2011 Vol. 36, No. 3
Intelligentization is one of the key themes of current world innovation.We introduce the web of intelligent age: smart sensors and smart sensor web,and discuss the configuration and properties of geoinfo-sensor-web.We also explain the primary intelligence characteristics of GNSS-based continuously operating reference stations(CORS) network and denote the reason that CORS can compose the ground infrastructure of geoinfo-sensor-web.Then we describe the future development and applications of CORS.
Ambiguity decorrelation is a useful technique for higher success rate and reliability of rapid ambiguity resolution.In this contribution integer Gaussian(Gauss),inverse integer Cholesky(LDL) and LLL decorrelation algorithm are studied from real number calculation order and real-to-integer transformation strategy,on which both ascending sort and integer inverse decoration algorithm are presented for enhancing the ambiguity resolution performance.An equivalence correlation coefficient is proposed by analyzing advantages and disadvantages of condition number,decorrelation coefficient and average correlation coefficient employed for decorrelation evaluation.The results show that the equivalence correlation coefficient outperforms three evaluation indexes mentioned above especially for high dimensional covariance matrix.All decorrelation algorithms is improved by ascending sort,and inverse-integer decorrelation algorithms is superior to integer-inverse ones with LDL,Gauss and LLL preferential order.
The elevations from ICESAT were compared with that from GPS RTK and static GPS measurements.For points along the satellite ground tracks,linear interpolation was used to get the heights of GPS points and compared to the geodetic heights,a minimum of 0.76 m is found for the GPS RTK points.For GPS points near the satellite ground tracks,the points along the ground track are selected by the longitude and latitude differences,5″,10″ and 20″ are used respectively.The heights of the GPS points were interpolated using different interpolation methods,such as arithmetic average,spline,etc..At the point where path of GPS RTK intersects with the ICESAT ground track,a minimum height difference of 0.381 m was found.Static GPS measurement about 45 minutes were carried out near the ICESAT ground tracks.And the coordinates were solved using the continuous operating reference stations with an accuracy better than 0.08 m.A minimum height difference of 0.103 m was found between GPS and ICESAT.
TEC is predicted by time series analysis through the ionosphere data sample released by International GPS Service.Time series of the TEC is decomposed into the trend item,periodic term and random item,then they are computed respectively.The number of model order is determined using the correlation coefficient and partial correlation coefficient.Predicted analysis of test is carried out through the data in 2008,the results indicate that the method can achieve high accuracy.The average relative accuracy of 7 d is 87.75%.The predicted value with its accuracy greater than 60% shares more than 95% of the total predicted value,greater than 60% has 72.6% of the total predicted value,and 95.3% of predicted residual is less than ±3 TECU.
We propose a new integer ambiguity resolution method,which enables one to determine DD carrier phase integer ambiguities at single epoch on the fly.Then receiver autonomous integrity monitoring(RAIM) based on carrier phase is researched.Test statistic is constructed using the method of least squares to test the integer ambiguity.Relationship between protection level,satellite geometry and miss detection probability is analyzed.Experiment results indicate that the success rate of integer ambiguity is 100% at single epoch on the fly.By adding one observation satellite,the vertical protection level will reduce about 0.2 m,the test statistic can detect cycle slips with 100% success rate.
In order to overcome the shortcomings of the traditional LIC algorithm with fixed filter length,an improved FLIC algorithm which filter length varied with the vector direction was proposed.And statistical variance of the intensities at all points in one texture streamline was employed to evaluate the visualization accuracy.The computational efficiency and visualization accuracy can be improved simultaneously by the new method.Furthermore,visualization of gravity vector field was carried out by improved FLIC algorithm,using color information to distinguish gravity anomalies or deflections of the vertical,and texture to express the direction of gravity vector.The results indicate that the visualization of gravity vector field has obvious advantages including multi-dimensional information expression,image cognition and vector direction details.
We present a new method,named baseline method,which used for recovering the Earth's gravity field and improving some parameters of initial orbital vector simultaneously.The principle of the baseline method is introduced.The formulae of transforming between baseline parameters and typical Cartesian parameters are deduced.And the sensitivities of inter-satellite observations to baseline parameters are analyzed.The two Earth's gravity field models are recovered from GRACE data by typical method and baseline method respectively,and their accuracies are evaluated.The results show that the accumulated geoid undulation error from baseline method is less 3 cm than the one from typical method.
Wavelet transform is widely used in deformation measurement data filtering,but there is still no definite solution to determine the best grading of its decomposition and reconstruction.Firstly,four parameters,the gap of mean square error variances between two adjacent gradings,correlation coefficient,signal-noise ratio and flatness,are integrated in as evaluation indexes.Then they are transformed to according to their contributions to the best grading.Finally,these four transformed parameters are added up to as a collective evaluation index,and the grading corresponding to the maximum value of the collective evaluation index is the best.Two examples,the one is a stimulant siginal,and the other is from actual deformation measurement,are used to test the new method,which prove the method is feasible and reliable.
According to experimentizing influence of environment temperature no at the instrument constant of a type of gyrotheodite,we analyze and deal with the data of instrument constant on each temperature,and summarize the diversification accord with linearity rule.Then it is possible to realize automatic compensation for influence of temperature to instrument constant drifting.This has practical value in improving the orientation accuracy and nicety of gyrotheodolite.
Marker-based watershed segmentation can effectively overcome the problem of over-segmentation and performs well in most cases.However it may not be suitable for remote sensing images due to the diversity of ground objects in local and regional scale.Regions with different characteristics can not be segmented with the same criteria arbitrarily.We propose regional adaptive watershed segmentation algorithm.First,a regional adaptive threshold image is constructed using the blurred gradient image combined with an evaluated statistic threshold.Then markers are constructed by threshold segmentation of the gradient image.Segmentation result is finally generated with an marker-based watershed segmentation.The algorithm is proved to be effective through a series of experiments including the comparison with the multi-resolution segmentation algorithm used in eCognition.
A new classification method is proposed for polarimetric SAR images.The H/Alpha/A decomposition is combined with Markov random field model.Firstly,H/Alpha/A decomposition is adopted to obtain 16 initial clusters based on their scattering characteristics.Secondly,the wishart distribution based on maximum likelihood is implied to update the classification result.Finally,by the adoption of wishart Markov random field model,iterated conditional model method(ICM) based on maximum a posteriori criterion is used to acquire the final classification result.Using fully polarimetric SAR images acquired by the NASA/ JPL,the experimental results verify the effectiveness of this improved algorithm.
Information extraction of land use from remote sensing image is one of the important ways for land use data acquisition.The spatial distribution of the land resources,the methods of the land-use,the resolution of the remote sensing data and number of the bands have a great impact on the information extraction results.The theory of decision tree methods are discussed.And the remote sensing data and other relevant data are used for the classification of land use information.By the statistic of the spectral characteristics,the characteristics of the normalized index of some type of the land-use,such as arable land,and the DEM model,the paper get a land-use decision tree branch for the information extraction,which can get a better solution for distinction the water bodies,construction shadows and other confused land-use type.The results proved that this method can achieve the initial classification and information quickly and effectively,which can make a feasible approach for land use information extraction by the comprehensive analysis of remote sensing data.
With there problems at hands,an improved random field,support vector machine conditional random field(SVM-CRF) was proposed.It uses SVM as its unary potential,combining the merits of SVM and CRF.Experiments using AVIRIS hyperspectral data as input were carried out,and SVM-CRF was analyzed extensively.Experimental results show that SVM-CRF is superior to SVM and classic CRF in terms of classification accuracies.
Remote sensing provides a valuable tool and enables to derive vegetation coverage over a large area using either empirical statistical models or sub-pixel analysis.However,to build an empirical regressive model needs a considerate ground truth data,while the parameters for the dimidiate pixel model(DPM) are often difficult to determine in a specific area.We combine the advantages of these two types of inversion models,using the regressive model to determine and validate the DPM parameters.The validated DPM was then used to derive the vegetation coverage information in a large-scale area.The Landsat TM and Terra MODIS data of Shihezi area of Xinjiang,China were acquired and used to derive vegetation coverage using the proposed approach.The scale effect in the inversion of vegetation coverage from the TM and MODIS data were also addressed in the study area.The results indicate that the proposedstrategy is robust in the retrieval of vegetation coverage over a large area from remotely sensed data.The empirical regressive model can help determine the DPM parameters and the DPM can then be applicable when there is no ground data available.The analysis of scale effect shows that the non-linearity of the inversion models mainly causes the scale problem in non-water land covers,while the heterogeneity in a pixel is a major reason that causes the scale effect of water-land boundaries in the retrieval of vegetation coverage.
Detecting moving ground vehicles from airborne video is a difficult problem because all pixels in the image are moving due to the self motion of the camera.We present a technique based on the normal component of the residual flow algorithm applied on using SURF operator stabilized frames to detect moving vehicles.Experiments show using SURF operator can improve the image stabilizing algorithm speed without loss of accuracy and this approach can effectively detect the moving vehicles.This method and airborne platform can be used as a new attempt to traffic information collection when the routine method can not be used.
We present a method for complicated planar building reconstruction from airborne LiDAR point cloud data.Firstly,the extracted building points are clustered into the different plane point sets.Then,those points in point set are fitted into plane.Every planar boundary is decided by multi-planes intersecting.From every planar boundary and the height of ground around building,we can calculate the 3D coordinate of every corner point parameter in building reconstruction.The building model can be reconstructed using these parameters.The experimental results show that this method is able to reconstruct effectively complicated planar building model.
A detailed study of choosing the best seam-line in image mosaic was carried out in LiDAR system.The tall ground object marked as obstacle can be identified by filtering laser points.And an improved A algorithm was proposed to make the seam-line keep away from the obstacle in the registered image automatically.We found that this method can optimize the seam-line exactly and intelligently to improve the quality of Ortho-image.
A new semi-blind high-resolution remote sensing image watermarking algorithm robust to general geometric distortions is proposed.The characteristics of high-resolution remote sensing images are given full consideration.In the watermark embedding procedure,the original watermark information is firstly processed by the mechanism of spread-spectrum.Then,in order to generate the key matrix,the watermark values and the most significant bit of the image are compared.Finally,the watermark is embedded into the most significant bit or the sub-least significant bit,and the key matrix is saved as the binary format.In the watermark extraction procedure,the template matching technique is employed to extract the optimal watermark information.Experiments confirm that the watermarked image has a very good invisibility and less effect on the original image.In addition,the watermark information embedded by the proposed scheme is robust to the general geometrical attacks such as cropping,rotation,translation,noise and filter and so on.
A new fusion method of remote sensing images based on fast discrete Curvelet transform(FDCT) is proposed.The FDCT is firstly done to the multi-spectrum(MS) and panchromatic(PAN) remote sensing images with space registration.Then fusion of low-frequency sub-band based on standard deviation weighted strategy,median and high frequency sub-band based on maximum replacement strategy,high frequency sub-band based on direct replacement strategy are implied.The fused image is produced by the inverse FDCT.The experimental results with IKONOS,QuickBird,WorldView-2 MS and PAN images indicate that the proposed method significantly outperform the traditional methods.
Phase unwrapping is one of the most important cases in PS-DInSAR.As it is similar with the GPS ambiguity resolution,the process to solve the ambiguities and deformation parameters by the algorithm of LAMBDA is presented based on the function models of PS-DInSAR.Two algorithms of LLL and inverse integer Cholesky are applied to decorrelate,which is the also named as z transformation in LAMBDA.With the criterions of iterative numbers of z transformation,condition number of variance-covariance matrix and mean correlation coefficient of ambiguity vectors,the performance of two algorithms are compared by random numerical simulation.The results show that the modified algorithm of LLL is equal to inverse integer Cholesky algorithm.The interpretation to the equality is given in theory.
We propose a quantized quality-guided phase unwrapping algorithm.The quality values of the quality map are denoted by integers.A priority queue consisted of a static array and many doubly linked lists is introduced.The relation between the integral quality values and the indexes of the array is established and priority queue keeps all the pixels by their quantized quality in a non-decreasing order by maintaining the doubly linked lists in an increasing order.The tests performed on interferometric SAR data and simulated InSAS interferogram confirm the high efficiency of the proposed algorithm.
Through introducing the three-dimensional Douglas-Peucker method for the geomorphology generalization to Level of Details of terrain and building the Quad_TIN model,we put forward a new algorithm for the geomorphic real-time dynamic rendering.The strategy of this algorithm is to classify the irregular sampling points from the perspective of the global importance firstly.Then the structural advantages of quadtree of GRID and TIN of the sampling points are combined in terms of visual expression and the dependence is removed by constructing quadtree spatial index in manner of isolated partition.After that,according to the viewing distance and simulation error,the divided level of quadtree and the inserted order and number of sampling points are quickly determined.Finally,gaps between nodes are eliminated through adding splitting-points.Experiments have proved that the simulated accuracy can be improved greatly with satisfactory speed.
An adaptive digital watermarking algorithm based on discrete Fourier transform(DFT) is proposed to protect the copyright of digital raster maps.This algorithm can utilize the human visual system(HVS) and the blocking DFT to determine adaptively the texture region where the watermarking information can be hidden.And the watermarking information is embedded to the middle frequency coefficient of the Fourier transform domain of the texture region.The algorithms with different size blocking(blocks) have different robustness.This problem is discussed,and these algorithms are analyzed with different size blocking,such as the 2×2 blocking,the 4×4 blocking and the 8×8 blocking.Experimental results show that the algorithm with 2×2 blocking not only has the good transparency,but also has the good robustness to the image intensity attack such as cropping,JPEG compression,noise adding,image enhancement,and so on.The algorithm with 2×2 blocking performs better than the algorithms with 4×4 blocking and the 8×8 blocking.
We propose an auto-calculation method for the catchment area of any given nodes in the municipal drainage pipe network.Firstly,the drainage pipe network graph is to be translated into directed graph.Secondly,the upstream nodes of any given node are to be obtained with the breadth-first search algorithm in graph theory.Thirdly,the calculation of the associated sides of the upstream nodes is carried out.Finally,according to the catchment area of upstream nodes and their associated sides,the catchment area of the given nodes can be calculated.And we put forward the correspondent programs for the new method.
A new algorithm for finding k-nearest neighbors based on spatial sub-cubes and dynamic sphere is presented.At first,the min-max box of the dataset is divided into a set of uniform sub-cubes,every point is distributed in a sub-cube.Then the dynamic sphere is builded using the test point as the center of sphere in order to reduce the searching scope.The radius of sphere can be estimated and adjusted according to the volume of min-max box,the numbers of spatial points and the numbers of k-nearest neighbors.The experimental results show that the algorithm can find k-nearest neighbors fastly.
We propose a new method for extracting loess shoulder-lines from grid DEM.The morphological characteristics of loess shoulder-lines are investigated firstly.By applying the edge detection approach,a new method for extracting loess shoulder-line candidate points is proposed based on the prominent height variation of the points.The algorithm then connects the candidate points to small line segments by morphological methods.Finally,precise,systematic loess shoulder-lines are extracted after refining the line segments.Experiments in the loess hill area show that the extracted lines from gradient based operators,such as Sobel,Roberts and Prewitt,have relatively poor results on both integration and matching precision with manually extracted loess shoulder-lines,while loess shoulder-lines extracted by LOG operator has very good matching ratio to the manually extracted loess shoulder-lines.So LOG is an ideal loess shoulder-lines extraction operator and can be used to extract loess shoulder-lines effectively and automatically.
With the increasing user requirements,the single geospatial information service is hard to satisfy business needs and this promotes the emergence of geospatial information services composition.Finding suitable geospatial service in the Internet for the composition is a crucial task.However,the research of geospatial service composition focuses matching the function of single abstract service,and neglects the effect among service composition.We propose a semantic matching method for geospatial information service composition,which can match services dynamic according to their practical situation of the service composition and the effect among the matching process of each atomic service in the composition service.Then an abstract template of service composition is converted into an actual chain of geospatial information service effectively and accurately,which can execute in the workflow engine.
The purpose of this study is to develop an information system for evaluating intensity of land use in development zone,so as to make the evaluating processes modeled and automated.After analyzing the requirement of land use intensity evaluation,database and software architecture and some key technologies such as 3D theme mapping,parsing engine of computing expression,integrative storing of multi-type data were designed according to requires of software engineering.The software was developed using the ArcGIS Engine Developer Kit and C# development language and was triumphantly applied in the case of land use intensity evaluating of Wuhan East Lake High-Tech Development Zone.It has been proved that the design methods and key technologies resulting in high efficiency and high-quality in the software should be introduced.