large-scale vegetation research; however, there have been problems when using remote sensing Vegetation Indexes (VIs) to monitor vegetation phenology due to the discrepancies between the vegetation greenness index and photosynthesis. Especially in evergreen forests, the periodic change in VI datasets is weak; thus, it is difficult to capture phenology metrics based on these VI datasets. Therefore, there is an urgent need to develop new technology to better monitor vegetation phenology. Recently, Sun-Induced Chlorophyll Fluorescence (SIF) has attracted increasing attention since it is strongly coupled with photosynthesis and has good performance in estimating vegetation Gross Primary Productivity (GPP). Due to its strong correlation with GPP, SIF is capable of capturing the rapid change in GPP over time and has great potential for monitoring vegetation phenology. Based on GOME-2 SIF, GOSIF, and CSIF data in the Northern Hemisphere during 2007—2018, this study mainly calculated the vegetation phenology metrics by using a double logistic model and analyzed the vegetation phenology change trends by using Sen’s slope trend test. The results showed the following: (1) The double logistic model used in this study could capture the start of the growing season (SOS) better than the end of the growing season (EOS), and vegetation phenology metrics derived from SIF data have stronger correlations with vegetation phenology metrics derived from GPP data than those derived from VI data, especially for SOS. (2) In the Northern Hemisphere, the multiannual average SOS was mainly (>90%) concentrated in 100—170 days, while the multiannual average EOS was mainly concentrated in 220—270 days. The SOS occurred later in high latitude areas and high-altitude areas, while the EOS showed the opposite trend. (3) From 2007 to 2018, the SOS derived from GOME-2 SIF data in the Northern Hemisphere showed a significant advancing trend (Senslope was -0.173), and the EOS showed an insignificant advancing trend (Senslope was -0.002). (4) The vegetation phenology in high latitude cold areas was mainly affected by air temperature, while the vegetation phenology in middle and low latitude arid areas was mainly affected by precipitation. SIF has great potential to calculate phenological characteristics based on vegetation photosynthesis, and vegetation phenology derived from GOME-2 SIF data showed a weaker change trend over the last 10 years compared with that of the period from 1980 to 2010. Overall, this study first analyzed the vegetation phenology change trends in the past 10 years based on long-term GOME-2 SIF datasets, and the results in this study could promote our understanding of global climate change and the terrestrial carbon cycle.关键词：remote sensing;solar-induced chlorophyll fluorescence;vegetation phenology;spatial pattern;trend;climatic factors- 1504
- |
- 462
- |
- 1
发布时间：2023-06-30
Ecology and Environment
-
摘要：The shapes of the channel Spectral Response Function (SRF) of Microwave Humidity and Temperature Sounders (MWHTS) are generally considered to be approximately rectangular. However, the SRF of each band channel of MWHTS shows certain in-band fluctuations based on the actual SRF test data. In this paper, the brightness temperature spectra of different scenarios for each channel at 118 GHz of MWHTS are simulated using an Atmospheric Radiative Transfer Simulator (ARTS). After inputting them into the MWHTS system simulation model we established in the previous stage, the output brightness temperature of the instrument is obtained after calibration, the influences of actual SRF on brightness temperature measurements and the retrievals of atmospheric temperature profiles are evaluated, and comparisons are further made with the actual satellite data of MWHTS in FY-3D to verify the results. The results show that the brightness temperature bias is linearly and positively correlated with the in-band fluctuations of SRF. The bias can reach 0.2—0.5 K in the 118 GHz channel when the actual SRF in-band fluctuations are larger than 3 dB. The in-band fluctuations of SRF will cause retrieval errors in the atmospheric temperature profiles, especially at an altitude of 1.8 km, where the retrieval error can reach the maximum value of 0.8—0.9 K. The simulation results are consistent with the results of the actual satellite data, so special attention needs to be paid to the bias in the simulated brightness temperature of the channels with large SRF fluctuations in applying simulations for Numerical Weather Prediction (NWP) using data assimilation methods, which has important research value for the future application of satellite data.关键词：remote sensing;FY-3D meteorological satellite;microwave humidity and temperature sounder;spectral response function;brightness temperature bias;retrieval of temperature profiles- 815
- |
- 356
- |
- 1
发布时间：2023-06-30 -
摘要：The Multiangle Implementation of Atmospheric Correction (MAIAC) is a new generic algorithm applied to MODIS measurements to retrieve Aerosol Optical Depth (AOD) over land at high spatial resolution (1 km). It is expected to have good potential in improving the AOD inversion of dark and bright surfaces of land. The high spatial resolution of the MAIAC retrievals enhances the capability to distinguish aerosol sources and determine subtle aerosol features. Retrieval of satellite aerosol properties is therefore often challenging due to considerable seasonal variations in surface reflectance and aerosol properties. To date, MAIAC AOD over arid regions under data-scarce environments has been evaluated. Considering these uncertainties, a systematic effort was made to evaluate the MAIAC AOD over arid areas using Aerosol Robotic Network (AERONET) ground-based AOD from 2000 to 2019. Considerable MAIAC-AERONET AOD matchups demonstrate the capability of MAIAC to retrieve AOD over varied ground surfaces and temporal scales. We employed a broader perspective and evaluated MAIAC performance under varying aerosol loading, aerosol types, surface coverage, and viewing geometry. The results show that (1) MAIAC performed well over various temporal scales, including monthly, seasonal and annual scales. Although underestimation is prevalent, MAIAC AOD in the spring and winter months correspond to the highest and lowest retrieval accuracies, respectively. (2) MAIAC performed well over four different typical land surface land surfaces, showing the highest retrieval accuracy over grassland, yet it slightly overestimated AOD. Construction land is most affected by the aerosol model, and farmland is strongly disturbed by surface reflectance and underestimated most obviously (Below EE = 46.5%). (3) The accuracies of the MAIAC AOD observations of Terra and Aqua are similar; R2 is more than 0.75, but both are underestimated, especially for Aqua. In general, MAIAC’s ability to provide AOD at high spatial resolution appears promising over arid areas and is expected to be helpful to study the characteristics of fine aerosols in arid areas and promote the study of local air quality.关键词：remote sensing;aerosols;AOD;AERONET;MAIAC;arid areas- 889
- |
- 324
- |
- 0
发布时间：2023-06-30 -
摘要：Earth energy comes from solar radiation, and solar radiation has a potential impact on the atmospheric boundary layer. During the solar eclipse, the solar disk is covered by the moon, and the solar radiation is reduced to reach the Earth, causing cooling in the surface layers of the atmosphere. Solar eclipse provides an ideal condition for studying the response of the atmosphere. On June 21, 2020, an annular eclipse occurred in China, and the partial solar eclipse can be seen except for the annular eclipse region. This solar eclipse provides a good opportunity for us to study the variation of the atmospheric thermodynamics. The most important parameter of thermodynamics during solar eclipse is the temperature in the different altitudes in the boundary layer. In this paper, to study the effect of the solar eclipse on atmospheric thermodynamics in the boundary layer, the atmospheric temperature and humidity profiles observed by the Ground-based multichannel Microwave Radiometer (GMR) in Xi’an, Zunyi, Nanning and Yibin, China. The GMR has high spatial resolution and high sensitivity and it is usually used to observe and study the atmospheric temperature and humidity profiles. Therefore, the variation of the atmospheric thermodynamics was studied at different regions and weather conditions by using the GMR during the solar eclipse. At these observation stations, the proportion of the sun covered by the moon is different; the annular eclipse can be observed in Yibin, and the covered proportion is maximum. This paper compared the results of the experiment and the observation results showed that the temperature and humidity profiles of the boundary layer had obviously changed. From the beginning of the solar eclipse to the maximum of the solar eclipse, the temperature of the boundary layer began to decrease because the solar radiation reaching the Earth was reduced. During the period from the maximum solar eclipse to the end of the solar eclipse, the radiation energy from the sun to the earth gradually increased, the temperature of the boundary layer began to rise, the variation in relative humidity was the opposite, the maximum variation in the temperature was approximately 4 ℃, the relative humidity was more than 10%, and the water vapor density profile had no obvious variation at these stations. The variation in the temperature and humidity lagged behind by approximately 15 to 20 minutes. The effect of the solar eclipse decreases with increasing height for the temperature and humidity in the boundary layer. In the process of the solar eclipse, the greater the solar disk is covered by the moon, the more obvious the effect of the solar eclipse on the temperature and humidity in the boundary layer. Due to the attenuation of solar radiation by clouds and rain, the influence of solar eclipses on the atmospheric temperature and humidity in the boundary layer will be weakened. This observation experiment had provided fine-scale variations of the atmospheric parameters both in time and height by using the GMR during the solar eclipse. These results may have important implications in understanding the response of atmosphere to the thermodynamics perturbations caused by the solar eclipse.关键词：remote sensing;the solar eclipse;Microwave radiometer;the boundary layer;temperature profile;relative humidity profile- 716
- |
- 261
- |
- 0
发布时间：2023-06-30 -
摘要：At present, the situation of environmental pollution in China is grim, among which regional compound air pollution dominated by PM2.5 is the most prominent. Aerosol Optical Depth (AOD) is a key physical quantity used to characterize the degree of atmospheric turbidity, which represents the intensity of aerosol light reduction. Many studies have shown that there is a strong correlation between AOD and PM2.5. Using the AOD data obtained by satellite remote sensing combined with other influencing factors to analyze the change mechanism of PM2.5 is of great significance to air pollution prevention and the protection of human health.The diffusion of PM2.5 is an extremely complicated process, and the PM2.5 prediction model based on the statistical regression method can only describe a relatively simple nonlinear relationship. However, the estimation of PM2.5 is considered to be a more complex multivariable nonlinear problem. Compared with statistical regression models, the PM2.5 prediction model based on traditional machine learning algorithms can deal with more complex nonlinear problems. However, its ability to process historical data is still limited, so it is difficult to mine the variation law of pollutant concentrations from the perspective of big data. Compared with the traditional machine learning method, the models based on deep learning can dig deep features hidden in historical data. However, the AOD remote sensing data are affected by image time resolution and pixel cloud pollution, which will greatly reduce the effective data. Because the construction of a deep learning method depends on a large amount of training data, less training data will seriously affect the model accuracy.Aiming at the problem that the traditional machine learning algorithm cannot deeply mine the hidden association features in data and the deep learning algorithm has a poor effect under the condition of less data, a combined model of PM2.5 prediction based on deep learning and random forest is proposed. The model builds a training dataset with AOD remote sensing data, meteorological reanalysis data and PM2.5 ground observation data. The deep hidden features in the training data are extracted by the powerful feature extraction ability of the deep learning model first. Then, the extracted hidden features are used in the training of the random forest model, and the predicted value of PM2.5 concentration is obtained by the random forest regression algorithm.To verify the effectiveness of this method, a series of experiments were carried out. The results demonstrate that PMCOM has better prediction accuracy in both overall prediction and seasonal prediction scenarios. The combination of random forest and long- and short-term memory neural networks is the best for this experiment. Even when only 35% of the data are used for training, R2 in the overall prediction experiment can reach 0.89, and R2 in each season prediction experiment is also above 0.75.The combination of deep learning and random forest can reduce the dependence of deep learning models on the amount of data by random forest and make full use of the high-level hidden features of existing historical data. In this way, it makes up for the deficiency of mining the internal associated features of data by a random forest model and improves the prediction accuracy of PM2.5 concentration.关键词：remote sensing;PM2.5;deep learning;Random Forest;LSTM;PMCOM- 3357
- |
- 553
- |
- 0
发布时间：2023-06-30
Atmosphere and Ocean
-
摘要：As an important parameter of vegetation canopy structure, the Leaf Area Index (LAI) has become a standard land surface parameter product for many earth observation systems and an important input parameter for several quantitative remote sensing models. Rapid and accurate acquisition of vegetation LAI is of great significance for the verification of remote sensing products and promotion of the development of remote sensing models. With the improvement of smartphone sensor performance and the functions of application software, smartphones have become a new alternative to vegetation LAI measurement instruments. However, due to the limitation of the narrow Field Of View (FOV) angle of the smartphone camera sensor, the existing algorithm relies on the assumption that the leaf inclination belongs to the spherical distribution, which is that the G function (the projection of a unit leaf area on a plane perpendicular to the observed zenith angle) is equal to 0.5. Therefore, the traditional algorithm cannot solve the problem of unknown leaf inclination distribution. In this paper, a G function estimation method based on shape matching was proposed. Based on the finite length method and the gap fraction of multiple images, the vegetation canopy clumping index in the quadrat was calculated, and the effective LAI (LAIeff) and the real LAI (LAItru) were obtained by using the Poisson distribution model. The algorithm was validated by data obtained from destructive measurements (LAIdes) of two crop types (maize and soybean) at Hailun Farm in Heilongjiang Province, China. The measured time covers the main growth stages of the crop. The results showed that the Root Mean Square Error (RMSE) of the estimated LAI using the algorithm before improvement was 0.84 (vertical shooting) and 1.33 (tilted 57° shooting), and the RMSE of LAIeff and LAItru after the improvement was 0.58 and 0.56, respectively. The LAI values retrieved by the new algorithm are more consistent with the growing trend of LAI in the time series. The algorithm in this paper extends the measurement method of crop LAI, which provides the possibility to quickly and accurately extract vegetation LAI from smartphone-captured images. Further research will be considered in two directions: analyzing the influence of external light environment changes on the measurement results and adding validation data of different vegetation types.关键词：remote sensing;smartphone;leaf area index;multi-angle gap fractions;G function;clumping index;effective leaf area index- 989
- |
- 271
- |
- 0
发布时间：2023-06-30 -
摘要：As ground features are affected by various factors, the problem of endmember variability will occur. Endmember variability greatly affects the accuracy of spectral unmixing results. This study is based on the vegetation-soil binary scene, and spectral unmixing is performed under the framework of NMF (Nonnegative Matrix Factorization). The PROSAIL model is used to describe the variability of vegetation endmembers from the mechanism so that the results of spectral unmixing have clear physical meaning. To improve efficiency, we set up two neural networks for model calculation and model inversion. In this way, the spectral unmixing algorithm can obtain the endmembers of the vegetation pixel by pixel, which can more accurately describe the variability of the vegetation endmember.In addition, there is diversity in the spatial resolution of remote sensing data products. The problem of the scale effect is widespread and is a key issue in the field of remote sensing. Analysis of the reason is largely due to the mixed pixels that universally exist. The method studied in this paper describes the variability in the vegetation endmember. A spectral unmixing algorithm that can describe the variability of vegetation endmembers pixel by pixel is obtained. This result can be used to invert vegetation parameters. Therefore, this study attempts to correct the scale effect of remote sensing products by considering the spectral unmixing method of vegetation endmember variability.This paper takes the LAI scale effect as an example. The effectiveness of the method is verified by an Unmanned Aerial Vehicle (UAV) image experiment. Three subimages were selected, and then they were resampled to two different levels of spatial resolution for experimentation. Among them, exponential function fitting is performed through the simulated spectrum of the PROSAIL model, and the relationship model is constructed to invert the LAI. The experimental results show that (1) the spectral unmixing method that uses the PROSAIL model to describe the variability of vegetation endmembers can obtain higher unmixing accuracy; (2) after using this spectral unmixing method, the Root Mean Square Error (RMSE) of the LAI scale effect is significantly reduced, and it has a certain effect on the correction of the LAI scale effect. This can improve the remote sensing scale effect problem to a certain extent.In summary, the spectral unmixing method has a certain effect on the correction of LAI scale differences and can improve the problem of the remote sensing scale effect to a certain extent. However, the research has the following issues that need further consideration: (1) This article only considers the vegetation-soil binary scene, but this method has not verified the multiple endmember scene. (2) The variability of soil and other background endmembers can be further considered. (3) The model can be further optimized to reduce the difference between the PROSAIL model spectrum and the real image spectrum, thereby improving the accuracy of unmixing. (4) In this paper, the evaluation is carried out by the method of upscaling. In the future, more complicated practical factors can be considered for evaluation. Simultaneously, real images collected at different flight altitudes can also be used for evaluation.关键词：remote sensing;spectral unmixing;endmember variability;prosail model;leaf area index;neural network;scale effect- 965
- |
- 361
- |
- 0
发布时间：2023-06-30 -
摘要：The rapid development of remote sensing image technology enables a large number of high-resolution remote sensing images to provide good data support for the accurate extraction of cropland and other ground features. However, high-resolution remote sensing images have large data volume and complex features, the artificial visual interpretation and traditional classification methods have limited extraction capabilities which cannot realized large-scale high-precision cropland extraction automatically. Deep learning technology has shown superior performance in the automatic extraction of remote sensing image information due to its strong ability to express features, providing a new idea for the automatic extraction of large-scale cropland. Exploring the application of different typical network models in the extraction of cropland with different landscape features is of great significance to the improvement of the quality and efficiency of cropland extraction. Based on above, the study uses the 2 m resolution data fused with GF-1 and GF-2 in 2015—2017 as the data source. Using Modified Pyramid Scene Parsing Network (MPSPNet) and UNet models applied to the fine automatic extraction of cropland in Shandong Province, and compared with the traditional object-oriented method, exploring the applicability of two deep convolutional neural network models in the automatic extraction of large-scale cropland. We also apply the trained models to the images of different regions and different time phases for the extraction of cropland, and explore the generalization ability of the models. The landscape features of cropland and uncertainty results are analyzed to explore the factors affecting the accuracy of cropland extraction by the models. Results show that: (1) MPSPNet and UNet models perform better than traditional object-oriented classification methods in the extraction of cropland at the district/county scale, the overall accuracy of the extraction of cropland at the provincial scale is better than 90% and there is no obvious difference between two models. (2) The landscape characteristic of cropland is an important factor that affects the effect of the two models, and the choice of the model has no obvious influence on the cropland extraction effect. The extraction effect is better in areas where the cropland landscape index is low and the plots are regular and flat, and the extraction effect is poor in the broken hilly areas of the plots with high cropland landscape index and in the noncropland plots whose characteristics are similar to the cropland, the UNet model is more likely to misclassify cropland in these areas. (3) The two models can obtain better cropland extraction effects in images of different regions and different time phases, and have strong generalization capabilities and temporal and spatial migration capabilities. This study proves the powerful feature learning capabilities of MPSPNet and UNET network models for high-resolution images, and the application potential of deep learning algorithms in fully automatic high-resolution cropland extraction.关键词：cropland;remote sensing;convolutional neural network;MPSPNet;UNet- 3106
- |
- 514
- |
- 4
发布时间：2023-06-30 -
摘要：The characteristics of remote sensing images, such as complex scenes, different sizes of targets and unbalanced distributions, increase the difficulty of target detection. However, feature pyramids that are suitable for detecting targets of different scales do not consider the importance of different feature maps when fusing the feature maps, let alone emphasize the features of target areas. For this purpose, this paper proposes a feature attention pyramid-based remote sensing image object detection method (namely, the feature attention pyramid network, FAPNet).First, the feature maps of different depths are fused by channel concatenation, and the features of different sized receptive fields are provided for the feature maps used for detection. The channel attention mechanism is used to recalibrate the fused feature maps in the channel dimension. The feature maps from different depths are adaptively adjusted according to the scale of the object to be detected to strengthen the feature that matches highly between the size of the receptive field and the object to be detected and weaken the feature with a low degree of matching. Second, the weakly supervised attention module uses the superimposed atrous spatial pyramid pooling structure and convolutional segmentation module to model spatial attention weights to adjust the feature distribution of the feature map used for prediction, strengthen the object area feature, and weaken the background area feature, which further improves the performance of object detection methods.The experimental results show that compared with RetinaNet, the proposed method improves the accuracy (AP) for car targets by 3.41% and 2.26% on the UCAS-AOD dataset and RSOD dataset, respectively, achieves better AP results on each target for multiclass targets and is superior to other comparative object detection methods on the mAP indicator for multitargets.A feature attention pyramid-based remote sensing image object detection method is proposed in this paper. Its contribution lies in the designed feature attention pyramid module and weakly supervised attention module. With the new modules, the proposed method can extract target features more accurately in complex scenes with targets of different sizes by channel attention and spatial attention, thus improving the performance of detection. The experimental results show that the proposed method is superior to the RetinaNet and FAN methods and is more suitable for remote sensing image object detection tasks with complex scenes and multiscale targets.关键词：remote sensing image;object detection;weakly supervised segmentation;attention mechanism;feature pyramid- 1091
- |
- 519
- |
- 2
发布时间：2023-06-30 -
摘要：Building extraction from remote sensing imagery is an important research direction for the interpretation of remote sensing imagery. UNet++ is constructed from U-Net by connecting the decoders, resulting in densely connected skip connections, enabling dense feature propagation along skip connections and thus more flexible feature fusion at the decoder nodes. However, the traditional standard convolution in the encoder fails to fully capture the multiscale features of remote sensing imagery because of the single path for semantic feature extraction, which affects the segmentation performance of the network to some extent. To address this problem, we propose a building extraction method to improve the accuracy of building extraction from remote sensing imagery.On the basis of UNet++, whose backbone is a deep residual network, we propose a building extraction network by replacing the standard convolution and max pooling in the encoder with depthwise separable convolution and applying an atrous spatial pyramid pooling structure (ASPP) to the end of the encoder. The network is referred to as the residual atrous spatial pyramid network (Res_ASPP_UNet++). On the basis of using dense and short connections to reduce the semantic gap between the encoder and the decoder, the Res_ASPP_UNet++ architecture uses multiscale ASPP made of several atrous convolutions with different sampling rates to sample the image in parallel to enrich semantic information by expanding the field of view and applies image-level features to encode the global context, avoiding segmentation errors caused by local features and improving the target segmentation performance.Experiments are conducted to validate the effectiveness of the proposed methodology. We compare the frequently used semantic segmentation networks with the Res_ASPP_UNet++ architecture using the WHU and Massachusetts datasets as data sources and take intersection over union (IoU), accuracy, precision and F1-score as evaluation indexes to evaluate the accuracy of building extraction. The experimental results are as follows: (1) The integrity of buildings extracted by Res_ASPP_UNet++ network is better than other segmentation networks on the whole, the boundary of the segmentation result is smoother and more accurate, and the result has less segmentation noise; (2) Compared with the compared semantic segmentation networks, the number of parameters of Res_ASPP_UNET ++ network is greatly compressed after improvement, the segmentation accuracy of the method is better on the whole, and it has great improvement to the UNET++ network through quantitative analysis; (3) Res_ASPP_UNET++ network has higher IoU index values than the existing building extraction algorithms listed in this paper; (4) Comparied with the proposed method without ASPP module, the value of each evaluation indexes of Res_ASPP_UNet++ with ASPP module is higher and the whole accuracy of the network with depthwise separable convolution is slightly higher than that of the network with max pooling; (5) Res_ASPP_UNet++ also extract the buildings on the Massachusetts dataset effectively and is better than other segmentation networks on the whole.The following conclusions can be drawn from the experimental results. Multiscale ASPP and depthwise separable convolution can improve the ability to express the detailed features of the model and thus effectively improve the building extraction performance from remote sensing imagery on the basis of greatly compressing the model parameters. In addition, the Res_ASPP_UNet++ architecture is robust to buildings of different scales and types and has strong generalization ability on datasets of different resolutions and sources. In future work, we will improve the segmentation for areas with a complex distribution of buildings and overcome the missing segmentation of small buildings.关键词：remote sensing imagery;building extraction;UNet++;depthwise separable convolution;deep residual structure;atrous spatial pyramid pooling- 2186
- |
- 529
- |
- 1
发布时间：2023-06-30 -
摘要：The photon counting LiDAR bathymetry system carried by UAVs is an important method for island reef mapping and shallow water bathymetry due to the characteristics of high detection sensitivity and high density. However, the high detection sensitivity also leads to the acquired photonic point cloud data with large background noise, a strong correlation between the signal-to-noise ratio and the type of ground objects, and large differences in the density distribution of photons, and the existing denoising algorithms cannot be well applied.In this paper, a denoising method for raw photon observation data is proposed. First, the effective signal interval of the raw photon observation data is calculated based on the histogram statistics method, and then the data in the interval are coarsely denoised by the grid statistics method. Finally, the local sparse coefficient method is improved, the horizontal ellipse search is used to calculate the local sparse coefficient value of each photon data in the grid, and the method of maximum interclass variance is introduced to determine the separation threshold of noise photons and signal photons, which improves the original photon observation data. Denoising accuracy. Jiajing Island and the adjacent shallow sea terrain in Hainan Province are selected as the research area to verify the denoising algorithm proposed.The results show that the average F1-score in the high signal-to-noise ratio areas, such as the island vegetation coverage area and the sandy intertidal zone, reaches 94.64% and 98.96%, respectively, and the average F1-score in the low signal-to-noise ratio area, such as the shallower and deeper water bodies near the coast, can also reach 93.04% and 90.74%, respectively. The overall F1-score is 94.34%, which can effectively remove most of the noise points and has strong adaptability to island vegetation, sandy land and underwater terrain of different depths with different signal-to-noise ratios.In addition, this paper also selects the spaceborne ICESat-2 photon dataset of coral islands in the South China Sea, which further verifies the availability and applicability of the denoising algorithm proposed in this paper on spaceborne photonic point cloud data.关键词：remote sensing;photon counting;UAV;lidar;local sparse coefficient;OTSU;noise removal- 732
- |
- 186
- |
- 0
发布时间：2023-06-30 -
摘要：The spatial density of high-quality monitoring points is an important indicator for time-series InSAR to carry out deformation monitoring. Relying on Distributed Scatterers (DS) to carry out InSAR deformation monitoring can effectively solve the defect of insufficient spatial density of traditional time-series InSAR monitoring points, but the interferometric phase of distributed scatterers is easily affected by decoherence, causing the interferometric phase distortion and unreliable, so the phase optimization of distributed scatterers is the key to DS-InSAR technology and is particularly significant. Aiming at this situation, this paper proposes a new DS phase optimization method based on singular value decomposition. This method reconstructs the phase matrix by using the time-series phase of homogeneous pixels, which belong to the same substance within the inspection window, and performs principal component analysis on the matrix to obtain the optimized phase. As a very necessary step, simulation data and 33 scene coverages of Sentinel-1A data in Baisha town, eastern Zhengzhou, are used to verify the reliability and validity of the proposed method. Using time-series average phase standard deviation, average phase gradient, and average number of residual points as the evaluation index of interferogram optimization effect, the index of the interferogram optimized by the proposed method is reduced by 15.61%, 25.81%, and 44.84% respectively compared with the original interferogram, which shows that these indicators have significant decreases. The results show that, compared with the contrast DS phase optimization method, the proposed method has a better effect on the interferogram DS phase optimization, especially in some areas with poor coherence and low signal to noise ratio. In addition, the proposed method can better maintain the detailed information of the ground features while reducing the DS phase noise. Besides, compared with the deformation monitoring results of the conventional PS(Permanent Scatterers)-InSAR technology, the number of high-quality monitoring points in this method has increased from 121471 to 644789, an increase of 4.3 times, and the density of high-quality monitoring points has increased more significantly than the comparison method. The experimental results of the simulation and real data confirm the effectiveness of the DS optimization method proposed in this paper, which can be used in DS-InSAR technology for surface deformation monitoring.关键词：remote sensing;DS-InSAR;distributed scatterers;phase optimization;singular value decomposition;deformation monitoring- 1192
- |
- 655
- |
- 0
发布时间：2023-06-30
Models and Methods
- 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