2008 Vol. 33, No. 5
The geostatistical basis for adaptive/progressive sampling is discussed,following an introduction to the necessary statistical background and developments in geographic information technologies.Where computational resources are limited,as they are in the field,strategies that combine heuristic and numerical approaches are the key to successful field implementation.A sequential algorithm for rapid location of further samples is formulated,using the criterion of maximum global reduction in Kriging variance.Results from a test confirm the effectiveness of the proposed algorithms.
A reconstruction method for complex object is developed using non-metric images.Firstly,a rotatable platform is designed to capture images and a planar grid board on its top is for camera calibration.With the help of the platform,images are captured and then processed with multi-baseline image matching algorithm to produce corresponding point features.Also benefiting from the rotation platform,the initial value of images' external parameter can be obtained.All these parameters and coordinates of corresponding point feature are seen as initial values for the block bundle adjustment program and then point clouds and accurate image parameters are obtained.In the end,a vision constrained Delaunay triangulation algorithm is used to produce surface model of the object.The method includes artful hardware design,efficient image matching method and rational 3D surface construction approach.Experiment results show that it is feasible and effective.
The method how to generate virtual image with integration based on field of outer view are discussed.The experimental results show that elevation precision achieves leading level in similar product and the plane precision achieves advanced level.Through the results verification of mapping precision to the test block of Beijing,topographic which is mapped using SWDC4 images under 1∶500,1∶1 000,1∶2 000,1∶5 000 scale satisfies the precision request and is superior to the limitation of specifications of surveys.
A new method to detect the change of two SAR images using cross entropy based on the statistical character of SAR cluster is proposed.Rayleight distribution model is used to describe the SAR cluster,and the different index of two images is derived using Kullback-Leibler divergence in detail.Constant false alarm rate is also introduced to segment the change areas.The method is demonstrated feasible using the RadarSAT SAR image.
An extension linear spectral model based on total least square algorithm is proposed,and iteratively abstract the better endmembers and physically based constrained algorithm with this model are presented. The results show that the proposed model provides better accuracy than traditional least square method,the total estimating precision improves by 10%-20%.
The theory and method of digital image inpainting technique is introduced.After analyzing the particularities and difficulties for restoring lost information on remote sensing images,a novel method based on accessorial GIS data is proposed.The method enhances automatization level of the algorithm,and improves the quality of the restoring.
The problem on scoring the image scrambling extent objectively for still gray-level images is studied,and the image scrambling extent can be expressed by the DSF(distance scrambling factor) and GSF(gray scrambling factor) is proposed.The evaluating equations for image scrambling degree are given and scrambling degrees of some commonly used scrambling methods are computed.Experimental results show that the proposed image scrambling degree can perfectly describe the image scrambling extent.
A novel model based on wavelet transform and support vector machine for dam deformation prediction is presented.Firstly,through the wavelet transform,deformation time series is decomposed into different frequency components.Then,according to the different characteristics of the decomposed components,different support vector machines are constructed to forecast the components.Finally,the predicted results of the components are reconstructed to be used as the final prediction result of deformation.The calculation result shows that this model has higher forecasting precision and greater generality ability.
A GPS IF signal computer-aided simulation method is proposed.The carrier Doppler frequency and the total propagation delay time can be modeled or calculated with the input GPS satellite ephemeris file.The simulated GPS IF signal is output to a text file for GPS software receiver post-processing and analyzing.The simulated GPS IF signal spectrum is compared with the received real one.The validity of the simulation result is verificated.
The principle to retrieve sea surface wave height using GNSS-R data is analyzed,emphasizing on the ICF and coherence time.The Oceanpal observational data is filtered,its autocorrelation function is computed,cubic polynomial interpolation is used to estimate the coherence time.A new experiential retrievable model to compute the wave height is educed.The result is consistent with the data of acoustic wave instrument.
TEC information from 2004 all-year GPS measurements on more than 40 continuous-operation stations over China is resolved,these TEC data are modeled by three commonly used regional models: poly model(Poly),trigonometric series model(TRI) and low-degree spheric function model(LSF).The results show that Poly and LSF models are consistent with GIM model(CODE),but the former is degraded by marginal effect,and TRI model have noticeable bias to GIM model,which alarms us this model is not suitable for the region,just for the local.Furthermore,the accuracy on high latitude area is better than that on lower latitude area.
In order to studying the means of monitoring and predicting to solar flare,the observation data of more than 60 global GPS trackingstation from IGS is downloaded,TEC that coming from two solar flare of X level is calculated and analyzed.The results show that the sudden increasing phenomenon of TEC is monitored in solar flare,which all GPS tracking-stations locate the high latitude area and the low latitude area,and it is consistent with the figure of X ray flux;difference TEC can gain more exact change of total electron content,so it adapts to the detailed study of solar flare;the phenomenon of TEC great range jumping is discovered in studying absolute TEC range between closer-day,so the experimental viewpoint and conclusion are obtained.
A dynamic approach is developped,which is based on the motion equation of GPS,to produce the high frequency GPS orbits.The method fits dynamic-integrated orbits to precise orbits at the sampling epoch.A model based on Helmert transformation is introduced.Results show that the fitted orbits have a same precision of IGS SP3 orbits.
The method for predicating the satellite clock error is presented through the optimization method of gray system theory.The GM(1,1) model establishing,the original sequence's transformation and model optimization method are discussed,and the parameter optimization method of gray system is given.Calculating results show that the parameter optimization method can improve the precision of the satellite clock error predicating.
Precision of satellite clock is one of the most important factors in GPS precise positioning,with widely applications of IGS precise ephemeris,which becomes a more and more attractive research orientation.Spectral analysis method is applied to analyze the precision of IGS precise ephemeris clock.The conclusions show that different satellite clocks have different precisions,and some of them have great magnitudes than the so-called precision.
Two methods of improving the generalization of neural network,using EKF and UKF algorithm,are described.In order to avoid the random selection of some parameters for UKF,a new adaptive UKF algorithm for training neural network,by using windowing residual vectors to adaptively estimate the covariance matrices of the observational vectors and the model errors,is established.The results show that EKF,UKF and adaptive UKF algorithms for training neural network can all improve the generalization of neural network,and the adaptive UKF algorithm is better than others.
The independent variables error and variables error are further studied.A new regression model is given,and calculation algorithms are deduced.Numerical experiments are used to demonstrate correctness of the new method,and the results are same as those of the total least square method.The algorithms proposed in this paper simplify the complicated matrix decomposing calculation,bring TLS into category of surveying adjustment.
For improving the digital map's quality,collocation method is a helpful tool to fit the systematic errors and weaken the effects of the stochastic errors.The main difference of the collocation applied in our situation is that the remained systematic errors resulted by the inaccurate covariance function will be fitted again and reduced from the estimated signals.In this way,the modified collocation can effectively resist the influences of systematic errors.An actual calculation and corresponding analysis show that the proposed method is effective in improving the quality of digital maps.
Based on the analysis of the factors which have influence on performance of navigation message and must be considered in the navigation message structure design,some criterion and methods of the performance evaluation of the structure of navigation message are proposed.The criterion and methods include the total number of data bits,the time to first user receiver position fix,the time for user receiver to acquire a set of almanac data,the flexibility and expansibility of navigation massage structure and the utilization ratio of the satellite communication channel.As an application example,the navigation message structure performances of GPS and GLONASS are analyzed.
Considering the movement character of discontinuous deformation in Jiashi strong earthquake swarm,the stress and distance fields of fore-and-aft earthquake are systemically studied using the spherical discontinuous deformation analysis method with the restriction condition of GPS observations.Through the second distance restriction,the process of the pregnant,happening and the congregation again of earthquake are numerical simulated,and the earthquake movement law in Jiashi area is opened out.
Graphic simplification can cause the transformation of local topological relations between lines,line and region,regions and that after combinatorial.On the basis of abstraction of local topological relations,the rule of local topological relation between lines is emphasized,and the disciplinarian of combining the adjacent local topological relations is investigated.The method of deleting local topological relations is proposed.
In order to identify the proximity relations between the constraining objects,the characters of the constrained Delaunay triangulation which is built while contour lines are analyzed,elevation points and terrain variation lines are constrains,and a detailed classification for edges of the constrained Delaunay triangulation is made,a method for constructing one kind of hierarchical structure of the contour lines is proposed.The technologies on adjusting the direction of the contour line based on this hierarchical structure are studied.
A cluster storage architecture called as GOS for spatial object based on object-based storage is presented,which providing a new method to storage and management of spatial data.GOS extends the idea of objectbased storage to repartition the function of database server and storage devices.In GOS,the storage device performes to directly access to spatial data based on object-oriented model while the server performe some administration for spatial object such as spatial query,spatial index,transaction processing and etc. This architecture promotes the scalablity and performance of the spatial database system.And the GOS support the object-oriented spatial data model on the logical layer to simplify the data model map and the parallel spatial query to improve the performance of spatial data query.
Service oriented architecture(SOA) is the development trend of interoperability and integration of geospatial information.Grid computing is a means to implement SOA,which has many virtues,and provides a new idea to implement sharing and integration of geospatial information in wide domain.A scheme to construct geospatial information service registry center is presented.Based on OGSA,WSRF and OpenGIS series standards,the middleware and framework provided by Globus Toolkit 4,that scheme can resolve the sharing problems of geospatial information service in grid environment,at the same time of assuring the standard compliance.
An approach for Web service semantic matching is proposed,two kinds of services selecting strategies of based on semantic matching and QoS are focused.The services selection approach towards the individual Web service and the services for composition is analyzed.The hybrid selection strategy is put forward and the corresponding algorithm is proposed.The results show that the hybrid selection approach can deal with different requests,increase the correctness and improve the efficiency of services composition.
According to the characteristics and needs of the emergency platform,a 1D mathematical model of flood routing is established and the equations are settled through the convenient Kulan's format of the Diagnostic line method.Using Matlab,a flood routing software is programmed and the visualization of the results is realized.Based on the application study,the utility and reliability of the established model is proved.
The relation of new features and computation complexity in real-time application are analyzed,a suitable profile for it is presented.According to the architecture and features of BSP-15,the design methods of H.264 encoder is proposed and implemented.Experimental results show that the encoder based BSP-15 can satisfy the real-time applications with acceptable video quality.