Loading...

Table of Content

    20 February 2024, Volume 53 Issue 2
    Ocean Satellite Altimetry
    Preliminary verification of dual-satellite tandem altimetry on board
    SUN Zhongmiao, ZHAI Zhenhe, GUAN Bin, RUAN Rengui, HUANG Lingyong
    2024, 53(2):  207-216.  doi:10.11947/j.AGCS.2024.20230264
    Asbtract ( )   HTML ( )   PDF (7787KB) ( )  
    References | Related Articles | Metrics
    The classic method to derive gravity field from satellite altimetry is first to calculate the deflection of the vertical using the sea surface height difference (SSHD), and then further to calculate the marine gravity anomaly and the marine geoid height, et al. Obviously, improving the measurement accuracy of SSHD can directly improve the inversion accuracy of ocean gravity field. The dual-satellite tandem altimetry principle is proposed in the paper. By designing the orbit of dual-satellite, the cross orbit distance (i.e. resolution) between the sub-satellite points of the dual-satellite can be reached about 1 arcmin. The dual-satellite simultaneously measure the SSHD along their orbital direction and in the cross orbital direction. At this time, the radial error of the orbit is showed as the relative orbital radial error between dual-satellite or between single satellite observation epochs, and the corrections related to atmospheric propagation and geophysical effects are approximately equal for dual-satellite with a ground orbit spacing of only 1 arcmin, and are rarely reflected in the SSHD. Therefore, the accuracy of SSHD will be significantly improved compared to that of the traditional single satellite condition. Using the actual observation data from our twin altimetry satellites A, B, the difference error between the relative orbit radial error and the eight corrections in the SSHD was preliminarily verified. The results show that for the calibration stage with a distance of about 25 km between sub-satellite points, the difference errors of corrections such as dry troposphere, wet troposphere, ionosphere, solid tide, polar tide, and reverse atmospheric pressure are all on the order of 5 mm. There are residual errors of approximately 1 cm and 2 cm in the difference in tidal correction and sea state deviation, respectively. For the sub-satellite point spacing of about 2 km in the business orbit, the relative orbital radial error is about 3 mm,and except for the residual error of about 0.52 cm in the sea state deviation difference, the difference error of other corrections is less than 0.05 cm and can be completely ignored.
    An improved retracker considering spatial and temporal characteristics of inland water level changes for SAR altimetry
    GAO Xianwen, JIN Taoyong, LI Jiancheng
    2024, 53(2):  217-230.  doi:10.11947/j.AGCS.2024.20230266
    Asbtract ( )   HTML ( )   PDF (15003KB) ( )  
    References | Related Articles | Metrics
    Synthetic aperture radar (SAR) altimetry has been widely used in inland water monitoring, but the echoes from inland waters are complex and need to be retracked to retrieve effective water levels. The SAR waveforms from inland waters usually contain multiple sub-waveforms, The main idea of retracking is to find the optimal sub-waveform. A reasonable water level may be obtained under specific conditions, but the wrong sub-waveform is likely to be picked in complex waters or small water bodies, which leads to abnormal water levels. To solve this problem, we proposed a spatial and temporal constrained multi-subwaveform (SaTCoM) retarcker considering time continuity and spatial stability of inland water level changes,which is different from traditional methods that finding the optimal sub-waveform by along-track or external data. Firstly, to find the sub-waveform reflected from water bodies effectively, SaTCoM interpolates and smooths the waveform. Secondly, the reference water level is obtained by filtering the sub-waveform level point clouds from all cycles in the time direction. Then, the along-track sub-waveform water levels are divided into several intervals and all along-track water levels are put into the nearest interval. Under the constraints of the maximum fluctuation from the reference water level, the indicator function characterizing the along-track spatial stability of the water level is used, which can guide the selection of the optimal water level intervals by maximum.
    In this paper, we selected 68 hydrological stations with situ level measurements of large, medium, and small water bodies in China as validation, the nearby Sentinel-3A/3B/6A satellite SAR waveforms data were used for five retrackers:SaTCoM, OCOG, ICE1, threshold, and MWaPP+. The results show that the mean RMSE of the inland water level obtained by using SaTCoM is 0.34 m, the relative RMSE is 7.8%, and the correlation with the situ level is 0.95, SaTCoM also performs excellently in most inland water surface and has an absolute advantage over the other retrackers, especially in small and medium-sized water bodies. At the same time, the analysis showed that the SaTCoM can still effectively capture the rapid water level changing signals under spatiotemporal constraints, so it indicates that the proposed retracker SaTCoM can effectively solve the difficulty of retrieving effective water levels in small water bodies, SaTCoM greatly expands the application of inland altimetry data.
    An estimation method of seabed topography based on Gauss surface function using ocean gravity data
    ZHAI Zhenhe, SUN Zhongmiao, GUAN Bin, MA Jian, LI Duan
    2024, 53(2):  231-238.  doi:10.11947/j.AGCS.2024.20230204
    Asbtract ( )   HTML ( )   PDF (6870KB) ( )  
    References | Related Articles | Metrics
    In this paper, 1'×1' grid disturbing gravity data in the South China Sea are derived from the satellite altimetry data of SARAL from January 2017 to December 2020. By comparing with the ship-borne gravity data, the accuracy is 5.5 mGal. A method is proposed to estimate the terrain model of the seabed by using gravity data and the regional characteristic parameters solved by Gauss surface function. For ETOPO-1 prior model, the accuracy of 1'×1' grid seafloor terrain estimated under a group of 10×10 grids is improved by about 10 meters. For DTU18 prior model, compared with the prior model, the 1'×1' grid seafloor terrain accuracy estimated under a set of 9×9 grids is improved by about 9 meters. To a certain extent, the results show that the five characteristic parameters obtained from gravity data retrieved from satellite altimetry can represent the surface features of the seabed topography in the corresponding area by solving the Gauss surface function, furthermore, the prior seafloor model can be refined by iterative cycle without relying on ship-borne data. Theoretically, for surface estimation, the smaller the mesh, the better the surface function can reflect the regional variation. Therefore, for the future development of satellite altimetry technology, it is expected that the gravity field detection technology with higher resolution will continue to improve the retrieval capability of seabed topographic details.
    Accurate verification and evaluation of on-board GNSS-R interferometric altimetry under on-shore conditions
    HUANG Lingyong, LI Shizhong, XIA Junming, WANG Haiyan, SUN Yueqiang, YANG Rixin, DU Qifei, HUANG Zhiyong
    2024, 53(2):  239-251.  doi:10.11947/j.AGCS.2024.20230290
    Asbtract ( )   HTML ( )   PDF (10947KB) ( )  
    References | Related Articles | Metrics
    The on-board GNSS-R interferometric altimetry technology utilizes wideband hybrid codes modulated on GNSS satellite signals to achieve high-precision wide-swath sea surface altimetry. In order to evaluate the performance of on-board GNSS-R interferometric altimetry, this article designs an accuracy evaluation strategy for on-board GNSS-R interferometric altimetry based on the results of on-shore GNSS-R interferometric altimetry. The evaluation results show that GNSS-R interferometric altimetry is related to the signal system and signal-to-noise ratio of satellite navigation systems. On-board GNSS-R interferometric altimetry can achieve centimeter-level altimetry accuracy. The research findings of this article provide technical reference and data support for the engineering practice of subsequent on-board GNSS-R interferometric altimetry technology.
    Global marine gravity anomalies recovered from multi-beam laser altimeter data of ICESat-2
    LI Zhen, GUO Jinyun, SUN Zhongmiao, JIA Yongjun, HUANG Lingyong, SUN Heping
    2024, 53(2):  252-262.  doi:10.11947/j.AGCS.2024.20230207
    Asbtract ( )   HTML ( )   PDF (10019KB) ( )  
    References | Related Articles | Metrics
    Satellite altimetry is one of the crucial techniques for the recovery of marine gravity anomalies. The along-track altimeter data is commonly used, while the cross-track data from conventional altimetry missions is not used to recover marine gravity anomalies due to the long time intervals or sparse ground-track spacing, which limits the improvement of the gravity anomaly model. The ICESat-2 laser altimetry mission operates with three pairs of laser beams, each pair separated by 3 km, which provides the possibility of using the cross-track altimeter data. The cross-track data processing method is presented and global marine gravity anomaly models (named IS2Gra_alo and IS2Gra_alo_acr) are recovered from the along-track altimeter data and the combination of along-track and cross-track altimeter data, respectively. The root mean square (RMS) of the difference between IS2Gra_alo_acr and global shipborne gravity is 5.54 mGal, which is 0.16 mGal better than that of IS2Gra_alo. According to the difference between ICESat-2 gravity models and released global gravity anomaly models, the RMS derived by IS2Gra_alo_acr is at least 0.10 mGal better than that of IS2Gra_alo. The above results confirm that the accuracy of the gravity model recovered from along-track data is improved by incorporating cross-track data, and the ICESat-2 altimeter data is reliable for the recovery of global marine gravity anomalies. Additionally, the combination of different cross-track and along-track altimeter data is discussed for the accuracy of the gravity model recovered from ICESat-2. The accuracy of the gravity model is able to be improved by incorporating the appropriate cross-track altimeter data with the along-track altimeter data. This research provides a reference for the recovery of marine gravity anomalies from the future altimeter data of the SWOT altimetry mission and the two-satellite tandem mode altimetry of China.
    Analysis of altimetry-derived sea surface observation anomalies for 2022 eruption of Tonga submarine volcano
    LI Qianqian, BAO Lifeng, WANG Yong
    2024, 53(2):  263-273.  doi:10.11947/j.AGCS.2024.20230277
    Asbtract ( )   HTML ( )   PDF (11065KB) ( )  
    References | Related Articles | Metrics
    On January 14 and 15, 2022, Tonga's submarine volcano erupted violently and caused a global tsunami, which attracted wide international attention. For the comprehensive observation of the marine environment anomalies caused by Tonga volcanic eruption, a method is proposed in this paper that the short-term and long-term effects of volcanic eruption on Marine environment are studied by using Jason-3 altimeter data. Aiming at the short-term effects, the repeated orbit of pass 186, which is closest to the volcano, is used to compare the observed values at different times to analyze the sea level change, significant wave height and ionospheric total electron content (TEC) anomalies during the volcanic eruption. The results show that the marine environment anomalies caused by the volcanic eruption can be observed by satellite altimetry. The volcanic eruption changes the sea level short-term, which is related to the changes of submarine topography, sterodynamic variability, and wave runup caused by volcanic activity. The volcanic eruption makes the significant wave height in some areas increase significantly, and the ionospheric TEC of pass 186 decreased significantly due to volcanic eruption. Aiming at the long-term effects, a 7-year regional mean sea surface observation time series from February 2016 to February 2023 in the study area was constructed, and the abnormal disturbance in the time series was detected by wavelet transform and singular spectrum analysis method. Then we analyzed the relationship between abnormal disturbance and volcanic activities, and inferred the long-term effect of volcanic eruption activities on marine environment. The results show that volcanic eruption may have a long-term effect on sea level change, the distribution of sea level anomaly is related to the location of subduction zone. Its effect is also coupled with that of EI Nino and La Nina events, so more data are needed for further analysis. In addition, the Tonga volcanic eruption has little influence on the long-term mean significant wave height and ionospheric TEC distribution in the study area.
    Multi-scale analysis of gravity anomaly models in sea area
    LIU Huanling, YANG Weiran, ZHANG Fang, WEN Hanjiang, HU Minzhang, JIANG Tao, LIN Wenqi, LI Chenxi
    2024, 53(2):  274-285.  doi:10.11947/j.AGCS.2024.20230309
    Asbtract ( )   HTML ( )   PDF (15645KB) ( )  
    References | Related Articles | Metrics
    Different from the traditional analysis method for gravity anomaly models, DOG (difference of Gauss) spherical wavelet is used to extract the gravity anomaly signals of DTU10, DTU17 and SIO V32.1 models in different bands for Mariana Trench area (140 °E-150°E, 10°N-20°N) as an example. The differences between the models are analyzed in depth. An initial attempt of multi-scale analysis in different depths and different resolutions based on RBF (radial basis function) are proposed. The results of multi-scale analysis using DOG spherical wavelet show that as the scale becomes smaller, the difference between the models becomes larger. The differences between DTU10 and DTU17 models are mainly concentrated in the band of 10.9~43.6 km, around the coast, trench and submarine mountain, which reflects the contribution of Cryosat-2, Jason-1/GM observation data and FES2014 ocean tide model. Due to the different methods for model construction, the increase of observations and the impact of waveform retracking, the difference between DTU17 and SIO V32.1 models is greater than that between DTU10 and DTU17. The traditional radial basis function is improved, and the multi-scale analysis of radial basis function under multi-depth and multi-spatial resolution is realized. The result is slightly better than that from single-depth and single-spatial resolution radial basis function. It is expected to be applied to the construction of gravity field model using multi-source data.
    Geodesy and Navigation
    Stochastic model refinement of GNSS advanced receiver autonomous integrity monitoring
    YANG Ling, ZHU Jincheng, SUN Nan, YU Yangkang, SHEN Yunzhong, LI Bofeng
    2024, 53(2):  286-295.  doi:10.11947/j.AGCS.2024.20210669
    Asbtract ( )   HTML ( )   PDF (8238KB) ( )  
    References | Related Articles | Metrics
    Global navigation satellite system (GNSS) receiver must have the capability of integrity monitoring in safety of life (SoL) applications. Advanced receiver autonomous integrity monitoring (ARAIM) which is expected to be extended to several application fields is the latest developments in the integrity monitoring of civil aviation, but the receiver-dependent term of stochastic model in ARAIM is usually established by an elevation-dependent model provided by radio technical committee for aeronautics (RTCA), which can only characterize the receiver noise of GNSS receiver of civil aviation. As a result, the performance of integrity monitoring would be adversely impacted. In this paper, the elevation-dependent model with adaptive coefficients to characterize the receiver-dependent errors is refined by the least square variance component estimation (LS-VCE) in order to extend the application scope of the ARAIM, and is verified by using the satellite-borne GNSS observation data of GRACE Follow-on (GRACE-FO) as an example. The results indicate that the overall performances of GNSS positioning and integrity monitoring are significantly improved by using the refined stochastic model. The ability of fault detection and exclusion is improved. Furthermore, protection level (PL) will decrease significantly, and so as to enhance the availability of the integrity monitoring system.
    Sliding window based GNSS dSTEC weighting method for real-time combination of global ionospheric maps
    WANG Ningbo, LI Zishen, LI Ang, ZHANG Yan, LIU Ang, WANG Liang
    2024, 53(2):  296-305.  doi:10.11947/j.AGCS.2024.20220351
    Asbtract ( )   HTML ( )   PDF (9292KB) ( )  
    References | Related Articles | Metrics
    Real-time global ionospheric corrections have been provided by several ionospheric analysis centers of the International GNSS Service (IGS) since 2017. Considering the potential unstable ionospheric streams from individual analysis centers in real applications, we propose a sliding window based differential slant total electron content (dSTEC) weighting technique for the combination of real-time global ionospheric maps (RT-GIMs). The combined RT-GIMs are generated using real-time ionospheric streams from the Chinese Academy of Sciences (CAS), Centre National d'Etudes Spatiales (CNES), Polytechnic University of Catalonia (UPC) and Wuhan University (WHU). The performance of combined RT-GIMs is validated during 15 February and 15 March 2022 in both ionospheric correction and positioning domains. The root mean square (RMS) differences between combined RT-GIM and IGS-GIM are 3.30 TECU for our combined one, and 3.20 TECU for UPC combined one. The positioning performance of combined RT-GIMs is also evaluated in ionospheric corrected single-frequency standard point positioning (SF-SPP) and ionospheric constricted single-frequency precise point positioning (SF-PPP), by analyzing the 95% quantile of positioning residuals. Compared to IGS-GIM corrected results, the positioning accuracy of combined RT-GIMs decreases by 7.7% and 4.9% in SF-SPP and SF-PPP analysis, respectively. Compared to BDGIM corrected results, the positioning accuracy of combined RT-GIMs increases by 15.9% in SF-SPP and 9.5% in SF-PPP, respectively. CAS combined RT-GIMs have been routinely provided to the IGS since 2022.
    GNSS-assisted FY-3 satellite atmospheric precipitable water retrieval method
    ZHAO Qingzhi, MA Zhi, YAO Yibin, DU Zheng
    2024, 53(2):  306-320.  doi:10.11947/j.AGCS.2024.20220538
    Asbtract ( )   HTML ( )   PDF (16119KB) ( )  
    References | Related Articles | Metrics
    Atmospheric water vapor is one of the important parameters in troposphere and has been widely used for short-term weather warning and long-term climate monitoring. The medium resolution spectral imager (MERSI) carried by FY-3 series satellites can be used for atmospheric water vapor monitoring, however, the atmospheric transmittance parameters are underestimated and the regression coefficients of water vapor and atmospheric transmittance are selected empirically when retrieving precipitable water vapor (PWV), which cannot meet the requirements of high-precision PWV applications such as short-term and imminent rainfall monitoring and numerical assimilation. Therefore, this paper proposes a high-precision PWV retrieval algorithm assisted by the global navigation satellite system (GNSS) for the FY-3 L1 data. This method introduces high-precision GNSS-derived PWV as the regression fitting parameter of the atmospheric transmittance calculation model, and assists the FY-3 L1 data to accurately estimate the model regression coefficients of PWV and atmospheric transmittance. In addition, this method considers the seasonal and elevation impact on PWV retrieval, and retrieves the PWV according to the season and corrects the FY-3-L1 PWV bias caused by the underestimation of some atmospheric transmittance parameters using digital elevation model. The L1 data of the FY-3 A satellite (FY-3A) and the data of 260 GNSS stations of the China crustal movement observation network in the Chinese region over the period of 2013 to 2014 were selected for the experiment. Results show that the GNSS-assisted FY-3 series satellite PWV retrieval algorithm proposed in this paper is superior to the traditional method (FY-3A-L2 PWV), and its overall accuracy improvement rate is 74.5%, and a more reliable and robust PWV grid product can be obtained,and it is of great significance for short-term and imminent rainfall monitoring and numerical assimilation.
    Photogrammetry and Remote Sensing
    Learned local features for SfM reconstruction of UAV images
    JIANG San, LIU Kai, LI Qingquan, JIANG Wanshou
    2024, 53(2):  321-331.  doi:10.11947/j.AGCS.2024.20220636
    Asbtract ( )   HTML ( )   PDF (13866KB) ( )  
    References | Related Articles | Metrics
    Reliable feature matching plays an essential role in SfM (structure from motion) for UAV (unmanned aerial vehicle) images. Recently, deep learning-based methods have been used for feature detection and matching, which outperforms traditional handcrafted methods, e.g., SIFT, on benchmark datasets. However, few studies have reported their performance on UAV images as these models are trained and tested using internet photos. By using UAV datasets with varying features, this study evaluated both handcrafted and learned methods in terms of feature matching and SfM-based image orientation. The experimental results show that even with the pretrained public-available models, more accurate and complete feature matching can be obtained through the combination of high-precision localization of handcrafted detectors and the high representation ability of learned descriptors, which has competitive or better performance in SfM-based image orientation when compared with SIFT-like handcrafted methods.
    Robust merging of subblock reconstructions for parallel structure from motion in photogrammetry
    XIAO Teng, WANG Xin, MEI Xi, YE Zhiwei, YAN Qingsong, DENG Fei
    2024, 53(2):  332-343.  doi:10.11947/j.AGCS.2024.20220321
    Asbtract ( )   HTML ( )   PDF (12626KB) ( )  
    References | Related Articles | Metrics
    In this paper, we proposed an improved parallel structure from motion pipeline in photogrammetry by robustifying the merging of subblock reconstructions in a better fashion. Specifically, the whole block represented by a view graph is divided into a number of overlapped subblocks via graph partition and expansion, and an improved incremental SfM is employed to generate the SfM reconstruction of each subblock. To merge these subblock SfM reconstructions in a more robust manner, a subblock graph indicating the overlapping relationship of subblock reconstructions is first built. By considering the geometry consistencies of subblock triplets, gross errors are detected. Then, we leverage the algebraic properties of subblock triplets, which aims to make them more geometrically consistent, to refine the relative transformations between subblock reconstructions. Finally, more accurate relative transformations between subblock reconstructions can be obtained to boost the subsequential merging. Experimental results using UAV images show that the proposed method can guarantee robustness in the subblock reconstruction merging stage. The precision of our SfM results is better than several state-of-the-art parallel SfM methods and the popular COLMAP. Furthermore, it has significant potential for use in photogrammetry and 3D Real Scene reconstruction.
    Monocular height estimation method of remote sensing image based on Swin Transformer-CNN and its application in highway road construction sites
    LIAO Zhaohong, ZHANG Yichen, YANG Biao, LIN Mingchun, SUN Wenbo, GAO Zhi
    2024, 53(2):  344-352.  doi:10.11947/j.AGCS.2024.20220607
    Asbtract ( )   HTML ( )   PDF (8376KB) ( )  
    References | Related Articles | Metrics
    At present, under the good condition of image geometry and radiation quality, the technology of 3D scene reconstruction by intensive matching of multi-view aerospace image is relatively mature, which has achieved good results both in accuracy and efficiency. However, when multi-view aerospace images with good geometric conditions are difficult to obtain, the geometric processing methods of classical photogrammetry and computer vision may face great challenges. In this paper, we study this problem and propose a monocular height estimation method of remote sensing image based on Swin Transformer and convolutional neural network (CNN). Swin Transformer is a hierarchical transformer structure with shifted windows. It combines the ability of convolutional neural network to process large scale image and extract multi-scale features, as well as the global information interaction ability of transformer. In addition, our method reformulates the height estimation problem into a classification-regression problem to improve model performance. Specifically, for each input image, our model classifies the height range into several discrete bins adaptively, where continuous height value is estimated via a linear combination of predicted discrete bins and height distribution probability. In experiments, we qualitatively and quantitatively demonstrate that the proposed method outperforms the state-of-the-art approaches, and it can also be applied to highway road construction sites with good generalization.
    A coupled DeepLab and Transformer approach for fine classification of crop cultivation types in remote sensing
    LIN Yunhao, WANG Yanjun, LI Shaochun, CAI Hengfan
    2024, 53(2):  353-366.  doi:10.11947/j.AGCS.2024.20220692
    Asbtract ( )   HTML ( )   PDF (15707KB) ( )  
    References | Related Articles | Metrics
    How to accurately monitor the planting of different types of complex farmland crops by remote sensing is the key to the realization of agricultural area survey and crop yield estimation in the area of smart rural agriculture. In the current pixel level semantic segmentation of crop planting in high-resolution images, the deep convolution neural network is difficult to take into account the spatial multi-scale global features and local details, which leads to problems such as blurring boundary contours between various farmland plots and low internal integrity of the same farmland area. In view of these shortcomings, this paper designs and proposes a dual branch parallel feature fusion network (FDTNet) that couples DeepLabv3+and Transformer encoders to achieve fine remote sensing monitoring of crop planting. Firstly, DeepLabv3+and Transformer are embedded in FDTNet in parallel to capture the local and global features of farmland image respectively. Secondly, the coupled attention fusion module (CAFM) is used to effectively fuse the characteristics of the two features. Then, in the decoder stage, the convolutional block attention module (CBAM) is applied to enhance the weight of the effective features of the convolutional layer. Finally, the progressive multi-level feature fusion strategy is adopted to fully fuse the effective features in the encoder and deco-der, and output the feature map to achieve high-precision classification and recognition of late rice, middle rice, lotus root field, vegetable field and greenhouse. In order to verify the effectiveness of FDTNet network model in high-resolution crop classification application, this paper selects different high-resolution Yuhu dataset and Zhejiang dataset and experimental results of mIoU reach 74.7% and 81.4%, respectively. The mIoU of FDTNet can be 2.2% and 3.6% respectively higher than the existing deep learning methods, such as UNet, DeepLabv3, DeepLabv3+, ResT and Res-Swin. The results show that FDTNet has better classification performance than the compared methods in two types of farmland scenes, which have single texture and large sample size, or multiple texture and small sample size. The proposed FDTNet has a comprehensive ability to extract effective features of multiple category crops.
    Cartography and Geoinformation
    Road section navigation attribute mining
    ZHANG Caili, XIANG Longgang, LI Yali, GAO Songfeng, PAN Chuanjiao
    2024, 53(2):  367-378.  doi:10.11947/j.AGCS.2024.20220649
    Asbtract ( )   HTML ( )   PDF (10020KB) ( )  
    References | Related Articles | Metrics
    A large number of automatic or semi-automatic road extraction methods have sprung up, but the generated products usually lacked navigation attribute information, such as road hierarchies, road speed limits, etc., which restrict intelligent navigation services such as "main road priority" and "speed limit alert". Therefore, taking road sections as the unit of analysis and considering the strong correlation between adjacent upstream and downstream road sections, a method for mining attribute information of road hierarchies and road speed limits was proposed. First, we preprocessed tracks and road networks and realized the connection between track points and road sections. Then, multi-modal features were designed based on the understanding of the data and the task. Finally, the random forest was used to recognize road hierarchies and road speed limits, taking into account the information of the target road segments and their upstream and downstream adjacency information. Compared with single-class features, the integration features of road networks and trajectories improve the classification accuracy of road hierarchies and road speed limits; compared with the classification of road hierarchies and road speed limits considering only the target road segment, the classification of road hierarchies and road speed limits considering spatial adjacency information is more effective.
    An adaptive road centerline extraction method for different trajectory data scenarios based on combinatorial optimization
    YAO Zhipeng, PENG Cheng, TANG Jianbo, LIU Guoping, YANG Xuexi, LIU Huimin, DENG Min
    2024, 53(2):  379-390.  doi:10.11947/j.AGCS.2024.20220606
    Asbtract ( )   HTML ( )   PDF (13808KB) ( )  
    References | Related Articles | Metrics
    Vehicle trajectory data is an important data source for road map update. Extracting road centerlines from the disordered trajectory points or trajectory lines, and generating a structured vector map is a key step for road network generation and update based on trajectory data. The existing methods of road centerline extraction mainly use a single curve fitting algorithm, which are not adaptive to different data scenarios, especially for complex road structures and trajectories of different quality. In addition, compared with the professional collected high-frequency trajectory data, road centerline extraction based on the low-frequency trajectory data collected by float cars is still challenging due to the noise, sparsity, and low position accuracy. Therefore, this paper proposes an adaptive road centerline extraction method for different trajectory data scenarios based on combinatorial optimization and divide-and-conquer strategy. Based on preprocessing and clustering of trajectory data, this method classifies the trajectory data according to its distribution characteristics. Then, the optimal fitting algorithm is matched according to different data scenarios, and the ideal road centerline is generated by combinatorial optimization strategy. This method integrates the advantages of different fitting algorithms, and can effectively solve the road centerline extraction problem for different data scenarios such as sparse data and complex road structures (e.g. self-intersection overpasses). Experiments on floating car data in Beijing, China, were conducted and results show that the average position accuracy of the roads generated by this method is 1.24 m, which is significantly better than the existing available methods.
    Summary of PhD Thesis
    Research on computing methods of separable non-linear least squares and application in LiDAR waveform decomposition
    WANG Ke
    2024, 53(2):  391-391.  doi:10.11947/j.AGCS.2024.20220697
    Asbtract ( )   HTML ( )   PDF (1623KB) ( )  
    Related Articles | Metrics
    An investigation of key technologies related to combining BDS-2 and BDS-3 observations in data processing
    HU Chao
    2024, 53(2):  392-392.  doi:10.11947/j.AGCS.2024.20220698
    Asbtract ( )   HTML ( )   PDF (1375KB) ( )  
    Related Articles | Metrics
    Study on the strata movement rule and regional control of deep mining with the super-thick and weak cementation overburden
    ZHANG Guojian
    2024, 53(2):  393-393.  doi:10.11947/j.AGCS.2024.20220704
    Asbtract ( )   HTML ( )   PDF (1635KB) ( )  
    Related Articles | Metrics
    Research on the key technologies of InSAR/GIS monitoring for illegal underground mining in coal mining area
    XIA Yuanping
    2024, 53(2):  394-394.  doi:10.11947/j.AGCS.2024.20220707
    Asbtract ( )   HTML ( )   PDF (1258KB) ( )  
    Related Articles | Metrics
    Study on the three-dimensional water vapor tomography method combining GNSS and RS data and its applications
    ZHANG Wenyuan
    2024, 53(2):  395-395.  doi:10.11947/j.AGCS.2024.20220710
    Asbtract ( )   HTML ( )   PDF (1351KB) ( )  
    Related Articles | Metrics
    Research and application of key technologies for high accuracy pseudolite system
    FAN Caoming
    2024, 53(2):  396-396.  doi:10.11947/j.AGCS.2024.20230001
    Asbtract ( )   HTML ( )   PDF (1355KB) ( )  
    Related Articles | Metrics
    Research on key algorithms for long-time-series deformation monitoring by Sentinel-1 InSAR
    MA Zhangfeng
    2024, 53(2):  397-397.  doi:10.11947/j.AGCS.2024.20230002
    Asbtract ( )   HTML ( )   PDF (1182KB) ( )  
    Related Articles | Metrics
    Point cluster generalization approaches taking into account the weights of the points
    LU Xiaomin
    2024, 53(2):  398-398.  doi:10.11947/j.AGCS.2024.20230010
    Asbtract ( )   HTML ( )   PDF (1340KB) ( )  
    Related Articles | Metrics