2007 Vol. 32, No. 10
Based on photogrammetric principals two methods for multi-baselines forward intersection are introduced: bundle forward intersection and linear forward intersection. The two methods for multi-baseline forward are compared and analyzed. And how the baseline length and sequential-stereo-image counts influence the forward intersection precision is analyzed. The analysis result can give directions for practical production.
Automatically extracting building information from dense “point cloud” is a critical research topic in the process of building reconstruction using laser scanning data.Based on the distribution density of the “point cloud” of the vehicle-borne laser scanning system, a new way to extract building information by using grids is proposed. This method is much more efficient and it can automatically extract building information from dense scanning data quickly and meet the precision requirement of city reconstruction.
According to the principle of analytical subsequent relative orientation, the formula of vertical parallax on corresponding point is deduced when stereo model is reconstructed by using the elements of exterior orientation of images. The method is verified by simulative images and tested by two kinds of different scale images taken from Xinjiang and Qianxi, China. Compared the vertical parallax of models reconstructed by using the elements of exterior orientation obtained by GPS-supported bundle block adjustment with that determined by position and orientation system (POS), the vertical parallaxes of models are analyzed. The experiments have shown that the vertical parallaxes of stereo models computed with POS elements of exterior orientation are intolerable. Therefore, how to eliminate vertical parallax is very important for the direct georeferencing and cartography of reconstructing stereo models.
The concept of granular computing into line detection in remote sensing images using Hough transform is introduced, according to the characteristic of granular computing that it can observe problems at multiple views and levels. An algorithm on line detection based on granular computing is established by analyzing linear edge features with granularity of various sizes. Different quantification interval of corresponding granular level is employed, so as to extract lines with different length and width efficiently. An application on airport runways extraction in the real IKONOS image proves the validity of this algorithm.
Clear and continuous edge feature is important to analyze and interpret SAR images. A new edge detection of SAR images is presented by analyzing regularization method. There are three steps. Firstly, using a modified regularization method reduce the speckle noise. Then, a statistical method is proposed to segment the target of interest. At last, edge detection is realized by a window method. Compared with traditional methods, experimental results with MSTAR dataset show that this method can maintain detail feature, resolve the broken edge and decrease noise efficiently. This algorithm has wonderful edge-detection performance.
2D Gaussian filter based on local fringe rates estimation is proposed. The method estimates the local fringe rates by using the maximum-likelihood method, and evaluates the intensity of the noise by using the local standard variance of the noise. The parameters of the filter are modified adaptively according to the above two factors, and then adaptive filtering is implemented. By using both, simulated and SIR-C/X SAR generated interferograms, the performance of the method is demonstrated and compared with the mean filter, the median filter and the Goldstein filter. The result shows that the method can filter the phase noise effectively, and preserve the details of interferometric fringes better.
Quality assessment for remote sensing image compression is of great significance in many practical applications. A muti-dimensional structure model is designed for image compression assessment, which consists of gray character distortion dimension, texture distortion dimension, loss of correlation dimension. Based on this model, a new index-Q is proposed and validated through a series of experiments. The experiments show that Q has a better consistency with subjective assessment results than conventional PSNR.
Using projection pursuit learning network (PPLN), a new classification for remote sensing image is proposed. The PPLN algorithm integrates the advantage of artificial neural network (ANN) with nonparametric statistical technique, projection pursuit algorithm (PP), which is capable of providing less network neurons and good robustness. In this study, the structure and improved learning algorithm of PPLN is introduced in detail. Using the TM image of Suzhou, an experiment of classification is done and the classification precision is superior to that of BP neural network and conventional maximum-likelihood.
The classical robust Kalman filter is derived by the conditions of white noise, but during the dynamic data processing, the colored noises may occur. For example, the three phase differences measuring errors of GPS carrier waves are colored noises and the measuring data between the adjacent satellite almanac is correlated, and the measuring data for the non-adjacent satellite almanac is uncorrelated, the covariance matrix of observational noises is triple diagonal block matrix. For achieving the precise dynamic positioning(or navigation) by the three phase differences of GPS carrier waves, the robust Kalman filtering with dynamic colored noises and observational colored noises are derived. It may be proved that the robust Kalman filter with white noises is the special case of robust Kalman filter with colored noises, the robust Kalman filter with colored noises is the extending of robust Kalman filter with white noises
A new algorithm for adjustment model with some nonnegative constrained parameters is suggested. In this algorithm, nonnegative constrained least-square problems are first transformed to convex quadratic programming problems, and then the programming problems are solved for the optimal solutions. The are necessary and sufficient conditions on the solvability for optimization solution are given, consequently the general form of least-squares estimation in adjustment model is given, as well as algorithm that are simple and easy to understand. A comparative calculation on a simulation example indicates that this algorithm can be applied to adjustment computation in the practical measurement.
In this study, two-pass differential InSAR technique was used to monitor ground subsidence. JERS-1 L-band SAR imagery, acquired from December 1992 to May 1998, was selected as experimental data. 6 interferometric pairs were selected, and 4 subsidence bowls were found in the research area. Based on the temporal series of height change map, maximal vertical subsidence rate of two bowls was obtained. The studies show that DInSAR technique has great potential in ground subsidence monitoring.
In InSAR height measurement system based on formation satellites, time synchronization, phase synchronization and space synchronization should be established by the radars cooperated with each other in bi-static mode. The space synchronization requirement is discussed on demand of the InSAR height measurement under the condition of formation satellites. The influence of space synchronization on InSAR system performance on the height measurement accuracy and the resolution is analyzed.
Now Loran-C is one of important navigation equipments in ship, and Loran-C/INS integrated system plays an important role in naval ship in special period. In order to increase the integrated system’s precision, the wavelet’s reducing noise technique is introduced into Loran-C/INS integrated system to increase the signal-noise ratio of Loran-C, and then into Kalman filter, the dynamic and static experiment results based on practically measured data show that this method can obviously increases every navigation parameter’s precision, and can remove the position fault of Loran-C in a short time.
The ocean loading tides corrections of GPS stations in Antarctica are described. Based on the theory of ocean loading tides, the displacement corrections of ocean loading tides on GPS stations in Antarctica are computed with CRS4.0 model. These corrections are also applied to GPS data processing. The GPS data are analyzed by the GAMIT software with and without these corrections. The authors compare and analyze the GPS baseline components to get the differences. The results show that the ocean tidal displacement corrections have obvious effects upon GPS baseline components. Therefore, we should not ignore the ocean loading tides corrections of GPS stations in Antarctica to obtain precise and reliable results.
The current mathematical model and parameter estimation method in precise point positioning(PPP) are introduced,then the adaptive filtering by selection of the parameter weights in the parameter estimation is applied.The selecting weight fitting method is employed to determine the initial parameters and the corresponding variance-covariance components.Two numerical experiments demonstrate that the position precision can achieve decimeter to centimeter in a short time,and that the feasibility and validity of the new method can be ensured.
Satellite precision orbit determination (POD) is always one of the focuses in geodesy and geodynamics. Integrated adjustment of LEO and GPS for POD is reported. That means the orbits of GPS satellites and LEOs are determined simultaneously. At first, the mathematical description of the integrated adjustment is given. Then 6 cases are described to illuminate the contribution of LEO to GPS POD. It is shown that the improvement can reach 40 on average when 3 LEOs are included if 43 stations are used. Even if only 21 ground stations plus 3 LEOs are used for GPS orbit determination, about 5.0 cm accuracy from IGS final orbit can be achieved with only one day’s data. In this process, the along and cross components of the GPS orbit benefit much and the systematic error vanishes or diminishes. It is mainly attributed to the strengthened geometry of GPS tracking network and no tropospheric delay for space-borne GPS observations.
A new model for grid task assign is discussed. This model can decompose the complicated computational task of 3D GIS into little blocks and assign to many computers to complete cooperatively and synchronously. By assigning 3D GIS roaming function to many computers to complete synchronously, which was originally done by only one computer, such questions as poor 3D vagile speed, high request for server equipment and roaming intermission, which exist in traditional 3D GIS roam and come of large amount of 3D data and complicated operation can be solved. In this system, grid computation was combined with 3D GIS roaming technique.
The geographic data set integration is an important content of geography information science researches. The traditional spatial data integration ways cannot effectively integrate the semantics of the different geography space data, thus has brought the information acquisition and the information understanding gap for integrated data sharing and the use. An ontology is an explicit specification of aconceptualizati. It may effectively integrate the geography space data and its semantics, and may overcome the flaw which the traditional space data integration method is unable to overcome. This study based on the ontology theory, the method and in the applied research around domestic and foreign geography information academic improved ontology-driven geographic data set integration’s frame, and carried on the confirmation explanation by the concrete example.
Recently, the study of moving objects has become an importance research issue in such areas as temporal GIS and spatiotemporal database. However, due to their special spatiotemporal conditions, trajectories are different from existing GIS data, and there is still a lack of efficient data-model and index methods to support storage and queries about trajectories. In view of this, a new spatiotemporal Index, named 2-dimensional spatio-temporal index for moving objects in network (2DSTI), and a set of spatiotemporal query algorithms were proposed. The 2DSTI was easy to implement and maintain; moreover it could support many query types on pass and current. Finally, 2DSTI and its spatiotemporal queries were implemented and validated through a series of experiments.
Adaptive Human-Computer Interface (AHCI) is an important field in the domain of user-centered intelligent user interface. Based on the discussion on the theory and approach of AHCI, this paper studies the adaptive expression of spatial information and the adaptation policy of the interaction elements on GIS interface emphatically, and then it puts forward a general framework of the AHCI specific GIS.
A weight determination algorithm is proposed based on fuzzy mathematics. It carries on the synthesis appraisal of fuzzy factor, which will affect the network weight with the help of fuzzy theory. So the result synthesizes the influence of various factor, and avoids the problem of traditional static weight algorithm. Compared with the static weight algorithm, the multi-factor weight algorithm can get better result for the driver’s need actually, and the algorithm efficiency has not been influenced.
On the basis of analyzing influence of spatial accuracy and relationship to other geographic features, researching method of error propagation of linear features’ simplification algorithms is proposed, error propagation model of linear features’ simplification algorithms is developed, and the model is visualized using error ellipse, finally, aiming at different linear features, error propagation of different algorithms is analyzed and evaluated.
A method to realize automatic calibration of binocular vision system with high precision is proposed. At the first step, coded retro-reflective targets is used to realize automatic correspondence of the calibration control points in the world space coordinates and image space coordinates. Then the orientations and its systematic distortion parameters are compated according to the Weng’s camera model. At last, based on the principle that the projection of every line in space onto the camera is still a line if a camera follows the pinhole model, a new method is proposed for calibrating the residual error of random camera with high precision. Experiments results indicate that this kind of calibration method have reached successful effect.