Most Down Articles

Published in last 1 year | In last 2 years| In last 3 years| All| Most Downloaded in Recent Month| Most Downloaded in Recent Year|

Most Downloaded in Recent Month

Please wait a minute...


For Selected: Download Citations EndNote Ris BibTeX Toggle Thumbnails

Select

The progress and trend of cartography and geographic information engineering in China (2019—2023) SUN Qun, REN Fu, LONG Yi, CHEN Xin, XU Gencai, LIU Jianjun, ZHOU Zhao Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 399-412.   DOI: 10.11947/j.AGCS.2024.20230562 Abstract （167）   HTML （40）    PDF（pc） （12949KB）（278）       Knowledge map    Save Based on the development of cartography and geographic information technology in China since 2019, this report summarizes the achievement in theories of cartography, digital cartography and press techniques, updating of China's National Fundamental Geographic Information Database, geographic information technology and industrial development, geographic information applications and services, atlas compiling and publishing, and the new online map. Finally, the report offers some prospects for the developing trend of cartography and geographic information engineering in China. Reference | Related Articles | Metrics

Select

GNSS ultra-rapid orbit and clock offset estimation method with the aid of the constraint of BDS-3 onboard clock HU Chao, WANG Qianxin Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 413-424.   DOI: 10.11947/j.AGCS.2024.20230168 Abstract （155）   HTML （28）    PDF（pc） （12929KB）（119）       Knowledge map    Save The high-stability BDS-3 onboard clock is one of the significant advanced technologies of BDS, which is underutilized in GNSS data processing. To solve the precision restriction of GNSS ultra-rapid orbit and clock offset parameters under the strict timeliness limitation, GNSS ultra-rapid orbit and clock offset estimation method with the aid of the constraint of BDS-3 onboard clock is proposed in this research. Firstly, based on the correlation GNSS clock offset and orbit, the GNSS orbit determination model is constructed with the consideration of BDS-3 onboard clock parameter. Secondly, impacts of onboard clock constraints on GNSS orbit determination are analyzed by taking BDS-3 precise clock offsets as example in experiments. Thirdly, to overcome the influence of the prediction clock offset and selected constraints on GNSS ultra-rapid orbit determination, the synchronously processed method of BDS-3 clock modelling and GNSS ultra-rapid orbit estimation is proposed. According to the experimental results, it is indicated that the accuracy of BDS-3 and GPS orbit and clock offsets can be respectively improved with 27.5%, 5.1% and 20.2%, 5.2%, under the optimal constraint on BDS-3 onboard clock. Meanwhile, compared with the traditional epoch-wise white noise strategy of GNSS clock parameter, the proposed one-step processing method for BDS-3 satellite clock modeling and GNSS ultra-rapid orbit determination can respectively improve the accuracy of GNSS ultra-rapid clock offset and orbit up to 4.8% and 34.2%, where the millimeter-level orbit accuracy improvements can be obtained. Therefore, the proposed GNSS ultra-rapid orbit and clock offset estimation method based on BDS-3 clock offset constraints can effectively utilize the information of BDS-3 highly stable onboard clock information, and realize the accuracy improvement of GNSS ultra-rapid orbit and clock offsets. Reference | Related Articles | Metrics

Select

Research on deep learning models for hyperspectral image classification PU Shengliang Acta Geodaetica et Cartographica Sinica    2023, 52 (1): 172-172.   DOI: 10.11947/j.AGCS.2023.20210203 Abstract （252）   HTML （19）    PDF（pc） （729KB）（714）       Knowledge map    Save Related Articles | Metrics

Select

A distributed GNSS/SINS/odometer resilient fusion navigation method for land vehicle MU Mengxue, ZHAO Long Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 425-434.   DOI: 10.11947/j.AGCS.2024.20220349 Abstract （89）   HTML （9）    PDF（pc） （8885KB）（111）       Knowledge map    Save To improve the fault-tolerance of a low-cost land vehicle navigation system in the complex environment, this paper proposes a distributed GNSS/SINS/odometer resilient fusion method based on the suboptimal gain fusion algorithm. First, a velocity compensation model for each odometer on four wheels is established according to the Ackermann steering geometry, which improves the accuracy of forward and lateral velocity measurement at the inertial measurement unit center. Then, a fault detection and classification criteria based on Chi-square test statistics is designed to make full use of the available observation information. Last, a resilient adjustment model for the stochastic model and information sharing factors (ISF) are proposed to mitigate the influence of abnormal observation from the sensor layer and the decision layer respectively and realize the resilient fusion of multi-source information. A real car test is carried out to verify the effectiveness of the distributed GNSS/SINS/odometer resilient fusion method. The experiment results demonstrate that the proposed method can effectively reduce the impact of subsystem faults on the global state estimation and improve the fault tolerance performance of the system in complex environments. Moreover, compared with the traditional federated Kalman filtering (FKF), the SGF algorithm can achieve the equivalent accuracy with significant computational efficiency improvement, which is conducive to the practical engineering application of multi-source information resilient fusion. Reference | Related Articles | Metrics

Select

GPS trajectory agglomeration and refined road network extraction WU Qunyong, WU Zufei, ZHANG Liangpan Acta Geodaetica et Cartographica Sinica    2019, 48 (4): 502-511.   DOI: 10.11947/j.AGCS.2019.20180256 Abstract （2008）   HTML    PDF（pc） （4008KB）（2001）       Knowledge map    Save Aiming at the shortcomings of low-accuracy in the use of GPS data to extract bidirection a roads and intersections, this paper proposes a trajectory agglomeration and refined roads extraction method that takes into account the position and travel direction to extracts refined road network. First,we remove the discrete and abnormal trajectory points from the original trajectory and insert the trajectory points into the trajectory segments by a certain step size, in order to improve the extraction accuracy of the intersection network.Second,we introduce the driving direction angle to express the driving direction of the vehicle at the track point, obtain its similar trajectory points set by considering the position and direction of the track point, calculate the offset distance of each track point in turn, and complete the track aggregation by iteratively offsetting the track points.Finally, we eliminate the track points that have not been successfully gathered, and use the Grid digitization method to extract the road network that can reflect the fine steering relationship of the roads from the trajectory data after gather. The trajectory agglomeration and road network extraction experiments were carried out with GPS data of Fuzhou taxis. The experimental results show that this method can effectively gather the GPS trajectories according to the direction of vehicle travel and the extracted road network is bidirectional roads, and can finely reflect the steering relationship of the roads at the intersections. Reference | Related Articles | Metrics

Select

Research Progress and Methods of InSAR for Deformation Monitoring ZHU Jianjun, LI Zhiwei, HU Jun Acta Geodaetica et Cartographica Sinica    2017, 46 (10): 1717-1733.   DOI: 10.11947/j.AGCS.2017.20170350 Abstract （4779）   HTML    PDF（pc） （1654KB）（9185）       Knowledge map    Save Deformation monitoring is one of the most mature applications of space-borne InSAR technique. Firstly, we introduce the basic principle of InSAR in the monitoring of deformation and the current SAR satellites. The deformation monitoring methods of InSAR are then classified into the groups of D-InSAR, PS-InSAR, SBAS-InSAR, DS-InSAR and MAI, which are analyzed in the aspects of technical features and application scopes. Subsequently, we analyze the research progress and deficiencies of InSAR in the investigation of urban, mining area, earthquake, volcano, infrastructure, glacier, permafrost and landslide. Finally, some advanced academic problems such as deformation monitoring in multi-demension and low coherence area, atmospheric and orbital errors mitigation, and accuracy assessment are concluded. Reference | Related Articles | Metrics

Select

Progress, Challenges and Perspectives of 3D LiDAR Point Cloud Processing YANG Bisheng, LIANG Fuxun, HUANG Ronggang Acta Geodaetica et Cartographica Sinica    2017, 46 (10): 1509-1516.   DOI: 10.11947/j.AGCS.2017.20170351 Abstract （4205）   HTML    PDF（pc） （1579KB）（9494）       Knowledge map    Save 3D LiDAR can perform an intensive sampling of the earth surface in a direct way, and yield the 3D point cloud that contains numerous and scattered points with the coordinates ( X, Y, Z) and attributes (e.g., intensity). As the vital 3D geospatial data for description of the world in the digital era, 3D point cloud plays an important role not only in earth science researches but also in national requirements (e.g., global change analysis, global mapping, and smart city). Inspired by sensor technologies and national requirements, 3D LiDAR has got great progresses in hardware, data processing and applications, and is facing new challenges. Following the history of 3D LiDAR, this paper first reviews the status of 3D LiDAR system, and introduces the development of key technologies in data processing. Then the typical applications of 3D LiDAR in surveying and other related fields are listed, and current challenges in point cloud processing are concluded. Finally, some future perspectives are presented. Reference | Related Articles | Metrics

Select

An InSAR phase unwrapping method based on R2AU-Net HE Yi, YANG Wang, ZHU Qing Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 435-449.   DOI: 10.11947/j.AGCS.2024.20230305 Abstract （90）   HTML （6）    PDF（pc） （17203KB）（87）       Knowledge map    Save The accuracy of terrain elevation or surface deformation retrieval relies heavily on the quality of InSAR phase unwrapping. Conventional phase unwrapping techniques, rooted in non-machine learning models (such as path-following or minimum norm), face challenges in producing accurate unwrapping outcomes within areas of low coherence or high phase gradients (dense interference fringes). Deep neural network models offer distinct advantages in nonlinear representation and feature expression, widely employed in digital image processing research, wherein InSAR phase unwrapping parallels image regression.This paper presents an InSAR phase unwrapping approach utilizing the R2AU-net. Initially, pairs of wrapped and unwrapped phases are simulated through mathematical fractal methods, circumventing inherent errors and artifacts introduced by integrating external DEMs into the phase. This approach maintains terrain feature diversity and complexity while providing the requisite dataset for model training. Subsequently, the R2AU-net phase unwrapping model, built upon the foundational U-net model, incorporates attention mechanisms to augment the model's convolutional feature selection capacity, thereby improving unwrapping performance in regions of low coherence or dense striping. The utilization of recurrent residual convolutional structures addresses the vanishing gradient issue, enhancing the model's feature representation capability.Ultimately, experimental analyses are conducted using both simulated and real data. The results demonstrate that the proposed R2AU-net phase unwrapping model effectively retains terrain elevation or real surface deformation information, thereby bolstering the reliability of unwrapping outcomes. In terms of performance, it surpasses established methods such as the Goldstein branch-cut method, SNAPHU method, as well as CNN and U-Net phase unwrapping models. Reference | Related Articles | Metrics

Select

Automatic Analysis and Mining of Remote Sensing Big Data LI Deren, ZHANG Liangpei, XIA Guisong Acta Geodaetica et Cartographica Sinica    2014, 43 (12): 1211-1216.   DOI: 10.13485/j.cnki.11-2089.2014.0187 Abstract （4653）   HTML    PDF（pc） （1314KB）（11056）       Knowledge map    Save With the diversification of the imaging methods and the growing categories, quantity, and observation frequency of remote sensing data, the ability of land-cover observation has reached an unprecedented level, which means a new era of big data in remote sensing is coming. However, the existing methods and processing techniques cannot fulfill the need of the big data application in remote sensing. Thus, to develop the automatic analysis and mining theory and techniques for remote sensing big data is among the most advanced international research areas. This paper investigates and analyses the domestic and overseas research status and progress around this field and points out its key problems and developing trends. Reference | Related Articles | Metrics

Select

Deep learning algorithm for feature matching of cross modality remote sensing images LAN Chaozhen, LU Wanjie, YU Junming, XU Qing Acta Geodaetica et Cartographica Sinica    2021, 50 (2): 189-202.   DOI: 10.11947/j.AGCS.2021.20200048 Abstract （3289）   HTML （237）    PDF（pc） （25649KB）（3275）       Knowledge map    Save Focusing on the problem of difficulty in matching due to the differences in imaging modality, time phases, and resolutions of cross modality remote sensing images, a new deep learning feature matching method named CMM-Net is proposed. First, a convolutional neural network is used to extract high-dimensional feature maps of the cross modality remote sensing images. The key points are selected according to the conditions that both the channel maximum and local maximum are met, and the 512-dimensional descriptors in corresponding location are extracted on the feature map to complete the feature extraction. In the matching stage, after completing the fast-nearest neighbor searching, in order to solve the problem of lots of mismatched points, a purification algorithm with dynamic adaptive Euclidean distance and RANSAC constraints is proposed to ensure that the mismatches are effectively eliminated while retaining the correct matches. The algorithm was tested using multiple sets of cross modality remote sensing images and compared with other algorithms. The results show that the proposed algorithm has the ability to extract similar scale invariant features in cross modality images, and has strong adaptability and robustness. Reference | Related Articles | Metrics

Select

Log-Gabor filter-based high-precision multi-modal remote sensing image matching CAO Fanzhi, SHI Tianxin, HAN Kaiyang, WANG Pu, AN Wei Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 526-536.   DOI: 10.11947/j.AGCS.2024.20230059 Abstract （64）   HTML （6）    PDF（pc） （12002KB）（74）       Knowledge map    Save A feature matching method based on Log-Gabor filtering is proposed to address the problem of high-precision matching for multimodal remote sensing images. The method adopts a multi-scale dense matching framework via a coarse-to-fine manner, which avoids the low repeatability problem of feature detectors in multimodal images and is able to extract a large number of accurate correspondences. The method consists of two main steps: first, a feature pyramid robust to non-linear radiometric differences between images is constructed using multi-scale multi-angle Log-Gabor filters; then, the coarse feature map is used for dense template matching to extract a large number of coarse feature correspondences; the feature pyramid is then used to achieve bottom-up refinement of coarse correspondences layer by layer. Furthermore, to address the problem of inefficient sliding window operation for template matching, a fast implementation method of dense template matching is proposed, which effectively reduces the running time of dense template matching. The results show that the proposed method can overcome the influence of non-linear radiation differences between images, and outperforms existing multimodal image feature matching methods in terms of the number of correct matches, matching accuracy and matching precision. Reference | Related Articles | Metrics

Select

Geographic Knowledge Graph Building Extracted from Multi-sourced Heterogeneous Data JIANG Bingchuan, WAN Gang, XU Jian, LI Feng, WEN Huiqi Acta Geodaetica et Cartographica Sinica    2018, 47 (8): 1051-1061.   DOI: 10.11947/j.AGCS.2018.20180113 Abstract （2671）   HTML    PDF（pc） （2571KB）（4346）       Knowledge map    Save As a new generation of geographic language,virtual geographic environments(VGE) needs to construct the virtual geographic knowledge engineering through the analysis and excavation of multi-sourced heterogeneous spatio-temporal data.Through the transformation about "data-information-knowledge-wisdom",it can promote the rapid conversion and integration of geographic knowledge for intelligent VGE system,which may provide a theoretical and technical support for intelligent processing of geographic information and intelligent service of geographic knowledge.Knowledge graph plays the role of a bridge between artificial intelligence and knowledge engineering for VGE system.This paper firstly reviewed the research status in knowledge graph and geographic knowledge graph from the domestic and abroad perspectives.Then,the construction process of geographic knowledge graph is proposed.Furthermore,the key technologies of geographic knowledge graph are discussed in depth.Finally,the application direction of geographic knowledge graph is discussed and stated.The study of geographic knowledge graph in this paper may help to realize the knowledgization of geographic information and promote the intelligent service level for VGE system. Reference | Related Articles | Metrics

Select

An Overview on Data Mining of Nighttime Light Remote Sensing LI Deren, LI Xi Acta Geodaetica et Cartographica Sinica    2015, 44 (6): 591-601.   DOI: 10.11947/j.AGCS.2015.20150149 Abstract （3690）   HTML    PDF（pc） （1051KB）（6444）       Knowledge map    Save When observing the Earth from above at night, it is clear that the human settlement and major economic regions emit glorious light. At cloud-free nights, some remote sensing satellites can record visible radiance source, including city light, fishing boat light and fire, and these nighttime cloud-free images are remotely sensed nighttime light images. Different from daytime remote sensing, nighttime light remote sensing provides a unique perspective on human social activities, thus it has been widely used for spatial data mining of socioeconomic domains. Historically, researches on nighttime light remote sensing mostly focus on urban land cover and urban expansion mapping using DMSP/OLS imagery, but the nighttime light images are not the unique remote sensing source to do these works. Through decades of development of nighttime light product, the nighttime light remote sensing application has been extended to numerous interesting and scientific study domains such as econometrics, poverty estimation, light pollution, fishery and armed conflict. Among the application cases, it is surprising to see the Gross Domestic Production (GDP) data can be corrected using the nighttime light data, and it is interesting to see mechanism of several diseases can be revealed by nighttime light images, while nighttime light are the unique remote sensing source to do the above works. As the nighttime light remote sensing has numerous applications, it is important to summarize the application of nighttime light remote sensing and its data mining fields. This paper introduced major satellite platform and sensors for observing nighttime light at first. Consequently, the paper summarized the progress of nighttime light remote sensing data mining in socioeconomic parameter estimation, urbanization monitoring, important event evaluation, environmental and healthy effects, fishery dynamic mapping, epidemiological research and natural gas flaring monitoring. Finally, future trends of nighttime light remote sensing and its data mining have been proposed from four aspects including new data source, knowledge discovery, in-situ observation, and national/global geographic conditions monitoring. Reference | Related Articles | Metrics

Select

A local encryption method for vector maps based on multilevel spatial index structure DING Chen, PENG Cheng, TANG Jianbo, DENG Min, YANG Xuexi, LIU Huimin Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 569-581.   DOI: 10.11947/j.AGCS.2024.20220564 Abstract （58）   HTML （1）    PDF（pc） （12120KB）（67）       Knowledge map    Save Vector maps are the basic data sources of location-based services. In the big data era, vector maps have been used in most of our daily life service applications. The development of big data technologies, which provide great convenience for sending, sharing, acquisition and dynamic update of vector maps, also make it is severe of the security risks such as information leakage and stealing in the process of map data transmission. Existing map encryption methods are usually based on traditional cryptographic algorithms and encrypt the entire vector map, without making full use of the spatial data structure and the distribution characteristics of geographical entities in maps, which make it is difficult for these methods to encrypt parts of the data and cannot meet the needs of applications such as crowdsourced data collection, personalized map services for encrypting parts of the data rather than the entire data. Therefore, this paper proposes a local encryption method for vector maps based on multilevel spatial index structure. Experimental results on real point, line and polygon vector maps show that: ① The proposed method can effectively perform local encryption and local decryption for vector maps; ② Compared with the state-of-the art encryption methods, the proposed method has higher encryption efficiency; ③ The security and the anti-attack ability of the proposed method are significantly improved by using different encryption parameters for different encryption units, which ensures the security of the encryption results. Reference | Related Articles | Metrics

Select

Urban Intersection Recognition and Construction Based on Big Trace Data TANG Luliang, NIU Le, YANG Xue, ZHANG Xia, LI Qingquan, XIAO Shilun Acta Geodaetica et Cartographica Sinica    2017, 46 (6): 770-779.   DOI: 10.11947/j.AGCS.2017.20160614 Abstract （2113）   HTML    PDF（pc） （6538KB）（3910）       Knowledge map    Save Intersection is an important part of the generation and renewal of urban traffic network. In this paper, a new method was proposed to detect urban intersections automatically from the spatiotemporal big trace data. Firstly, the turning point pairs were based on tracking the trace data collected by vehicles. Secondly, different types of turning point pairs were clustered by using spatial growing clustering method based on angle and distance differences, and the clustering methods of local connectivity was used to recognize the intersection. Finally, the intersection structure of multi-level road network was constructed with the range of the intersection and turning point pairs. Taking the taxi trajectory data in Wuhan city as an example, the experimental results showed that the method proposed in this paper can automatically detect and recognize the road intersection and its structure. Reference | Related Articles | Metrics

Select

The positioning method of lunar observations by Jilin-1 Guangpu satellite JING Zhenhua, HU Xiuqing, LI Shuang Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 503-511.   DOI: 10.11947/j.AGCS.2024.20220597 Abstract （65）   HTML （13）    PDF（pc） （7475KB）（65）       Knowledge map    Save The clarification of pixel positions in remote sensing images is an essential prerequisite for data applications, and in combination with the high-resolution Moon views obtained by Earth-orbiting satellites, it is promising to calibrate satellite instruments directly by the radiance of specific regions of the lunar surface. This paper proposes a method to calculate the lunar image coordinates without relyingstrictly on the instrument parameters for the Jilin-1 GuangPu02 (JL1GP02) lunar observations. The selenographic coordinates of the two-dimensional image pixels are calculated by the satellite view libration at the observation moment and the orientation relationship between the Moon and instrument scans, and the geometric positioning quality is evaluated by creating simulated images. Results show that the geometric deviations of the lunar observation images at different resolutions are usually within 1.3 pixels, and the time series results for about two years agree well and generally follow a normal distribution, indicating the reliability of the positioning results. The method can avoid complicated coordinate transformation and achieve image-level geometric positioning of lunar images when the interior orientation parameters are unknown. Reference | Related Articles | Metrics

Select

Research progress of geodesy in China (2019—2023) DANG Yamin, JIANG Tao, YANG Yuanxi, SUN Heping, JIANG Weiping, ZHU Jianjun, XUE Shuqiang, ZHANG Xiaohong, YU Baoguo, LUO Zhicai, LI Xingxing, XIAO Yun, ZHANG Chuanyin, ZHANG Baocheng, LI Zishen, FENG Wei, REN Xia, WANG Hu Acta Geodaetica et Cartographica Sinica    2023, 52 (9): 1419-1436.   DOI: 10.11947/j.AGCS.2023.20230343 Abstract （799）   HTML （112）    PDF（pc） （1294KB）（1378）       Knowledge map    Save From July 11 to 20, 2023, the 28th International Union of Geodesy and Geophysics (IUGG) general assembly was held in Berlin, Germany. According to the tradition of IUGG, the Chinese National Committee for International Association of Geodesy (CNC-IAG) organized dozens of domestic institutions to compile the “2019—2023 China National Report on Geodesy”, which summarized the research progress of various branches of geodesy in China from 2019 to 2023. This article summarizes the overall progress of China's geodetic discipline in recent years, focusing on representative progress in six research directions including reference frame, comprehensive PNT and resilient PNT, gravity field and vertical datum, precise GNSS products, multi-source sensor integrated navigation, and marine geodesy. Moreover, in light of the development trends of international geodesy and related interdisciplinary disciplines, several suggestions are proposed for the future development of geodesy in China. Reference | Related Articles | Metrics

Select

PolSAR image registration combining polarization whitening filtering and SimSD-CapsuleNet XIANG Deliang, DING Huaiyue, GUAN Dongdong, CHENG Jianda, SUN Xiaokun Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 450-462.   DOI: 10.11947/j.AGCS.2024.20230197 Abstract （62）   HTML （3）    PDF（pc） （21773KB）（64）       Knowledge map    Save Polarimetric synthetic aperture radar (PolSAR) image registration has a wide range of applications in feature classification, change detection, and image fusion. Existing PolSAR image registration methods, whether based on deep learning or conventional methods, use PolSAR magnitude image information for processing. This processing leads to a large amount of polarization information loss, and at the same time, the registration accuracy and reliability perform poorly under the influence of the inherent coherent speckle noise of PolSAR images. To this end, this paper first develops a novel and effective key point detector based on polarization whitening filter (PWF) refinement processing, which uses PWF to suppress coherent speckle noise in PolSAR images and selects significant and uniformly distributed matching key points by threshold constraint, morphological erosion, and non-extreme value suppression. Further, in this paper, we design a Siamese simple dense capsule network (SimSD-CapsuleNet) to quickly extract the shallow texture features and deep semantic features of the data, and we use the polarization covariance matrix as the input data in order to make full use of the polarization information. In this paper, the distances between the capsule form feature descriptors are calculated and fed into a hard L2 loss function for the training of the model. The method in this paper is validated on PolSAR images acquired by different sensors with different resolutions, and the results show that the method can acquire more uniform and a larger number of matching key points in a shorter time, and the combination of PWF and deep neural network can achieve fast and accurate PolSAR image registration. Reference | Related Articles | Metrics

Select

A self-supervised pre-training scheme for multi-source heterogeneous remote sensing image land cover classification XUE Zhixiang, YU Xuchu, LIU Jingzheng, YANG Guopeng, LIU Bing, YU Anzhu, ZHOU Jianan, JIN Shanghong Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 512-525.   DOI: 10.11947/j.AGCS.2024.20220553 Abstract （72）   HTML （7）    PDF（pc） （18983KB）（63）       Knowledge map    Save Deep learning has revolutionized the remote sensing image processing techniques over the past few years. Nevertheless, it is laborious to annotate high quality samples, thus limiting the performance of deep networks because of insufficient supervision information. To resolve this contradiction, we investigate the self-supervised pre-training and fine-tuning paradigm for multi-source heterogeneous remote sensing image land cover classification, aiming to relieve the urgent need for manually annotated data. Specifically, the proposed generative feature learning model consists of asymmetric encoder-decoder structure, in which the deep encoder extracts high-level key characteristics contained in multi-source data and task-specific lightweight decoders are developed to reconstruct original data. To further improve the feature representation capability, the cross-attention layers are utilized to exchange information contained in heterogeneous characteristics, thus learning more complementary information from multi-source remote sensing data. In fine-tuning stage, the trained encoder is employed as unsupervised feature extractor, and learned features are utilized for land cover classification through the designed lightweight Transformer based classifier. This self-supervised pre-training architecture is capable of learning high-level key features from multi-source heterogenous remote sensing images, and this process does not require any labeled information, thus relieving the urgent need for labeled samples. Compared with existing classification paradigms, the proposed multimodal self-supervised pre-training and fine-tuning scheme achieves superior performance for remote sensing image classification. Reference | Related Articles | Metrics

Select

Analysis of interferometric mapping accuracy for spaceborne distributed SAR dual-frequency alternative bistatic mode WANG Yuan, XU Huaping, LI Chunsheng, ZENG Guobing, LIU Aifang, GE Shiqi Acta Geodaetica et Cartographica Sinica    2024, 53 (3): 463-472.   DOI: 10.11947/j.AGCS.2024.20220657 Abstract （56）   HTML （3）    PDF（pc） （7726KB）（62）       Knowledge map    Save Due to flexible baseline and minimal temporal decorrelation, spaceborne distributed synthetic aperture radar (SAR) systems have been rapidly developed and widely applied in interferometric mapping. Dual-frequency alternative bistatic mode is an emerging mode, it can simultaneously acquire multiple SAR images at two frequencies compared with the conventional mode, which enables higher accuracy of interferometric mapping. In this paper, the fundamental of dual-frequency alternative bistatic mode is briefly introduced. Then, the phase ratio relationship and coherence characteristics of SAR interferometric pairs in this mode are investigated in depth, and the interferometric phase accuracy and mapping accuracy are analyzed theoretically, which provides a theoretical basis for the design of the SAR systems and the study of the interferometric processing. Finally, simulation results demonstrate the superiority of the dual-frequency alternative bistatic mode over the conventional mode. This mode can increase the interferometric phase data of different frequencies, thereby improving the accuracy of interferometric mapping. Reference | Related Articles | Metrics