Listed by UNESCO in 1987 as a World Heritage site,the world-famous Ming Great Wall stretches several thousands of kilometers across northern China,and served as a massive military defensive system which in recent time...Listed by UNESCO in 1987 as a World Heritage site,the world-famous Ming Great Wall stretches several thousands of kilometers across northern China,and served as a massive military defensive system which in recent times has a unique historical,artistic and scientific value.Due to historical reasons and lack of advanced technologies,construction resources and conservation status of Ming Great Wall have not been investigated in any systematic manner;indeed,the extent of the Great Wall has not even been measured.This has resulted in a shortage of reliable first-hand scientific information on actual size,spatially resource distribution and preservation status of this World Heritage site.Driven by the urgent need to establish protection,research,renovation and management of Ming Great Wall,a comprehensive investigation and spatial mapping was jointly organized and completed by the State Bureau of Survey and Mapping and State Administration of Culture Heritage.High resolution digital stereo models at 1:10000 map scale covering the whole length of the Ming Great Wall have been created by photogrammetric reconstruction using nearly ten thousand aerial images.Spatial distribution and attributes of the wall sections,trenches and various subsidiary facilities in the surroundings of the Great Wall were measured with the help of digital photogrammetry workstations and results from field studies.Reliable and precise information about the Ming Great Wall has now been obtained and documented,including surface lengths,resource distribution,and preservation status.For example,the total length of Ming Great Wall is 8851.8 km,of which 6259.6 km is of actual wall,2232.5 km of natural terrain,and 359.7 km of trenches.In category lengths,1828.8 km is constructed of stone,3411.3 km of earth,249.6 km in brick,197.5 km of cliff wall and the rest 572.4 km of other means.Such information provides the scientific basis and strong platform in helping to delineate areas needing protection,in planning conservation and renovation programs,as well as digital archiving for posterity and web-based applications for modern promotions of one of the world's great attractions,the Ming Great Wall.展开更多
In recent years, there have been a lot of interests in incorporating semantics into simultaneous localization and mapping （SLAM） systems. This paper presents an approach to generate an outdoor large-scale 3D dense s...In recent years, there have been a lot of interests in incorporating semantics into simultaneous localization and mapping （SLAM） systems. This paper presents an approach to generate an outdoor large-scale 3D dense semantic map based on binocular stereo vision. The inputs to system are stereo color images from a moving vehicle. First, dense 3D space around the vehicle is constructed, and tile motion of camera is estimated by visual odometry. Meanwhile, semantic segmentation is performed through the deep learning technology online, and the semantic labels are also used to verify tim feature matching in visual odometry. These three processes calculate the motion, depth and semantic label of every pixel in the input views. Then, a voxel conditional random field （CRF） inference is introduced to fuse semantic labels to voxel. After that, we present a method to remove the moving objects by incorporating the semantic labels, which improves the motion segmentation accuracy. The last is to generate tile dense 3D semantic map of an urban environment from arbitrary long image sequence. We evaluate our approach on KITTI vision benchmark, and the results show that the proposed method is effective.展开更多
In the paper, an approach is proposed for the problem of consistency in depth maps estimation from binocular stereo video sequence. The consistent method includes temporal consistency and spatial consistency to elimin...In the paper, an approach is proposed for the problem of consistency in depth maps estimation from binocular stereo video sequence. The consistent method includes temporal consistency and spatial consistency to eliminate the flickering artifacts and smooth inaccuracy in depth recovery. So the improved global stereo matching based on graph cut and energy optimization is implemented. In temporal domain, the penalty function with coherence factor is introduced for temporal consistency, and the factor is determined by Lucas-Kanade optical flow weighted histogram similarity constraint(LKWHSC). In spatial domain, the joint bilateral truncated absolute difference(JBTAD) is proposed for segmentation smoothing. The method can smooth naturally and uniformly in low-gradient region and avoid over-smoothing as well as keep edge sharpness in high-gradient discontinuities to realize spatial consistency. The experimental results show that the algorithm can obtain better spatial and temporal consistent depth maps compared with the existing algorithms.展开更多
Block Adjustment(BA)is one of the essential techniques for producing high-precision geospatial 3D data products with optical stereo satellite imagery.For block adjustment with few ground-control points or without grou...Block Adjustment(BA)is one of the essential techniques for producing high-precision geospatial 3D data products with optical stereo satellite imagery.For block adjustment with few ground-control points or without ground control,the vertical error of the model is the decisive factor that constrains the accuracy of 3D data products.The elevation data obtained by spaceborne laser altimeter have the advantages of short update periods,high positioning precision,and low acquisition cost,providing sufficient data support for improving the elevation accuracy of stereo models through the combined BA.This paper proposes a geometric positioning model based on the integration of Optical Satellite Stereo Imagery(OSSI)and spaceborne laser altimeter data.Firstly,we elaborate the principle and necessity of this work through a literature review of existing methods.Then,the framework of our geo-positioning models.Secondly,four key technologies of the proposed model are expounded in order,including the acquisition and management of global Laser Control Points,the association of LCPs and OSSI,the block adjustment model combining LCPs with OSSI,and the accuracy estimation and quality control of the combined BA.Next,the combined BA experiment using Ziyuan-3(ZY-3)OSSI and ICESat-2 laser data was carried out at the testing site in Shandong Province,China.Experimental results prove that our method can automatically select LCPs with high accuracy.The elevation deviation of the combined BA eventually achieved the Mean Error(ME)of 0.06 m and the Root Mean Square Error(RMSE)of 1.18 m,much lower than the ME of 13.20 m and the RMSE of 3.88 m before the block adjustment.A further research direction will be how to perform more adequate accuracy analysis and quality control using massive laser points as checkpoints.展开更多
文摘Listed by UNESCO in 1987 as a World Heritage site,the world-famous Ming Great Wall stretches several thousands of kilometers across northern China,and served as a massive military defensive system which in recent times has a unique historical,artistic and scientific value.Due to historical reasons and lack of advanced technologies,construction resources and conservation status of Ming Great Wall have not been investigated in any systematic manner;indeed,the extent of the Great Wall has not even been measured.This has resulted in a shortage of reliable first-hand scientific information on actual size,spatially resource distribution and preservation status of this World Heritage site.Driven by the urgent need to establish protection,research,renovation and management of Ming Great Wall,a comprehensive investigation and spatial mapping was jointly organized and completed by the State Bureau of Survey and Mapping and State Administration of Culture Heritage.High resolution digital stereo models at 1:10000 map scale covering the whole length of the Ming Great Wall have been created by photogrammetric reconstruction using nearly ten thousand aerial images.Spatial distribution and attributes of the wall sections,trenches and various subsidiary facilities in the surroundings of the Great Wall were measured with the help of digital photogrammetry workstations and results from field studies.Reliable and precise information about the Ming Great Wall has now been obtained and documented,including surface lengths,resource distribution,and preservation status.For example,the total length of Ming Great Wall is 8851.8 km,of which 6259.6 km is of actual wall,2232.5 km of natural terrain,and 359.7 km of trenches.In category lengths,1828.8 km is constructed of stone,3411.3 km of earth,249.6 km in brick,197.5 km of cliff wall and the rest 572.4 km of other means.Such information provides the scientific basis and strong platform in helping to delineate areas needing protection,in planning conservation and renovation programs,as well as digital archiving for posterity and web-based applications for modern promotions of one of the world's great attractions,the Ming Great Wall.
基金supported by National Natural Science Foundation of China(Nos.NSFC 61473042 and 61105092)Beijing Higher Education Young Elite Teacher Project(No.YETP1215)
文摘In recent years, there have been a lot of interests in incorporating semantics into simultaneous localization and mapping （SLAM） systems. This paper presents an approach to generate an outdoor large-scale 3D dense semantic map based on binocular stereo vision. The inputs to system are stereo color images from a moving vehicle. First, dense 3D space around the vehicle is constructed, and tile motion of camera is estimated by visual odometry. Meanwhile, semantic segmentation is performed through the deep learning technology online, and the semantic labels are also used to verify tim feature matching in visual odometry. These three processes calculate the motion, depth and semantic label of every pixel in the input views. Then, a voxel conditional random field （CRF） inference is introduced to fuse semantic labels to voxel. After that, we present a method to remove the moving objects by incorporating the semantic labels, which improves the motion segmentation accuracy. The last is to generate tile dense 3D semantic map of an urban environment from arbitrary long image sequence. We evaluate our approach on KITTI vision benchmark, and the results show that the proposed method is effective.
基金the Science and Technology Innovation Project of Ministry of Culture of China(No.2014KJCXXM08)the National Key Technology Research and Development Program of the Ministry of Science and Technology of China(No.2012BAH37F02)the National High Technology Research and Development Program(863)of China(No.2011AA01A107)
文摘In the paper, an approach is proposed for the problem of consistency in depth maps estimation from binocular stereo video sequence. The consistent method includes temporal consistency and spatial consistency to eliminate the flickering artifacts and smooth inaccuracy in depth recovery. So the improved global stereo matching based on graph cut and energy optimization is implemented. In temporal domain, the penalty function with coherence factor is introduced for temporal consistency, and the factor is determined by Lucas-Kanade optical flow weighted histogram similarity constraint(LKWHSC). In spatial domain, the joint bilateral truncated absolute difference(JBTAD) is proposed for segmentation smoothing. The method can smooth naturally and uniformly in low-gradient region and avoid over-smoothing as well as keep edge sharpness in high-gradient discontinuities to realize spatial consistency. The experimental results show that the algorithm can obtain better spatial and temporal consistent depth maps compared with the existing algorithms.
基金supported by the National Science Fund for Distinguished Young Scholars[grant number 61825103]the Fundamental Research Funds for The Central Universities[grant number 2042022kf1002].
文摘Block Adjustment(BA)is one of the essential techniques for producing high-precision geospatial 3D data products with optical stereo satellite imagery.For block adjustment with few ground-control points or without ground control,the vertical error of the model is the decisive factor that constrains the accuracy of 3D data products.The elevation data obtained by spaceborne laser altimeter have the advantages of short update periods,high positioning precision,and low acquisition cost,providing sufficient data support for improving the elevation accuracy of stereo models through the combined BA.This paper proposes a geometric positioning model based on the integration of Optical Satellite Stereo Imagery(OSSI)and spaceborne laser altimeter data.Firstly,we elaborate the principle and necessity of this work through a literature review of existing methods.Then,the framework of our geo-positioning models.Secondly,four key technologies of the proposed model are expounded in order,including the acquisition and management of global Laser Control Points,the association of LCPs and OSSI,the block adjustment model combining LCPs with OSSI,and the accuracy estimation and quality control of the combined BA.Next,the combined BA experiment using Ziyuan-3(ZY-3)OSSI and ICESat-2 laser data was carried out at the testing site in Shandong Province,China.Experimental results prove that our method can automatically select LCPs with high accuracy.The elevation deviation of the combined BA eventually achieved the Mean Error(ME)of 0.06 m and the Root Mean Square Error(RMSE)of 1.18 m,much lower than the ME of 13.20 m and the RMSE of 3.88 m before the block adjustment.A further research direction will be how to perform more adequate accuracy analysis and quality control using massive laser points as checkpoints.
