2013 Vol. 38, No. 3
A modified orbit algorithm with sparse ground control points (only three points for high accuracy geo\|rectification of space\|borne SAR image is proposed, which is based on range\|doppler position model and high precision orbit ata of GPS. The algorithm decreases the number of ground control points through simplifying the orbit parameter odel, so it can improve the work efficiency on the basic of high accuracy correction.
Levees are constructed to protect coastal cities from typhoon, flood and sea tide. Since the stability of the levees is important during the extreme weather, it is necessary to monitor its deformation regularly. Repeat\|track space\|borne SAR images are useful for environment monitoring, especially ground deformation monitoring. Persistent scatterer technique is carried out to monitor the deformation of the levees. To make the subsidence map, we use 24 scenes recorded from Envisat satellite. The experimental results show that the levees around Pudong airport and Lingang town which were built on the land reclamation suffered serious deformation.
An improved SIFT operator based on the theory of zero\|crossing is proposed for automatic remote sensing image matching. Firstly, in the improved SIFT algorithm based on zero\|crossing theory, the array of scale space pyramid zero\|crossing point detection is constructed. Then the search window around the center pixel for the zero\|crossing point in the adjacent three layers from the array of scale space detection pyramids is established. If the result of zero\|crossing detection is greater than the detection threshold, the center pixel will be marked as a feature point. Feature extraction based on the theory of zero\|crossing in Gaussian scale space introduced the image geometry feature in scale space detection, resulting in higher repeatability and more stable characteristics. The experiments carried with close\|range digital stereo image which has large rotation angle and low altitude aerial stereo\|image showed significant increase in the aspects of feature repeatability rate, correct number of corresponding points and matching correct rate with the improved SIFT operator applied to automatic remote sensing matching image.
We proposed a hybrid detector based on spectral matching and tensor analysis, which is designed for hyperspectral and high resolution remote sensing images. Firstly, a spectral matching is performed in vector space using adaptive coherence/cosine estimator (ACE). Then, the result pixels of whose spectral are similar with targets spectral are further processed by support tensor machine (STM) to detect the real targets from the background pixels. The experimental results on CRI dataset demonstrate that the proposed approach could obviously reduce the processing time and improve the targets detection precision.
A batch\|parallel pyramid\|building algorithm based on the MapReduce framework is proposed.The formal description of the pyramid\|building task as well as the decomposition algorithm is given in the first place.Then, the details of the processing steps in the map and reduce phases are depicted. The experimental results show the feasibility, efficiency and scalability of the proposed approach. The tile pyramid building on massive remote\|sensing data, which is too complicated to be done on a commodity computer, can be efficiently accomplished on a server cluster. Furthermore, the proposed algorithm can be used as the basic framework for processing massive remote\|sensing images, and be applied to promote the efficiencies of some other remote\|sensing processing algorithms, e.g. the feature extraction, the image coaddition, the tile image delta,etc.
The impact both satellite three\|line array CCD and control geopositioning because of drift angle correction and residual drift angle is analyzed.The method that eliminates impact of residual drift angle in the control geopositioning of satellite three\|line array CCD is proposed. The experimental result shows that the control geopositioning of satellite three\|line CCD has great impact because of residual drift angle. It must be corrected in practical applications.
For compensating the shortcoming of multibeam sonar image in resolution and that of side scan sonar image in position, and meeting with the requirement of acquiring high\|quality seabed relief image in practice, a new method, by merging MBS image and SSS image and forming high\|quality seabed relief image, is presented. Firstly, speeded\|up robust feature (SURF) algorithm is studied deeply and an improved algorithm ISURF is given out. The ISURF algorithm obviously improves the matching accuracy and decreases the matching time of MBS image and SSS image relative to SURF algorithm. Then, four image\|fusion algorithms are studied. By comparing and analyzing in virtue of independent evaluation parameters and associated evaluation parameters, the optimum image fusion method, which is Laplace Pyramid method or wavelet transform method, is given out for fulfilling the fusion of MBS image and SSS image. Based on above researches, a complete fusion procedure of MBS image and SSS image is presented and used for forming a high\|quality seabed relief image. These methods depicted in this paper have been proved by experiments.
Aiming at the problem in the conventional GPS single epoch data processing if no priori information, a new algorithm for solving GPS integer ambiguity using single epoch data was proposed. It used one epoch code observation to calculate initial ambiguity by least square method. The original search space of ambiguity was constructed by using standard deviation of the initial position. The ambiguity transformation by using two different linear combination of expansion wave was carried out to make the original search space of ambiguity smaller. After the new search space of ambiguity was confirmed, ambiguity N1 and N2 were solved by means of the inverse linear transformation. Furthermore, ambiguity function method was used for the searching of the true value using single epoch data.Two sets of data with different baseline length was used for analysis and test. The experimental results show that the proposed method was feasible and reliable.
We propose a new method for ambiguity fixing between network RTK reference stations, which can improve the ratio value and success rate for the ambiguity test when ambiguity fixed needs long time passing the ratio test. Firstly, wide\|lane and L2 ambiguity, with its variance\|covariance matrix are formed from the original ambiguity by integer transformation. Secondly, LAMBDA method is used to sequentially fix the transformed block ambiguity. The results show that the new method not only can fast and reliably fix the wide\|lane ambiguity, but also improve the ratio value and success rate when ambiguity fixed correctly,which reduce the initialization time of ambiguity fixing.
An exponent fading factor adaptive Kalman filter algorithm instead of standard Kalman filter is introduced to improve the ground\|based GPS tomographic method of wet refractivity. Two different periods of GPS observation data and meteorological data from Hong Kong CORS are processed to determinate the 3D distribution of wet refractivity with tomographic method. The results show that this method can retrieve the wet refractivity vertical profile and 3D distribution precisely within a long tomographic procedure without any divergence.
From the engineering point of view, we propose a novel dual sink\|based continuous disaster tracking (DCDT) algorithm, which consists of a static sink and a mobile sink guided by GIS. The two sinks can cooperatively work in an intelligent model in which the mobile sink can calculate the right position to collect the information from boundary nodes; in which the static one can reduce much message transmission as well as energy consumption based on centroid algorithm. The simulation result shows that our solution can distinctly prolong the network life time than single sink network in continuous object tracking.
A heuristic algorithm integrating simulated annealing and Voronoi long edge guided optimization is proposed for large scale vehicle routing problem. The simulated annealing strategy is employed to accept some bad neighborhood solutions in local search. Voronoi long edge guides to improve the unreasonable local structure in routes. A large scale vehicle routing benchmark instances is designed to test the proposed algorithm. The results show that the proposed algorithm can provide high quality solution in a short time.
Link dynamics of vehicular ad hoc networks (VANETs) have great difference with that of the traditional mobile ad hoc networks (MANETs) due to the impact of simulation scenarios. Based on Manhattan mobility model, the influence factors of vehicular nodes in urban environments are analyzed, including traffic light and drivers’ behavior. Analytical expression of expected link lifetime is derived. The impact of traffic light and drivers’ behavior to the expected link lifetime is analyzed. The simulation results verified the accuracy of the analytical results.
Currently, in the highly dynamic and heterogeneous P2P systems, load balancing, as one of the core issues of performance improvement, faces increasingly serious challenges.Based on Gossip protocol, we use topology maintenance messages among nodes to exchange load information periodically, present a load balancing solution based on the distributed genetic algorithm. This solution evaluates the fitness with the physical network topology to minimize the cost of virtual nodes migration. Sub\|populations are processed in parallel by P2P nodes to overcome the disadvantages of excessive computation of the genetic algorithm. The experimental results show that our approach has made great improvements in the speed and the quality of solution.
In order to map the semantic heterogeneity among the spatial features, and implement the integration of spatial data, we presented a method of building ontology and applied an algorithm to map the semantic of spatial features between the charts and topographic maps. In the process of ontology construction, we established the concepts trees of domain ontology as well as the rule constraint among the concepts by manually, and automatically extracted the application ontology from the spatial features. For the semantic mapping among the spatial features, we proposed a semantic mapping algorithm based on concept depth constraint on rules and edit distance. The algorithm calculated the spatial features similarity by means of the method of semantic and syntax from both charts and topographic maps, and covered the shortage of which the HowNet only calculated concepts similarity not an instance. Finally, an example to validate has been given. It is tested that the proposed method of ontology construction and ontology mapping algorithm has superior usability, could achieve dynamic association on the level of ontology among the spatial features, and provided a new approach for spatial data integration from the different domain.
The problems existed in the traditional methods of Web services classification are analyzed, the concepts of manifold and manifold learning and the purpose of introducing the manifold learning into the Web services are described. The algorithm for the visualization and classification of geo\|spatial Web services(GWS) based on manifold learning is proposed.During the process of dimension reduction, the similarity between GWS is preserved and the data manifold is unrolled. In order to improve the precision of classification, we gain the mapping rule from the GWS to the 2D data and the initial number of clusters according to the visualization of 2D mapping data. The experimental results prove the validity of the improved visualization and classification algorithm for GWS proposed in this paper.
Based on Web mining technologies, we examine the SDI user intensions including the background of domain users, access behaviors and the relationships of cartographic rules. A Web mining method is presented to construct the self\|adaptive domain space model and to analyze the user intension behaviors based on the service log documents focusing on the user preferences of spatial scale, feature semantic and feature object. By the analysis of the real service log information verifies the feasibility and effectiveness of this Web mining method.
Through integrating atmospheric dispersion model, GIS spatial analysis and population dasymetric mapping model (PDMM), we firstly conducted high\|resolution simulations of air pollution concentration and population spatial distribution. Then, a model for assessing population relative risks of air pollution exposure (MAPRRAPE) was proposed and implemented for spatial zoning of population exposure to SO2 at various levels by taking Tarrant County as a case. The results show that the population exposure to SO2 for Tarrant County in 2000 detected by air pollution concentration method obviously differs from those produced by MAPRRAPE.While the results disclose the defect of the air pollution concentration based population exposure, it thereby conversely confirms the significance of MAPRRAPE in decision making for preventing and controlling regional air pollution exposure.
We detail the implementation of a new solution—increment transmittance mode, for the exchange of update information from the spatial data producers to the specialized departments. In order to transfer update spatial data with the increment transmittance mode, the incremental spatiotemporal changes of the spatial object should be confirmed. A new hierarchy model with four computing operators is proposed in order to self\|contained distinguish the incremental spatiotemporal changes of the spatial object. The model confirms the change types with the whole object delta of the spatial object, the size, dimension, shape and position delta of the spatial object. By computing results of the four operators, the basic changes of the single and multi\|features can be given. The complex incremental spatiotemporal changes can be composed based on the basic changes.
We describe the principle of QTM sphere solid partition based on great circle arc octree and unequal octree and the features of their grid system, derive the volume formula of grid cells, discuss the size of the volume deformation among the grid cells at the same layer and in the same column，and compare them with degenerated octree subdivision on the shape deformation, and explain the sphere octree encoding and decoding methods. The results show that， at the full use of great circle arc characteristics, global 3D\|grids based on QTM sphere octree of great circle arc have a simple shape, arranged in neat and moderate deformation. It can be used for the organization and management of spatial data of integrated earth.
Considering the importance of PSHA model in earthquakes forecasting, by using the catalog in the 200 years in China, we tried to improve the original PSHA model with combining seismic energy distribution model, and computed the probability of earthquakes in a specific area in China with time\|span T based on both the improved PSHA model and the original model. Finally the experimental results show that the improved model is more efficiency and reliable than the original one.
We present a new method for simplifying terrain using terrain significance information (TSI) index from DEMs. Firstly, TSI index of each grid cell based on DEMs is extracted. Secondly, critical elevation points are extracted from original DEMs for constructing target DEMs with proper TSI index threshold values. Finally, 1∶50 000 DEMs taking as reference target, and very important points (VIP) and three\|Douglas\|Peucker (3DD\|P) being as verify methods, the terrain simplification effect of the DEMs reconstructed with the method is analyzed and evaluated. The results show that the terrain simplification method based on TSI index is not only with better numerical precision, but also effective in reducing the terrain distortion by keeping terrain skeletons to meet the demands in different levels of multi\|scale terrain digital modeling.
3D point pattern of cluster analysis is presented based on 3D Voronoi. 3D Voronoi cell is used to represent the spatial region that the spatial point effects. Through a quantitative description of the spatial parameters about 3D Voronoi cell there exists the potential to distinguish the weight and effective quantity of each point in the 3D space. Spatial neighborhood relationship among points is extracted according to 3D Voronoi cells to delimit the candidate points that will be clustered. The method illustrates segmentation and cluster distribution of 3D points based on the underlying density and spatial relationships, and actual analysis is imposed on the point cloud of 3D rabbit (Bunny). The ability to make quantitative description of each 3D Voronoi cell gives insights into spatial controls and cluster process on 3D points.
In the automatic river generalization, a double line river need to be transformed into a single line river sometimes. An improved algorithm is given to solve this problem. First, the medial axis of a double line river is extracted based on Delaunay triangulations, and the double line river is splited to segments by its skeleton. Then, spatial graphic conflicts are detected by means of buffers of the small\|scale segments’ skeletons. At last, the spatial graphic conflicts will be solved and the generalization result will be corrected.
To improve the efficiency of data query and rendering speed of navigation map display, it is necessary to organize the display objects of map into structures. A dual level grid structure of fixed rendering scope for multi\|scale map database was proposed, and based on this structure we designed a forecast loading strategy to improve the efficiency in directional moving navigation, and thus the average data loading time can be reduced. In the grid interior processing, a vision retrench rule was presented based on ocular distinguish characteristic. Using this rule, we retrenched the display objects inside the grids by appending, simplification, and merging, and thus we reduced the data amount in the grids and improved the rendering efficiency with map vision effect still maintained.
Due to the iterative algorithm exists depending on initial value and easily divergent for lightning location, we propose a three\|dimensional grid search algorithm using for positioning and calculating of intracloud lightning . The procedure of the three\|dimensional grid search algorithm to find the lightning location is introduced. The lightning space range is defined according to the prior condition. Through dividing space grid reasonably and using multi\|level grid search strategy to calculate the cost function value of each mesh point, the most appropriate solution of the cost function is the maximum likelihood estimator of lightning position. The validity and stability of this method is confirmed by the simulation and numerical analysis of a lightning detection network data .When the grid is divided reasonably and properly, the results must be close to the true solution after multilevel searching.
We compare, both theoretically and practically, bi\|phase inter\|pulse coded pulse train radar waveforms containing different pulse shapes. These waveforms can be employed in the Wuhan ionosphere comprehensive sounding system (WICSS). Compared with bi\|phase inter\|pulse coded rectangular pulse train waveforms, bi\|phase inter\|pulse coded Gaussian pulse train waveforms can achieve higher spectral effciency and range accuracy, better range resolution, and the same Doppler resolution. It is specially suitable for ionosonde. The simulating and experimental methods used in this paper are also suitable for other radar waveform comparison.