最新刊期
卷 24 ， 期 2 ， 2020
-
摘要：Ship detection on spaceborne optical images is a challenging task that has attracted increasing attention because of its potential applications in many fields. Although some ship detection methods have been proposed in recent years, many obstacles still exist because of the large-scale and high complexity of optical remote sensing images. Identifying ships from interferences, such as the features of clouds, waves, and some land architectures that are similar to ships, is difficult. Therefore, an accurate and stable deep-learning based method is proposed in this work.The method involves three steps: First, the image feature pyramid is extracted using convolution to detect multiscale ship targets. Second, a multilevel attention feature mapping structure is constructed from top to bottom using the fine-grained features of the top layer from the pyramid to improve the expressive ability of shallow features. Finally, Softmax classifier is used for multilevel ship detection.The experimental results based on real remote sensing images are shot by “JL-1” satellite, Google satellite, and NWPU VHR-10. The result proves that the performance of our algorithm is better than the three other state-of-the-art methods. In addition, the network was cut while ensuring accuracy. The complexity of our algorithm is reduced, and its practicality is improved by experiments and analysis.This work proposes an attention-based method called A-FPN. However, unlike traditional algorithms, A-FPN has higher robustness and wider range of use. Furthermore, we effectively cut the network to reduce the complexity of the algorithm, thereby exhibiting the significance of our algorithm in practical applications.关键词：optical remote sensing;ship detection;JL-1 satellite;neural network;attention features- 1083
- |
- 299
- |
- 9
发布时间：2021-10-18
Chinese Satellites
-
摘要：As an important target at sea, the realization of automatic recognition of ships is of great siguificance. The diversity of ship classes and complexity of marine environment exert higher requirements for fine-grained recognition of ships in satellite remote sensing images. Many domestic and foreign experts and scholars have carried out extensive research on ship recognition in remote sensing images. However, no universal method can recognize all kinds of ships. Most of the existing methods have strong specificity; poor portability, that is, these methods can only recognize certain classes of ship targets; and complex processing methods. This work proposes a multiscale deep learning model, which avoids complex modeling methods, such as template matching and background modeling. The network can adaptively determine the network parameters and realize the fine-grained recognition of ships by training large number of samples.First, this work uses negative sample enhancement learning to train the model, and fog and coastal backgrounds are sent as negative samples to the network for training to solve the influence of complex sea conditions, such as cloud–fog occlusion, and coastal background. At the same time, a multiscale sample training method is used in this paper in view of the problem that the size of some targets in the image is small and affects recognition accuracy. The images are compressed into multiple scales and sent to the network for training so that the network can fully learn the features of various ship sizes, thereby solving the difficulty of small target recognition. Second, the pre-trained ZF model is used for feature extraction, and the feature maps are sent to the region proposal network to generate proposal areas. Finally, the generated candidate areas are sent to the fully connected layer for ship fine-grained recognition.Experimental results show that the precision and recall of our method increased by 6.98% and 18.17% respectively, and the accuracy of ship recognition can reach 92.27% compared with Faster R-CNN. The method can guarantee real-time requirement based on high recognition accuracy and recognize ships under various conditions.The trained model can realize the fine-grained recognition of ships. The model not only solves the problems of cloud–fog occlusion to ships, but also the difficulty of small target recognition. The accuracy and real-time performance of the model meet the actual requirements and has strong robustness. However, the experimental results also show that the method used still has shortcomings. First, the constructed network structure has high complexity and excessive network overhead, which increase the processing time of ship target recognition; Secondly, the accuracy of the trained ship recognition model can still be improved. These two points are also the key tasks for the follow-up work.关键词：remote sensing;ship recognition;complex sea conditions;multi-scale;fine-grained;deep learning- 862
- |
- 265
- |
- 7
发布时间：2021-10-18 -
摘要：The Land Surface Temperatures (LSTs) obtained from remote sensing data are of great value for land surface process and climate change research. However, the spatial resolution of retrieved LSTs are relatively coarse, and most of them are mixed pixel. The spatiotemporal variability and angle effects of LSTs limit the accuracy of many scientific studies, including land surface process study.The ENVI-met model is coupled with the RAPID model to quantitatively evaluate the effects of three-dimensional (3D) heterogeneous scenes on brightness temperature and brightness temperature distribution. The key to model coupling is to maintain the consistency of the scene. The component temperature output of the ENVI-met model is used as the component temperature input parameter of the RAPID model. The surface temperature distribution of the 3D scene is simulated on the basis of the reanalysis data. The global numerical forecast product called NCEP is used to provide the boundary conditions required by ENVI-met, and the horizontal temperature distribution and the thermal radiation directionality simulation test of heterogeneous scenes were carried out. In the horizontal distribution simulation study, the onboard G-LiHT data (optical imagery, Lidar data, LST observation) provides 3D scene construction input parameters and temperature field verification data. Six different heterogeneity scenes in a certain wetland area in the United States were taken as examples for the simulation and verification in this study. Moreover, the impact of broadcast direction, road, and spatial heterogeneity on the simulation results were analyzed. In the simulation study of thermal radiation directivity, a three-dimensional scene is constructed on the basis of airborne WiDAS multiangle, multispectral data. Moreover, two heterogeneous scenes in Heihe region are taken as examples for simulation and verification.Results show that the bright temperature value from the vertical observation angle obtained by simulation is close to the LST bright temperature value of G-LiHT (RMSE is 1.1K), indicating that the coupled model can effectively simulate the bright temperature distribution under different spatial heterogeneities. Bare soil simulation error is the largest (2.31K), broadcast direction has minimal effect on the simulation (less than 1.2K), and the road width affects the simulation results (approximately 1 K). The simulation accuracy of the coupled model will decrease with the increase in spatial heterogeneity. The multi-angle simulation results of the coupled model are consistent with the variation of the directional brightness temperature in the WiDAS dataset with the viewing angle, but differences in the rate and magnitude of the change are observed.The simulation method used in this work can be utilized to predict the thermal radiation directionality of the land surface during satellite transit.关键词：remote sensing;ENVI-met model;RAPID model;LST;thermal radiation directionality;heterogeneous vegetation fields- 552
- |
- 206
- |
- 0
发布时间：2021-10-18 -
摘要：Oceanic lidar is a powerful tool that can detect the depth-resolved profiles of the upper ocean water. Seawater optical properties are usually retrieved according to the single scattering lidar equation. However, the accuracy of the simplified equation cannot be guaranteed because of the multiple light scattering in the ocean, which calls for an accurate and effective lidar return model. The accuracy of Monte Carlo (MC) simulation gains recognition because of few assumptions. However, MC simulation is limited by its low efficiency. The computation cost in the analytical model decreases. This phenomenon has not been verified in the oceanic lidar. Therefore, evaluating the accuracy of the analytical model is essential.The principles of the analytical model, conventional MC simulation, and semi-analytic MC simulation were introduced. The analytical model generally depended on the quasi-small-angle approximation. Under such approximation, the radiative transfer equation was solved in the Fourier space using the small angle approximation, which reduced the complexity of the calculation. The conventional MC simulation is based on the purely stochastic construction of an ensemble of photon trajectories through the medium of interest. The semi-analytic approach is used to reduce the statistical error of the conventional MC simulation by combining stochastic and analytic techniques.The effects of operating parameters, such as the height, field of view, water types, and distribution of phytoplankton layers, on the lidar signals were analyzed and compared. The results showed that the analytical model agrees well with the MC simulation in the homogenous and stratified water. However, in terms of the calculation efficiency, the semi-analytic MC is faster than the conventional MC, and the analytical model is faster than semi-analytic MC. As a result, high accuracy and remarkable efficiency make the analytical model superior in the simulation of the oceanic lidar return.Methods for simulating oceanic lidar signals, including the analytical model, conventional MC simulation, and semi-analytic MC simulation, were introduced in this paper. Simulations based on these methods were performed under different operating parameters and demonstrated the high accuracy and remarkable efficiency of the analytical model. These advantages make the analytical model superior in the simulation of the oceanic lidar return. The physical mechanism of laser propagation in the water and retrieval of optical parameters based on oceanic lidar will be the prospective objectives based on the method and result in this work.关键词：optical remote sensing;lidar;multiple scattering;analytical model;Monte Carlo simulation- 665
- |
- 297
- |
- 2
发布时间：2021-10-18
Technologies and Methodologies
-
摘要：The night light images obtained by the visible light imaging linear scanning business system (e.g., DMSP/OLS) and the visible light near infrared imaging radiometer (e.g., NPP/VIIRS) carried by the national polar orbit satellite are the main data sources for monitoring the social and economic activities and natural phenomena (e.g., forest fire, oil, and gas combustion). However, the existing nighttime lighting data lack the radiation of calibration, the pixel saturation, the discontinuous time scale, and the inconsistency of the multisource night light radiance.Thus, this work proposes a method of extracting invariant target area based on linear fitting and performs the mutual correction between two kinds of data between DMSP/OLS and NPP/VIIRS images by taking the Kashi of China and Pakistan Economic Corridor as the research area. Subsequently, the correction results of the economic corridor of China and Pakistan are tested on different spatial scales, including the total amount of regional gray scale, the standard difference index, and the total of standardized difference index.Results show that the goodness of fit of two calibration models is above 0.78. The correlation between the total gray level of the corrected DMSP/OLS image and the GDP and population data significantly increased (GDP:R2=0.7689; population: R2=0.9033), and the standard difference index significantly decreased after being calibrated. The calibrated NPP/VIIRS images are more consistent with DMSP/OLS in temporal and spatial distribution, and the spatial details are more prominent. This phenomenon enhances the consistency of multisource nighttime light images and is more suitable for the analysis of long-term development trends in socio-economic factors.关键词：remote sensing;DMSP/OLS;NPP/VIIRS;nighttime lights;invariant region;inter-calibration;China-Pakistan Economic Corridor- 993
- |
- 395
- |
- 11
发布时间：2021-10-18 -
摘要：In the global warming background, glaciers around the world remarkably shrunk. Glacier monitoring is an important part of cryosphere science. High-quality DEM and DOM were the primary data in glacier research. In recent years, the rapid development of the technicality of UAV provided a new platform for observing glacier. The control point was usually unevenly distributed on glacier surface because of complex and inaccessible surface conditions. Thus, this study set 17 control points on the lower part of the Laohugou Glacier No. 12. Aerial photography images were acquired by low-altitude micro-UAV covered ice tongue. Pix4D Capture was used as route planning software, and all images and control points were processed by Pix4D Mapper software. In processing, different numbers and distribution modes were used, and the accuracy of every DEM and DOM were checked. Pix4D Capture has two route planning modes, namely, single grid and double grid. The two route planning modes have the same accuracy. In ice tongue, five control points are evenly distributed, and high precision image data could be obtained. If enough control points are distributed along with the main flow line of the glacier, then the image precision is acceptable. If the control points are mainly distributed at the mid-upper or mid-lower parts of the glacier, then control point should cover the fluctuation region of the ice surface.关键词：remote sensing;UAVs;photogrammetry;DEM;DOM;the Laohugou Glacier No. 12- 745
- |
- 183
- |
- 5
发布时间：2021-10-18 -
摘要：Aerosols have important effects on global energy balance, cloud properties, rainfall frequency, and atmospheric circulation. To understand the impact of aerosols on climate change, this paper explores the quantitative relationship between fine particle aerosols and gases.Aerosol Optical Depth (AOD), Fine Mode Fraction (FMF) from MODIS, and trace gas (SO2, NO2, and HCHO) from OMI were used to analyze the quantitative relationship between the fine aerosols and trace gases over the Yellow Sea, East China Sea, and South China Sea between 2006 and 2015.First, the mean values of the aerosols and trace gases were analyzed. The mean values of AODfine, SO2, NO2, and HCHO decreased orderly in the Yellow Sea, South China Sea, and East China Sea all decreased. Meanwhile, the sensitivity analysis of the relationship between aerosols and trace gases revealed that: AODfine is most sensitive to SO2 with a sensitivity of 0.424, which may be ascribed to the anthropogenic emissions from the coastal cities of Eastern China. Meanwhile, East China Sea and South China Sea demonstrate high sensitivity to HCHO (0.664 and 0.545, respectively), which can be ascribed to the biomass combustion in Southeast Asia and Southern China. The seasonal correlation analysis of aerosols and trace gases in these three regions reveal that the AODfine in the Yellow Sea has a strong correlation with SO2 during summer and autumn (R>0.5) mainly due to the high temperature and humidity. A significant correlation is observed between HCHO and AODfine was significant in the East China Sea (R=0.57), and a relatively good correlation is observed between HCHO and AODfine in the South China Sea was relatively good (R=0.57) due to regional and seasonal changes.In sum, fine particle modal aerosol has a significant correlation with trace gases, and such relationship provides a scientific basis for understanding the aerosol processes, especially those of artificial aerosols dominated by fine particles.关键词：remote sensing;anthropogenic emissions;AODfine;aerosols;trace gases;China’s neighboring sea;spatiotemporal change- 529
- |
- 103
- |
- 0
发布时间：2021-10-18 -
摘要：In the process of urbanization, the expansion of impervious surface has a significant impact on the urban surface hydrdogical cycle, studying the temporal and spatial correlation between impervious surface and surface runoff can alleviate urban waterlogging, which can provide some suggestions for the construction of "sponge city" and scientific basis for the improvement of urban drainage system. In this work, the combination of linear spectral decomposition method with BCI index and SCS—CN model are used to simulate the surface runoff changes caused by impervious surface in the main urban area of Hangzhou in 1990—2017. Moreover, the landscape pattern index is calculated to explore the relationship between impervious surface and surface runoff.Results show that: (1) In the main urban area of Hangzhou, the impervious surface is high and increasing during the past 30 years, and the growth rate tends to be stable. The impervious surface area of Shangcheng District and Xiacheng District has the highest proportion. Jianggan District, Xihu District, and Gongshu District have a large growth rate and intensity. (2) In the past 30 years, the surface runoff of Hangzhou’s main urban area continued to increase, and the 220—240 mm runoff radiated from the center of the city to the surrounding area. The surface runoff of Shangcheng District and Xiacheng District was the highest, which was roughly consistent with the evolution of impervious surface. (3) The spatial pattern of impervious surface is highly related to surface runoff. The number and fragmentation degree of impervious surface are negatively correlated with surface runoff, and the proportion of patch occupied landscape, patch density, and patch aggregation are positively correlated with surface runoff. Surface runoff can be controlled by optimizing the space patterns of impervious surface, such as reducing large areas of impervious surfaces and increasing the number of impervious surface patches.关键词：remote sensing;hydrological model;impervious surface;spatial pattern;surface runoff;regressive analysis;Hangzhou- 886
- |
- 299
- |
- 9
发布时间：2021-10-18 -
摘要：In recent years, the problem of urban waterlogging has become prominent, which has brought many inconveniences towards social production and human life. Taking Tianjin as an example, the coupling on the spatial evolution of lakes and waterlogging was studied.Based on the Landsat series data, multi-ban water index was used to extract the lake from 1984 to 2017 year by year. Combined with the data on waterlogging, a comprehensive evaluation index system on lake evolution and waterlogging systems was proposed. Meanwhile, a coupling degree model between lake evolution and waterlogging was established.Results indicated that the lake area in Tianjin volatility increased from 1984 to 2005 and intermittently decreased from 2005 to 2017. From 2005 to 2017, the lake area in Dongli District decreased the fastest and the greatest. The proportion of lakes in most districts also decreased significantly. The waterlogging areas of Tianjin were mainly distributed in the six districts of the city centre and nearby districts. Areas without lakes were more likely to produce waterlogging than areas with lakes; areas with few lakes were more likely to produce waterlogging than areas with many lakes; the smaller the proportion of lakes, the greater the density of waterlogging points. The lake group moved from the centre towards the whole Tianjin, and the opposite direction of the lake’s centroid movement was likely to produce waterlogging.Lake areas in Tianjin have decreased in recent years, and the reduction is mainly due to human interference. The decrease of lakes is one of the important causes of waterlogging. Therefore, a primary coupling and coordination relationship exists between lake evolution and waterlogging.关键词：remote sensing;Landsat;urban waterlogging;lake area;lake centroid;coupling degree- 600
- |
- 123
- |
- 0
发布时间：2021-10-18
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