2010 Vol. 35, No. 8
We propose a new method of the calculation of the semantic similarity of land use types based on ontology,which improves the generalization methodology of land use data by introducing ontology-based semantic similarity into polygon amalgamation.Ontology-based semantic similarity integrates human spatial cognition.The improved methodology of land use data generalization can reflect and generalize the characters of land use better.It was verified by a case study.
The resulted data handing and database construction of the second land resource survey includes the land-use data generalization and map design of different administrative levels.From the perspective of practical works,we explore a technical method of land-use data generalization,including the detection of polygons with closest relationship from neighborhood candidates,the establishment of the generalization rules,the aggregation and amalgamation methods for polygon combination and the evaluation of the generalization results.
Lines do not represent the routes completely and effectiveness based on arc-node model.In this paper,a new model generalization of schematic network maps based on dynamic segment is proposed.Routes are constructed according to the road segment attribute similarity and geometry continuous.Based on route networks,a new method comprising shapes simplification and displacement is proposed to schematize road network quickly.A case study for schematic map was carried out and analyzed.Initial experimental studies show that this approach is more effective in evidence on making features maximum clearly on screen.
Building with parallel structures is the most popular kind in modern city.As the widely used 3D city model,more and more generalization methods for this kind of building have been developed.This paper begins with the structuring of 3D building data and then a data structure that benefits the generalization is proposed.Orthogonal of building,introduction of polygon and acquisition of feature information are used in building construction.According to the requirement of visualizations,LODs of city buildings are divided into three levels.Based on which,a simplification algorithm of shifting facet based on minimal distance is realized in the simplification of the building with parallel structures.Experiment was done to prove the good efficiency and usability of the algorithm and the algorithm works well in simplification of buildings with parallel structures.
This paper derives a rapid algorithm that enables the determination of the average node degrees of a vehicular ad hoc network(VANET) in highway scenarios.Extensive experiments were undertaken to verify the derived analytic expression via realistic mobility traces.Results demonstrate that the mean degree increases linearly with respect to traffic density and radio transmission range.
In this paper,the cloud genetic algorithm(CGA) was introduced by the combination of cloud model theory and genetic algorithms(GA).The Y-conditional cloud generator for the normal cloud model is used as cross operation in this hybrid genetic algorithm,and the basic cloud generator is used as the mutation operator in the optimization operation of GA.Both the cross and the mutation operation make use of the randomness and stability of cloud to improve the algorithm convergence,robustness and the solutions quality.And also it overcomes the traditional GA shortcomings such as slow searching,easy to local optimization solutions.The simulation test shows that the proposed hybrid algorithm can improve the vehicular scheduling efficiency and feasibility for the public transport.The scheduling solutions from this algorithm were also validated through the service quality evaluation and show well application prospect.
An efficient algorithm for generating Voronoi diagram of 3D scattered point-set is proposed,which takes the data structure named "point-face-polyhedron" to store Voronoi cell,initializes the Voronoi cell of the first point,and the Voronoi cell of the new inserted point was generated after whose adjacent cells were changed accordingly through fission and reconstitution of cells.The experimental results show that this method can generate Voronoi diagram of 3D scattered point-set veraciously and efficiently.The topological neighbors inquiring for 3D scattered point-set can be realized efficiently based on this algorithm.
A hierarchical approach is proposed to match the corresponding nodes from two different datasets.In this approach,node types are first of all distinguished and described,and the limitations of existing description methods are pointed out.On this basis,structural properties of node and its linked segments are considered to build the three geometric constraints,including distance,topology,and direction.They are further used to filter the candidates for matching reference nodes.Indeed,this process is of a hierarchy.Hence,it is also called a hierarchical approach.An experiment is employed to illustrate the rationality of the approach proposed in this paper.
This paper presents a public and light interpretation to an old method,and uses a new name—rendering and pickup,proposes the procedure detail and key sentences about the method implementation,discusses some techniques deeply about the method in correctness,precision,limited for memory bitmap,partition rasterization and time-spend.The authors think that the old method could become a generally industrial method.
In the application of video retrieval by sample,video data is described by multiple high-dimensional features,and the weights of these features are changed in different queries.We propose a new indexing structure called multi-feature index tree(MFI-Tree) to index multiple high-dimensional features of video data for this retrieval application.MFI-Tree employs tree structure which is benefit for browsing application,and travels the last level aggregate node in retrieval application to improve the performance.And more,aggressive decided distance for kNN search algorithm which fast reduces the distance to prune the search space more effectively is proposed.The experimental results show that MFI-Tree and ADD-kNN algorithm offer performance advantages over sequential scan.
This paper focuses on the design and development of a collaborative air pollution simulation system,especially for the data collaboration,modeling collaboration,visualization collaboration and analysis collaboration.Based on the design of the collaborations,a prototype system was developed to study sulfur dioxide dispersion within area of Pearl River Delta.The experimental result shows that the collaborations is smooth and system is efficient.
This paper explores approaches for Toponym resolution in Chinese text,and proposes a geo-parsing approach based on conditional random fields and discourse toponym relations,and a geo-coding approach based on partial fuzzy matching and cognitive salience calculation.The proposed geo-parsing approach deals with the recognition of toponym in three major steps.The experiment shows that the key factors that may influence the performance of toponym resolution in Chinese text are the coverage of gazetteer,the performance of geo-parsing and the performance of semantic disambiguation of toponyms.In our experiment,there are about 17% toponyms can not locate their semantics in the gazetteer.Ambiguity in geo-parsing and geo-coding are the next prominent factors that affect the overall performance of toponym resolution.
A new approach of automatic water body extraction based on LiDAR data and aerial images is proposed.The LiDAR intensity image and aerial images are co-registrated by SIFT operator.Transformation parameters from LiDAR image to aerial image can be computed by the matched conjugated points.Black regions that have no reflection on height image generated by LiDAR data are automatically extracted.Geometric constraint conditions are used to remove non-water body areas.Edge information is used for region growing with the projected inilial position as initial value.Finally,mathematical morphology techniques are used to get more precise water body areas.Experimental results show that the proposed approach can achieve very good water body information.
A new unsupervised classification method of fully polarimetric SAR data based on the PSO algorithm and H/α/A is presented.Firstly,the result obtained by H/α/A classification is used to initialize the clustering centers,and then the fully polarimetric SAR data is classified by the PSO algorithm.Meantime,the Wishart distance measure is employed to improve the performance of the PSO-H/α/A classifier.Experimental results show that the new scheme proposed in this paper can effectively classify the fully polarimetric SAR data.
An effective method for high resolution synthetic aperture radar(SAR) images feature extraction and target recognition is presented.According to the different response characters of targets to multi-scales Gabor filters,multi-scale information and the pertinence between multi-scales are adopted to extract new multi-scale features.This new feature combines target's macroscopical and microcosmical characteristic together,that makes it easy to obtain more essential features of SAR image.Nonlinear transform that
Traditional target's optimal polarization theory was based on the scattering matrix.But the scattering matrix is usually difficult to obtained.Thus,the calculating process always is complex.In this paper,we deduce the formulae optimal receive power based on Stokes matrix and polarization synthesis.The algorithm could easily be carried out and the programming process is much directly.The experimental results show that the ideal results could be obtained by the proposed algorithm.
We propose a Gabor filter and Markov random fields(MRF)-based method for color texture image segmentation.First,we analyze color and texture feature,transforme RGB space to LUV space to get color feature vector,and then do Gabor filtering and Gaussian smoothing processing on original color image and MRF model is used to represent the regional relationship.Finally,we combine color and texture information and use Bayesian method to estimate maximum a Posteriori(MAP).The experimental results show that this algorithm is efficiently doing color texture image segmentation.
Based on the 6S radiation transfer model,the influences of aerosol optical depth,solar zenith,sensor zenith and ground altitude changes on atmospheric corrected reflectances were calculated.The 6S model's sensitivities to these four parameters were discussed.A new correction algorithm based on look-up table was introduced.The look-up tables of correction parameters for different aerosol optical depth,solar zenith,sensor zenith and ground altitude values were established by the 6S model offline computation.Each pixel of MODIS data was corrected by interpolation from the look-up tables to get surface reflectance.The comparison results between our method,online method and parameter-uniform method show that the result of our method was closed to that of online method,which indicates that our method effectively reduces the influence of parameters spatial heterogeneity on atmospheric correction.
We present a background image building method to deal with the problem of color images motion object detection in real-time.The distance between color images is defined to compare the frame image and the background image according to the frame sequences.The motion state matrix is extracted to update background image in pixels level.The method need not background models and fit variety changes.Actual color images were used to test the method.The experimental results show that our method is effective and accurate.
A non-cooperative game model for QoS-aware optimal composition for multi-task is proposed to analyze services optimal composition strategies of tasks in competition situation,the competition relationships among concurrency tasks are described,and defines the best reply functions for QoS-aware optimal composition of remote sensing information services.Theoretical analysis shows that the proposed non-cooperative game model can reduce the conflicts among concurrency tasks effectively and maximize expectation utility of all tasks.
Through comparising the advantages and disadvantages of sensor-related modeling standards: IEEE 1451,ECHONET and SensorML,a general remote-sensing satellite sensor modeling approach based on SensorML and ISO 19130 is proposed.The procedures of the proposed approach include sensor parameters analysis,dictionary definition,system modeling,process modeling,process chain modeling,model validation and model rectification.The modeling tool,namely SensorModel,is designed and implemented using.NET framework based on the proposed approach.The SPOT-5 satellite platform modeling was used to test the feasibilities of the proposed approach.
A novel forward and backward diffusion based image denoising and enhancement method is proposed.In the proposed method,the direct gradient computation is replaced by a nonlinear structure tensor(NLST) to increase the robustness to noise;The shock filter item along the gradient direction is replaced by a forward or backward diffusion according with the adaptive threshold value of gradient;The corner structure is also enhanced by performing a backward diffusion in corner region.Two experimental results show that the proposed algorithm not only removes noise,but also efficiently enhances image texture details and no artifacts are produced.
We present a cluster matching method of artificial targets in close range photogrammetry based on the known points and epipolar plane constraint.Firstly,all the images are grouped according to the known points such as points on the AutoBar and codes.Secondly,image triplets per known point and their geometrical quality are calculated.Finally,parts of the image triplets with best geometric goodness are selected to match following the epipolar plane constraint.The experimental results show that more than 95% image points of about 10 000 in round 100 pieces of images are matched in less than 5 s,with matching error rates less than 0.1‰.
Based on predicted residual analysis and estimation of filtering errors,a new Kalman filter approach to correct model errors is proposed.Considering the expression of predicted residual,there is no need to discuss whether model errors are from measurement model or dynamic model.Based on the estimation of filtering errors,we just need to correct the measurement model to clear up the whole effects of model errors on the filtering value at current time.A GPS kinematic navigation calculative experiment is given to prove the feasibility of this method.It not only eliminates the whole effect of Kalman filtering model errors efficiently,but also has a good result.
The EOP parameters are solved out by using the data of 2008 from IVS,and the precision of the internal and external aspects are discussed respectively.Then,the data of CONT08(continuous VLBI 2008) is analyzed,and the results were compared with the EOP parameters from IERS and IVS.
Based on time-varying requirements of prediction model for dam monitoring data,the forgetting factor is introduced to set up a forgotten matrix to give prominence to the contributions of recent data.Then the IWRLS algorithm is made to achieve updating model parameters at real-time.On this basis,in order to reflect the physical meaning and complete the filtering operation at the same time,a statistical model and ARMA are introduced into the Kalman filter equations.In the equations,state equation is established by self-variable which reflects the state characteristics with ARMA,and observation equation is established by dependent variable which reflects physical meaning with statistical models.So considering the white noise,the time-varying Kalman prediction model is established with the comprehensive functions.Case analysis shows that the fitting and forecast accuracy of time-varying Kalman model are superior to those traditional statistical models.
A multi-scale decomposition and reconstruction architecture is defined on the basis of the empirical mode decomposition(EMD) theory,and the systematic errors mitigation model is demonstrated.A standard of the scale selection for the systematic errors elimination is given in terms of the means of the accumulated standardized modes.Then,the scheme of the GPS/PLs baseline solution based on the EMD is suggested.The experimental results show that the proposed scheme dramatically improves the reliability of ambiguity resolution and the precision of baseline vector after systematic error eliminated.
The reference epoch based relative clock was estimated using the method of epoch-difference and satellite-difference.An experiment is carried out using the data of different sampling rates collected from Hong Kong and Shanghai GNSS reference networks,in order to study the influences of region and sampling rates to the accuracy.The results show that the difference between our results and IGS final precise clock error products is less than 0.2 ns,and the accuracy of the reference epoch based relative clock is less than 0.9 ns.The accuracy is at the same level in different regions,and the influence of sampling rate is very small.
At 2002,Japan decided to establish the quasi-zenith satellite system(QZSS).The system will supply advanced methods for the positioning and navigation and for the next-generation mobile communication.The construction of the system will be accomplished at 2011.The establishment background and the functions of the system,the roles of the system in GPS augmentation,and the time tasks in the system are analyzed.