2010 Vol. 35, No. 6
GNSS radio occultation technique has been demonstrated to be a powerful tool for remote sensing of the Earth's neutral atmosphere and ionosphere.It receives GPS signal through a dual-frequency GPS receiver onboard LEO satellites.Affected by the vertical refractive index of the communication medium,the propagating path will be curved,which would result in a time delay when arriving at the receiver.From this time delay,we could obtain the ionospheric refractive index.Then the ionospheric electron density profiles could be derived.Relying on its characteristics of all-weather,high vertical resolution and broad coverage,which other ionospheric surveying methods do not have,the radio occultation method has laid foundation for the new ionospheric remote sensing technology,and offered a unique chance for improving our knowledge of ionospheric behavior.Firstly,we review different inversion methods and application examples of GNSS ionospheric radio occultation.Finally,we investigate and summarize the existing problems as well as its future research directions.
We studied a so-called crossover point approach for low-cost LEO cluster precise orbit data to recover the global time-variable information without accelerometer systems (ACC) or k-band ranging system (KBR) measurements.This approach can reduce the non-gravitational force effects on LEO,obviously.We analyzed the 20 d data coverage and the orbit radial error effects on this approach.With 20 d precise orbital data,the crossover points have well global coverage.On the other hand,the radial orbit accuracy of COSMIC data is about 15 cm at 800 km orbital altitude.It is not sufficient for the time-variable gravity field recovery with crossover point approach.If the orbital radial component error is reduced to about 1 cm,the time-variable gravity field can be recovery with acceptable accuracy.
With one year of radiosonde observations as reference,the precision of GPS radio occultation (RO) atmospheric parameter profiles,including refractivity profiles,temperature profiles and pressure profiles,in four different austral seasons are compared and analyzed at 0～30 km altitude.We found that in average,the precision of RO profiles is worse in the austral spring than in the other three seasons.But in different altitude regions,the precision of RO profiles varies greatly.Generally,RO profiles are more precise at 5～25 km than at 0～5 km and at 25～30 km.In the lower troposphere of 0～5 km,the precision of RO refractivity profiles is the worst in the austral Summer and best in the austral winter.This is that the water vapor content in the atmosphere is the highest in summer and is the lowest in the winter.Above 5 km,the precision of RO pressure profiles is the highest in the Austral summer.The precision of RO temperature profiles doesn't vary much in different seasons and in different altitude regions.
The launched COSMIC (constellation observing system for meteorology ionosphere and climate) satellite in 2006,operated by Taiwan of China and America,provides about 2 000 occultation data everyday,which is abundant for atmospheric probing research.The inversed electron density profiles through Abel integral method and the onion peeling method both have been compared with the results obtained by traditional ionospheric observation methods.Based on this,the statistical analysis has been made on the comparison results of the important ionosphere parameters,and the problems existed in the inversion process,together with the impact of fitting and interpolation methods on the occultation have been discussed.The numerical results indicate that the electron density profiles obtained by the two kinds of inversion methods accord well with the results observed by ionosond.The computed foF2 is consistent with the results of ionosond.However,large error exists between hmax and hmF2.In order to obtain a reliable distribution of the ionosphere over some places,we need to combine the two methods for analysis,or to improve the inversion methods so as to acquire high-credible hmax values.
The fault detection and identification methods of Snapshot and Filtering are applied to detect and identify the single and multiple faults GNSS observations.Compared the performances of false alarm rate,miss detection rate,correct identification rate,computation consuming time,availability and precision between Snapshot and Filtering methods,the Filtering FDI method performs,especially in multiple faults identification,is better than Snapshot method.
At present,space-borne GPS receivers have gradually evolved into a standard tracking system for low Earth orbiting (LEO) satellites.In order to produce high accuracy onboard navigation solutions continuously,an extended Kalman filtering model of real-time orbit determination with space-borne GPS measurements is firstly derived based on the dynamic model compensation method,and relevant software named SATODS is developed subsequently.Then a simulative test was carried out using broadcast ephemerides and space-borne GPS dual frequency pseudo-range data from CHAMP satellite with SATODS.The results compared with JPL RSO demonstrate that real-time orbital accuracy about 1.0 m for satellite position and 1.2 mm/s for velocity (3 d RMS) is feasible with dual-frequency code measurements.In addition,it is hardly subjected to data quality of GPS observation such as the number of GPS satellite and the value of PDOP,and so on.
We analyze the reason that zero-differenced ambiguity cann’t be fixed in a standard mode.We proposed the FBI(fractional bias isolation) mode and method for zero-differenced interger ambiguity resolution.We found that with 30 min observations,the accuracy of mm-level in horizontal and cm-level in elevation direction can be achieved,which has respectively enhanced one order of magnitude and has improved approximately 50%-60% in horizontal and elevation direction compared to float solution.In the meantime,the accuracy of ZPD(zenith path delay) estimation has been also ameliorated about 20%-30%.
We studied the precise satellites clock error estimation algorithm for navigation satellites.The computation of precise satellites clock error based on the non-difference phase was realized and the results were compared with the products provided by IGS analysis centers.The results show that the post precise satellites clock products calculated by the software are equal to those provided by IGS Final,CODE,JPL and MIT.Their bias are less than 0.1 ns and the precision is equal to the precision of IGS Final.In addition,the real-time precise satellite clock error is also estimated with a high accuracy about 0.1 ns compared with that of CODE.
Because of the mass redistribution of the geophysical fluid loading,obvious non-linear variation exists in most of the global GPS sites(mainly seasonal variations of the vertical component).And its transient motion with respect to the reference frame can not be completely described by a linear velocity.We consider and deal with the non-linear and seasonal variation characteristics of frame sites appropriately,together with combination of the linear velocity model are of great importance in the establishment and maintenance of the dynamic terrestrial reference frame.Firstly,the current status of the global and regional reference frames have been briefly introduced.Then,the disadvantages of the linear ITRF model has been investigated,leading to the thoughts of building velocity model considering different types of non-linear variation,so as to establish and maintain a dynamical reference frame.Based on this,method of regional reference frame establishment and maitenance has been discussed.Finally,we look into the future developments of ITRF under the GGOS background.
We discuss the future configuration of the Chinese astrometric and geodetic VLBI network.The technical specifications of equipments at observation stations and the corresponding necessarily required investment are proposed based on the history,current status and near future development of the domestic and international VLBI technology from the view point of astrometry,space geodesy and deep space exploration.This configuration is only for the criticism and reference in order to minimize the investment,maximize the efficiencies and the integration of various efforts and resources.
Space very long baseline interferometry(SVLBI) is the unique space technology that can directly interconnect the three reference systems,which include the conventional inertial system (CIS) fixed in space and defined by radio sources,the conventional terrestrial system (CTS) fixed to the Earth and defined by a series of observing stations on the ground,and the dynamic reference system defined by the movement of a satellite.An SVLBI mathematical model with nutation parameters is discussed,and its feasibility for estimation of EOPs is discussed.Simulations have been done for the SVLBI observations and related parameters for the RADIOASTRON.With the results of simulation,the sensitivity of EOPs to the priori errors of the associated parameters which include the coordinates of the ground VLBI stations,the orbit parameters of RADIOASTRON satellite,clock parameters and the coordinates of radio source is studied.
We introduced the principles and influencing factors of the zero-difference kinematic POD,and analyzed the impacts of GPS orbits and clock products on the kinematic POD of GRACE.The results indicate that the precision of the kinematic orbits of GRACE is at the same level for the IGS precise ephemeris and the rapid ephemeris,which means that the precision of the GPS orbits have no influence on the GRACE's kinematic orbits.When using high-rate clock products with 30 s and 5 s interval,the RMS in radial direction are 3.8 cm and 3.4 cm respectively,which indicates that high-rate clocks product can improve the precision of kinemaitc orbits.
Both static and kinematic tests were investigated by using IGS 5 min,IGS 30 s,CODE 30 s and CODE 5 s-interval precise satellite clock product in precise point positioning (PPP) solution.Test results show that the sample rate of IGS satellite clock has little effect on the static PPP solution.Using a higher sampling rate can achieve high precision in a short time but they can all obtain the same precision nearly with the time going.While in the kinematic PPP,the sample rate of satellite clock has a significant impact on the PPP solution.The higher the interval of satellite clock is used,the better accuracy it can achieve.The accuracy of kinematic PPP achieved by using 30 s-interval precise satellite clock is improved by nearly 40%-50% in the plane and 20%-30% in the vertical direction with respect to the solution by using 5 min-interval precise satellite clock.While using 5 s and 30 s-interval satellite clock almost produce the same accuracy of kinematic solution.When achieve the same precision,the time to get convergence by using 30 s and 5 s-interval clocks is shorter than that of using 5 min-interval clock significantly.
On the basis of the theory of Blewitt,a new method is introduced.The new method uses the ionosphere-free combination LC.Differencing LC data between pairs of satellites is not effective to the ionospheric delay and receiver clock variations.Therefore it is useful for cycle slip detection.The wide-lane combination,the ionospheric combination and satellite difference of the ionosphere-free combination are all used to detected cycle slip.The results show that the combination of the three methods can accurately detect cycle slip.
We study the ground-based GPS zenith water vapor remotely sensing theory,and propose the double difference method for ZTD estimation with IGS precise ephemeris provided and precision clock error.We use a Kalman filtering method and a double-difference method to estimate the ZPD,respectively.We compare the results with zenith delay provided by IGS stations.The results calculated with two methods and provided by the IGS trend of the same epoch with the largest difference doesn’t exceed 2 cm.According to the solution of ZTD,we use empirical models to solve ZHD,and isolate ZWD.Finally,we carry out the solution of sub-atmospheric precipitation.
GPS precise data processing requires accurate antenna phase-center corrections.In the last decade,the traditional relative antenna phase-center correction was replaced by absolute antenna phase-center correction.We present the recently developed absolute antenna phase-center corrections.Then,we study the influence of this model on GPS precise data processing that is studied by analyzing GPS network with different antenna type,difference scale and difference elevation.Based on these analyses,we demonstrate the difference between relative and absolute antenna corrections,and the factors that impact the baseline solution by absolute antenna corrections.Finally,some suggestions are also given.
We design a set of processing methods for navigation data of offshore towed streamer exploration.Firstly,we introduce the key technologies of preprocessing,such as gross error detecting of crossline acoustics observation and detecting the problems of RGPS which likes cyclic slip.Then,we present the processing methods through three aspects which are constrained network adjustment,curve integration,and stable adjustment.Also,we establishe the constraint equations and give the strategy of weight fixing.We also give the results compared with the software used commonly aboard to proved validities of the whole processing methods.
We discuss how to fix the offshore towed streamer shapes.Firstly,we introduc the arc Model,the polynomial model and the curve integral model.Then,we compare and analyze each model.Finally,with an example,we show how the curve integral model works.
We study the navigation and positioning data processing of two-dimensional explo-ration,and design a set of navigation and positioning data processing systems.For the characteristics of two-dimensional exploration,we use the integral model to simulate the shape of the cable to calculate the detection point coordinates.Under the premise of without loss of accuracy,this method simplifies the calculation process and improves the calculating speed.The main components of this method include gross errors detection and the repair for DGPS data and compass data,data epoching,curves adjustment and detecting node calculation.Compared with some abroad mature deep-sea exploration and navigation data post-processing software,the calculation accuracy of this method meets the demand of deep-sea exploration work.
Aiming at the condition that among observation equations there are errors in both observation vector and structure matrix,we propose an improved algorithm of total least squares(TLS).Meanwhile,when observation vector and structure matrix are with unequal precision,we deduce the calculation formula using an adjustment model with constraints.Analysis and calculation show that the algorithm that improves approximation accuracy of recursive TLS effectively,can satisfy the conditions needed by fitting equation.
The conventional solution method for TLS is based on matrix singular value decomposition.This method is rigorous in theory,but is complex and not easy to be programed.The complexity of the method is an important reason restricting TLS application in the field of Geomatics.By introducing the total least-squares adjustment standard,we derive the total least-squares iterative method,which is simple and easy to be programed.Through the introduction of multi-function derivative knowledge of the implicit function to determine the linear information of the parameters to observational data,we solve the problem of assessing the accuracy based on TLS.Finally,we apply the new method to the fitting function determination and coordinate transformation parameters determination based on measured data,and verify the feasibilities of the new method.The new method has a great significance to the popularization and application of TLS.
To solve the problem of independent baselines search,we introduced four classical algorithms generalized into two categories.We proposed a new algorithm based on vertex incidence matrix.And we analyzed the optimality and relative merits of these five methods.The simulation results verify the effectiveness and advantages of the proposed algorithm.Moreover,we studied the application of the new method to independent double-difference ambiguity selection.The study has demonstrated that the proposed algorithm outperforms the existing methods and can be extended to a higher dimensional space.Besides,we mertioned a new concept of vector weight.
We introduced Meanshfit algorithm to the road extraction in small scale aero-borne image by estimating the local maximum of the possibility distribution nearby the specified points.With the shadow kernel of the kernel function,the searching process will climb to the top along the gradient direction of the possibility densities with a fast convergence speed.We extracted the road points nearby the old road vectors for change detection,and traced the new road with a single initial point.Experiments prove the validities of this algorithm.
A two-step progressive texture retrieval algorithm for multi-source remotely sensed imagery based on Contourlet transform and spectral histogram is proposed.Firstly,the Contourlet transform is applied to decompose texture features of remotely sensed imagery at different scales and different directions.Then,low-pass signal and high-pass signals are adopted to realize elementary retrieval and exact retrieval relatively.The experiments on texture database obtained from USC and QucikBird images show that the proposed algorithm not only utilizes the advantages of Contourlet on multi-scale and multi-direction texture feature extraction,but also makes full use of the efficiency of spectral histogram on distributed statistical feature representation.On the aspect of precision and efficiency,Contourlet transform outperforms wavelet transform.The proposed method provides a powerful tool for texture retrieval of remote sensing image.
We propose a new unsupervised classification method,GMM-UC,for remote sensing image based on Gaussian mixture model.This method is based on the theory of finite mixture model that the remote sensing data is mixed by a finite number of sub Gaussian distributions to a certain percentage.Through the improved EM algorithm,GMM-UC automatically determines the number of sub-Gaussians and its parameters,then restores all land objects (a sub-Gaussian is corresponding with a class of object).The improved EM algorithm solves the problems of high dependence on the initialization parameters and easily converging to the boundary of parameter space in standard EM algorithm effectively.Due to the adaptive design of the number of sub-Gaussian,the EM algorithm can find the shapes of the data distributions in a feature space furthest at a certain range,and the sub-Gaussians and and their mixed distributions is considered optimal.Experiments indicate that this method overcomes the fault of traditional unsupervised classification method to some extent,and expands the application field of it.
The rigorous geometric model of SPOT-5 images is introduced firstly,and a new geopositioning method based on line-of-sight rectification is proposed.Then,we deduce the bundle adjustment of line-of-sight calibration.Test results using SPOT-5 images show that using a little GCPs the systematic error can be reduced effectively after adjusting the look angles,and the object positioning accuracy can be impoved.Comparative experiment shows that line-of-sight rectification can provide similar position accuracy in case of 5 control points,with direct correcting of orbit and attitude error in case of 10 control points.
Pattern recognition is a focused technique of computerized image analysis.One difficult problem of which is the uncertainty on the expressional face recognition.A new method on pattern recognition is proposed on the basis of data field.In the methodological framework of pattern recognition,the method is designed upon the data field and hierarchical modeling.This method is applied to recognize JAFFE,a famous expressional face database.The whole recognition rate may be about 94.3%.This result indicates that the method is suitable and effective.
GeoVideo,the video with geographic description,is a new approach to the geo-spatial representation.While VideoGIS is the information system to collect,manage,distribute and utilize the GeoVideo resources.We aim to design and implement a Web service-based VideoGIS using Abode Flex.These types of web services,i.e.web mapping services,web video services and geographic XML services,were mashuped in rich Internet application.The key techniques of VideoGIS system design include GeoVideo data collection and geographic descriprion,costribution of web services,and system integration in Flex Builder.Video location and semantic descriptions were edited in XML document and the video paths were created in map layer with multiple video metadata.video service and mapping services were published using Adobe FMS and ArcGIS server,and the XML documents were stored in IIS virtural directory.In Flex Builder,Geovideo player was designed using VideoDisplay Control,and the map operations were implemented using ArcGIS Server Flex API.The system development shows that VideoGIS design and implmentation in Flex are simple and efficient,and the interfaces are dynamic and effective.We believe that VideoGIS has application potentials in facility management,geographic education,tourism promotion and other fields.
China Geodetic Coordinate System 2000(CGCS2000) was formally released to be used in China on July 1 in 2008.So all products related to reference ellipsoid should be transferred to CGCS2000.We main discuss how to transfer maps from reference coordinate system to CGCS2000 with 3D four parameter coordinate transformation model.In order to test the proposed method,the experimental areas which could stand for characters of whole country were chosen.Computation and analysis show that the recommended method can fully meet the accuracy of basic scale map transformation in China,and is better than the existed methods while it is simple and convenient.
We constitued a fuzzy similar matrix with characteristic function of samples of evaluated object,and computed transitive closure of fuzzy similar matrix.Accoding to transtive closure,different λ-cut of fuzzy set are clustered,and we may get rational classification result.With an example of evaluation of district econmic develop level,we present the steps and method of fuzzy clustering for multi-factors evaluated system.