2010 Vol. 35, No. 1
Combining prior information with observing information,Bayesian methods for blunder detection are imposed.Especially a lot of effective measures are used to overcome the masking and swamping.When multiple blunder influence each other,the Bayesian method for blunder positioning based on the posterior probabilities of classification variables sometimes gives birth to masking and swamping which leads to the failure of positioning blunder.Hence,on the basis of seeking the reason of masking and swamping,and analyzing the eigenstructure of sampling correlation matrix of classification variables,the Bayesian unmasking method for positioning multiple blunder is introduced.The corresponding algorithm-adaptable MCMC sampling algorithm is implemented.
This paper emphasizes on the modeling of fuzzy neural network(FNN),proproses three modeling styles,namely,single point modeling,forecasting modeling,and prediction modeling,and provides new idea for deformation analysis and prediction.With landslide deformation instance,it points out the feasibilities for engineering deformation analysis and prediction with FNN.
A combined PPP(Precise Point Positioning) mode is implemented using GPS and GLONASS observations.The performance of the combined GPS and GLONASS PPP was assessed using observations from one of the IGS tracking network(IRKJ) and the currently available precise orbit and clock products.The corresponding positioning accuracy and convergence time were compared between GPS-only and GPS/GLONASS combined precise point positioning.The results show that GLONASS has caused a slight impact on positioning accuracy and convergence time in case of more than 8~9 GPS satellites are tracked.However,when the number of available GPS satellites is not enough,especially is less than 4~5,adding a couple of GLONASS satellites can improve positioning accuracy and shorten convergence time significantly.
Presently The algorithm of two-way satellite time transfer is one of the time synchronization methods with the highest accuracy.On the basis of two-way satellite time transfer algorithm,the characteristics of the propagation delay of two-way time transfer signals between moving satellite varying with intersatellite range were analyzed,and the rule of satellite clock offset varying with intersatellite range obtained with this algorithm was deduced.This paper presents a dynamic two-way time transfer algorithm with which high-accuracy intersatellite clock offset is solved through the combination of intersatellite pseudo-range polynomial fitting and clock-offset polynomial fitting.Simulation data of actual satellites showed that with the algorithm the intersatellite time synchronization error can be controlled within 5ns.As a result the algorithm can be used for intersatellite high-accuracy time synchronization.
In the areas such as Japan and eastern and middle part of China,the performance augmentations of QZSS to GPS on navigation precision,availability and integrity are simulated and compared.The results show that QZSS consist of only three satellites can provide effective augmentation to GPS in above fields.With the mask angle increasing,the performance improvements will become more obvious.The results also show that QZSS can effectively reduce the performance deterioration of GPS in the wartime,because of results from the invalidation of part satellites.The analysis results can be used as guidance for the effective application of GNSS resources in Chinese middle part and eastern key area.
According to the deficiency of the LLL algorithm(A.K.Lenstra,H.W.Lenstra,L.Lovasz algorithm),the adaptive integer orthogonal transformation algorithm is proposed in the paper.It can be used to improve the LLL algorithm with sort ascending matrix.Based on the condition number,comparison has been made between the LLL algorithm and the improved LLL method by using 600 symmetric and positive definite matrixes derived from random simulation,and as 30 sets of actual measuring data.Numerical results show that the improved LLL algorithm has better performance in decreasing matrix condition number and the number of candidate integer ambiguity.Thus it could decrease the searching time of the right integer ambiguity resolution.
A Ms 8.0 earthquake struck Wenchuan on 12 May 2008 and anomaly of observation curve of fixed deformation stations is little.After the earthquake observation data before earthquake have been analyzed and processed.The result shows that deformation reveals anomalies before Wenchuan Ms 8.0 earthquake.The frequency of anomalies are high when stations are near from the earthquake epicenter and the frequency are low when stations are far from the earthquake epicenter.Finally the phenomena have been primary explained.
The error of high-accuracy inertial navigation system(INS) caused by the gravity disturbance can't be neglected.Starting from the physical geodesy and Newton's second law,an error dynamics equation of INS,including the influence of gavity disturbance,is firstly presented in this paper.Then,taking single-axis INS as an example,the position error caused by the deflection of vertical and corresponding characteristics of error propagation in three situation were analyzed.Finally,simulation was done on 1′×1′ deflection of vertical database,and from the simulation result we can see that the horizontal error of INS caused by deflections of the vertical on the sailing course can reach as large as 3 km.
The finite element model of Three Gorges area was built based on DEM and the actual GPS stations' coordinates.The crustal vertically deformation of the Three Gorges reservoir area was calculated under the condition of 135 meter,156 meter and 175 meter water level impoundment by means of the elasticity finite element analysis method respectively.Based on the Farrell's globosity deformation theory,the vertically deformation fields of the Three Gorges reservoir area was simulated under the condition of 135 meter water level impoundment.By comparing the two simulated fields,the mechanism of crustal vertically deformation are analyzed.The deformation isoline covered the valley line with the maximum in Xiangxi segment,and the regional sedimentation is synthetically geological function and the water body gravity as the main reason.
The spatial distribution of road networks is a fundamental issue in spatial analysis of GIS for Transportation.Betweenness Centrality is extensively used to evaluate the importance of nodes for the analysis of complicated networks.In this paper,six typical urban road networks were selected and transformed to direct graphs for the analysis of betweenness centrality.The results show that the distribution of betweenness centrality shows consistency and has a hierarchical property.The statistical results of betweenness centrality illustrates that the majority of road segments in high administrative levels characterize high value while the majority of road segments in low administrative levels have low value in urban road networks.The experimental results show that there is a correlation between the Betweenness Centrality hierarchy in mathematical measure and the administrative levels.
According to the problem that different conditions of traffic flow of different turns around crossings are neglected and uniform traffic conditions on a link is supposed on link dividing methods for traffic information collecting based on GPS equipped floating car at present,which leads to traffic data with low quality and poor effect of the road traffic conditions identification system and vehicle dynamic navigation system.The directional link dividing method can be used to collect traffic data distinguishing different turns around crossings and sub-link dividing method can be used to collect traffic data distinguishing different places of a link are designed,which improves application effect of the road traffic conditions identification systems and vehicle dynamic navigation systems.
The random uncertainty and fuzziness uncertainty exist in spatial data at the same time.Ant that it is not suitable for using hybrid entropy to deal with the uncertainty of some geographical objects which contain random uncertainty and fuzziness uncertainty.The uncertainty of fuzzy geographical object was discussed by using type-2 fuzzy theory.The error model of membership of vague geographical object was built by using λ-cut of fuzzy set and error band of circle based on positional erros.The measure method of the error of membership was analyzed,Finally,the membership grade error cube was proposed to synthetically describe the uncertainty of fuzzy point.
The progressive transmission of vector map data requires an efficient multi-scale data model to process the data into a hierarchical structure.This paper presents such a data structure of river network without redundancy of geometry for progressive transmission.For a given scale,the river network display has to settle two questions.One is which river objects to be selected and the other is what detail to be visualized for the selected rivers.This study combines the T pfer law and the BLG-tree structure to answer the above two questions simultaneously.At the level of object element,the river branches are sorted on descending significance grade decided by watershed area to support the river selection by the T pfer law.At the level of geometric detail,the river branch is splitted into segments by joint points with the organization of the linear BLG-tree to export a good graphic representation at a given scale.Based on the data structure,a WebGIS is established to provide progressive transmission services of river networks.
The generalization of contour lines is an active issue in the field of multi-scale representation of terrains.This paper considers the contour lines as the 3D segments constructed from elevation points which are sorted in a special rule,and then selects the significant feature points from those spatial points in 3D space using the 3D D-P algorithm.Based on the spatial distribution characteristic of contour line data ses,several methods are designed and given comments on their results.It is concluded that the generalization of contour lines using 3D D-P algorithm is a promising way to meet the cartographic generalization according to its quality and speed in the generalization.
The raster terrains can be deformed to sparse networks and dense networks according to different neighborhood structures.The least-cost path problems in sparse networks are analyzed,modeled and resolved using the Dijkstra algorithm.Simultaneously the ones in dense networks are analyzed,modeled and resolved by simulated annealing algorithm.The accuracy and efficiency of the results got from the experiments of the two methods were compared and analyzed thoroughly.And the Dijkstra algorithm is proved to be more suitable for least-cost path planning in raster terrains.
Considering many geographical spatial objects and influence factors in the process of loss-of-life evaluation caused by dam breach,an evaluation method of loss-of-life caused by dam breach based on GIS and neural networks optimized by genetic algorithms is proposed.The calculation model for estimating loss-of-life caused by dam breach is established based on GIS spatial information grid model.The influence factors of loss-of-life caused by dam breach were analyzed.The evaluation index system is established by grey relation degree model.A fast evaluation system of loss-of-life caused by dam breach based on GIS is implemented.Scheme and GIS technologies for the system were analyzed.Application shows accuracies and good effects of the evaluated method.
This paper studies dynamic composition and reconstruction of geographical information services through programming the geographical information services composition as the optimization problem upon on the discrete and limited data structure.This model defines and describes the semantic information of the geographical information services quality including the times,expenses,availabilities and so on.The paper acquires the weight of the geographical information service quality and evaluation function.Then this model uses these parameters to process the genetic operations and requires the optimal composition project.Finally,the paper analyzes and tests the services dynamic composition model from the composition times and composition success rates.The result inclidates that the services dynamic composition model is feasible.
Spatial relationship reasoning is an importance topic in the spatial relation studies.The mixed spatial relation qualitative reasoning between topological relation and directional relation in three dimensional space is studies in the paper.The directional regions based on the projection system and the topological relation based on the nine intersection matrix model are described with Allen's interval relation pairs.The presented method is used to make the mixed spatial reasoning between topological relation and directional relation,and reasoning results were shown by the composition tables.
Firstly,this paper defines the concept of pseudo 3D visualizations in mobile maps.Secondly,the paper proposes a pseudo visualization model and designs a rapid 3D projection transformation algorithm.Thirdly,in order to show a realistic 3D scene with 2D data,a kind of method of pseudo 3D pre-processing is proposed and a necessary enhanced 3D technology is adopted.Finally,pseudo 3D visualization from 2D vector data was implemented for mobile maps.Live pseudo 3D maps are produced and shown on mobile devices at real time.The experimental results are given in the paper.
In this paper,language features of spatial information in web texts are firstly analyzed.Then essential methods about online extraction and matching of spatial information and are studied.On the basis of GIS spatial databases,a framework and several methods about expanding the abilities of spatial information query are proposed by keeping data acquisition method in an open style.Related case studies were performed to verify the correctness of the methods.
The slant-range imaging principle of space-borne SAR sensor is discussed,and from the Range-Doppler equations,the direct georeferencing model without ground control points(GCPs) is formulated.In the conditions of the lack of GCPs,the model for adjusting the model parameters,including the track model parameters and image geometric processing parameters,is derived in detail.Direct georeferencing is made on a scene of ENVISAT ASAR imagery of Beijing area,and planimetric accuracy of 170.966 m(±8.7 pixels) is obtained.After using one GCP to adjust some model parameters,the planimetric accuracy is improved by ±54.665 m(±2.8 pixels).The experiment shows that in the condition of the lack of GCPs the positioning accuracy for space-borne SAR imagery can approximate to two ground sampled distance,and will has a good application prospects,especially for the georeferencing of the difficult areas.
Flat-phase is the linear phase which is caused by flat ground in the interferometric processing of SAR.It increases the difficulties of phase unwrapping,and its removal accuracy also affects the processing results of InSAR.So flat-phase removal is one of the critical steps of InSAR processing.Generally,flat-phase can be calculated accurately if precise orbit state vectors are available.However,the orbit sampling interval of ALOS satellite is 60-second.Generating the flat phase exactly is one of the difficulties in interferometric processing for ALOS/PALSAR data,which also limits the InSAR application of PALSAR data.In this paper,by analyzing different methods for estimating orbit state vectors,we utilize the Hermite interpolation on equidistant nodes and apply the estimated results to remove flat-phase.ALOS/PALSAR data of Changzhou City,Jiangsu Province were used to evaluate the proposed method.Finally,we compared Hermite interpolation on equidistant nodes with cubic spline interpolation and least squares polynomial regression.
Based on a large typically representative survey data(N=335) in Taihu Lake,a significant linear relationship was established between analytically measured total suspended solids(TSS) and MODIS B1-B3 at a resolution of 500 m.However,this paper also demonstrates why MODIS B1-B3 is applied by exploring how to identify the sensitive bands for detecting TSS.Finally,a general and robust regression model was developed to estimate TSS in Taihu Lake.
Image denoising is an important and widely studied problem in machine vision and image processing.However,a large number of image denoising methods eliminate noise and discard textures and edges,at the same time.To overcome the shortcomings,the paper makes its improvement on a basis of unsupervised,information-theoretic,adaptive image filter under analysis of difficulties on image denoising.According to the principles of signal energy,a detection operator of textural features is proposed to check the filter residue of the image in the time-domain and the frequency-domain respectively.Textures and details filtered out by mistake during the process of denoising will be extracted as much as possible.After the filtered image is compensated for missing information,the final denoised image is obtained.Experimental results show that the method can retain the image's textures and details,effectively eliminate image noise,increase the signal-to-noise ratio(SNR),reduce the mean square error,and significantly improve the image's visual effect.Thus,it is practicable.
The uncertainly of artificial mark locations in digital photogrammetry is badly affected by imaging system.Aiming at mark location models with overlapped zero-mean Gaussian white noise and considering the physical process of imaging,a theoretical error performance Cramér-Rao lower bound based on the uncertainty theory of mark location was derived through the theory of maximum likelihood estimation.And its corresponding confidence interval was also determined on a certain confidence level.With the same confidence level,the uncertainty of any practical locations operator can be estimated with the degree of closeness between its own confidence interval and theoretical confidence interval of CRLB.To assess the uncertainty of practical locations operator and design a new more precise mark locations operator,circular mark is taken as an example to illustrate the effects of noise level and mark size on CRLB.The results show that the precision based on CRLB can be achieved 0.015 pixel by choosing suitable mark radius and small noise standard deviation.
In order to deal with the kernel optimization,a new intelligent data-dependent optimizing kernel method is proposed.In this scheme,a new kernel cluster balanced K-means is firstly presented,which can effectively overcome the singularity of matrix for Mahalanobis distance in local Fisher criterion(KO-LKFC).Then,the total objective function based on local Fisher criterion is given.Finally,genetic algorithm is used to find the global optimization and the intelligence is improved.Experimental results based on UCI data and the MSTAR SAR data show that it is effective and feasible.
An approach for extracting outline of residential area through perceptual organization and boundaries of textured regions is presented.In the proposed approach,the concept of free angle is applied to locate boundaries of residential regions.Then multiple clues are integrated into the procedure of curve perceptual organization to obtain close and long boundary.Experimental results show that extracted boundaries of residential area from high resolution remote sensing imagery are accurate and complete.
This paper studies the situation of the lower accuracy of Position and Attitude Measurement Device in InSAR system and puts forward a method of topographic change detection base on terrain matching.This is a new method in topographic change detection with InSAR data.It can improve the degree of automation data processing in topographic change detection with poor Direct Georeferencing of InSAR system.
In order to improve the speed of obtaining terrain data in 3D terrain roaming systems,an Unstructured Peer-to-Peer Terrain Data Storage System(UPTDS) is proposed on the principle of terrain data characteristics and Unstructured Peer-to-Peer System's virtue.Massive terrain data is stored in the peers of the UPTDS.And UPTDS is divided into two layers,data sever layer and data download layer,according to peer's property and function.The two layers are incharge with data searching and supplying data respectively.Compared with the traditional storage system,UPTDS advances the searching efficiency as well as eliminates server performance bottlenecks.