2013 Vol. 38, No. 6
According to the character of the CCD sensors which set up on HJ-1A/1B satellites,a new imaging model which is extended from the collinear equation for linear sensor is built.In this model,quadratic polynomial and Hermite interpolation polynomial are employed for fitting the variation of the exterior orientation elements,and high-order polynomial is used as distortion model,plus engaging iteration method to solve the frequent problem,ill-conditioned system.Furthermore,we also present a method based on bipartite graph matching for correspondence points matching to meet the requirement of the distribution of control-points.The empirical results show that the proposed model compared with existing models such as quadratic polynomial and rational function model is able to achieve a higher accuracy,and requires for smaller condition number,that is sufficient for keeping pace with the demand of the research for global land cover.
A new variational model for segmenting the images corrupted by various noise is proposed.Firstly,the energy function of the presented model based on Chan-Vese model is modified,and a new auxiliary variable is introduced to integrate some fitting terms.Secondly,it is extended to convex optimization by convex relaxation.And the solution is decomposed into solving a few functional optimization sub-problems,which can be obtained by applying Split-Bregman algorithm and additive operator splitting(AOS) numerical algorithm,the effective results are obtained.Comparing with the model in which a auxiliary variable is not introduced,experimental results verify that the proposed model for segmenting the noisy images reduce computational time and has better results.
Nonlinear dimensionality reduction on hyperspectral imagery can be achieved using the isometric mapping(Isomap) method.We explore the spectral interpretations of Isomap manifold coordinates through observing and comparing the changing trends between manifold coordinates and spectral signatures.The study aims to extract desired low-dimension manifold features from Isomap manifold maps.Two cases study are designed to testify the capacity of manifold maps in extracting low-dimension manifold features.The results show that the Isomap manifold maps can be used to extract low-dimension manifold features.Moreover,the results prove that the spectral interpretations of manifold coordinates are feasible.This will be helpful for the applications of the Isomap method in hyperspectral imagery fields.
Considering the influence of the posterior and the statistic distributions of full-polarimetric SAR data,we proposed a new classification method of full polarimetric SAR data.First,the covariance matrix of polarization SAR data was converted to nine intensity quantities with normal distribution.Then,the probability of occurance for each class was calculated with iterative initial classification.Finally,the nine intensity images were classified with maximum likelihood classification method taking the probabilities of occurance for the classes into account.We applied the developed method to the ALOS PALSAR full-polarimetric data of Xunke County,Heilongjiang Province.The overall accuracy is 81.34% and the Kappa coefficient 0.84.The developed method showed higher accuracy than that from the traditional maximum likelihood classifier.This indicates that our method can improve the accuracy of classification.
A novel local saliency features of interest extraction algorithm based on visual saliency measurement theory is provided.Firstly,the focuses of attention are extracted by an iteration process,which is fitted to the attention shift mechanism.Then,we treat it as local saliency points,and the saliency edge points are got according to the combination of Canny edge graph and saliency map.After that we employ main hue,texture features of saliency points and edge orientation histogram features of edge points as feature vector for interesting objects representation.This algorithm emphasizes the salience areas in the scene,so that it can improve process efficiency and fit to human visual perceptual characteristic,otherwise this multi-channel feature describe method ensure the algorithm’s adaptive to various scenes.
A remote sensing image object detecting of aeroplane based on hierarchical adaptive part-based model is proposed.Because the present approaches take less of consideration on subtype choice and multi-resolution structure relationships,a new algorithm was put forward.First,an extern part-based model is built up.Then,we optimize its effects from subtype adaptive choice,subtype hierarchical modeling and weighted distance transform.Finally,HAPM takes model deformation and together with hierarchical structure,which greatly improve the detection result and its application.By collecting the remote sensing images from ten airports,the effects of the new algorithm also has been tested.In a consequence,the results show that the new approach is worthwhile.
We present a new algorithm based on RFM for epipolar generation between multi-source high-resolution remote sensing images.In this algorithm,an inverse RFM is iteratively built using positive RFM.Then the approximation epipolars are generated by the inverse RFM and projection tracking method.Experimental results verify that the presented method is adaptive and available and provides a feasible scheme to solve the problems for auto-matching between multi-source high resolution satellite images.
Considering of the problems of serious effects caused by vegetation on automatic landslide monitoring in close-range photogrammetry,we present a method of detecting vegetation regions in landslide images based on texture features and naive Bayes classifier.Some meaningful discussions and analysis have been done mainly for the effectiveness of this algorithm,image contrast stretching and the generality problem of samples training.Comparing with another detection method based on visual cognition features,we prove the availability and the validity of this method.The experimental results show that the vegetation detection method can almost detect the vegetation regions from close-range images and the result is satisfying.
We propose a new change detection approach based on value and shape optimized combination by analyzing robustness of several traditional change detection algorithms.It was tested in Xi’an City of Shaanxi Province to analyze land cover change from 2000 to 2009.The overall accuracy and Kappa coefficient of change detection result are 92.313% and 0.844 respectively,which outperform other methods.
The vegetation coverage is an important indication of the situation of ground vegetation,and it is the foundation data for describing ecological system,and is also the main factors affecting soil erosion.Based on 372 images of SPOT4 VEGETATION(S10) data from April 1998 to July 2008,the change trend of the yearly maximum NDVI,the inter-annual changes of yearly maximum NDVI and monthly maximum NDVI,and the dynamic change and spatial distribution of vegetation coverage in Shaanxi Province are analyzed by means of maximum value composites(MVC),one-dimensional linear regression and differential approach.The results show that the yearly maximum NDVI of Shaanxi Province is on the whole improved from 1998 to 2008,but the inter-annual change trend of the monthly maximum NDVI has a great difference among different months,the changes during the adjacent two years of both the yearly maximum NDVI and the monthly maximum NDVI have a great difference,and the vegetation degradation and improvement appear fluctuantly.The change of vegetation coverage is obviously during 1998 to 2008,and the vegetation is best in August and September.Vegetation coverage of most regions is increased,the obviously increased regions mainly appear in northern parts of Shaanxi Province(southeast parts of Yulin and north part of Yan’an),in south central of Baoji,parts of Xi’an,Shangluo,Ankang are also improved.
Taking a priori precise ZPD derived from CORS network to augment PPP can reduce the number of unknown parameters,accelerate its ambiguity resolution in PPP and shorten the convergence time.We investigate the applicability and interpolation effect of spatial regression model in which precise ZPD derived from regional CORS network are used to interpolate ZPD for user site.We also demonstrate that H1QM3 model and MLCM model are equivalent,and propose the height constraint strategy to improve interpolation precision.It is shown that the H1QX1 model and H1QM3 model are superior to other models,interpolation precision is 10 mm in the flat region and 20mm in the undulated region.With height constraint,the precision of intepolated ZPD can be improved by 60% in the undulated region.The optimal number of reference stations for intepolation is 9 or 10.
In real-time navigation and positioning of the satellite,the zenith tropospheric delay model is used to decrease tropospheric delay by requiring meteorological parameters,but this requirement reduces the accuracy of the model.Therefore,we analyze the variation of troposphere in the elevation and the horizontal respectively on the global scale by requiring the high-precision zenith tropospheric delay data which come from IGS.Based on the analysis,we put forward a new zenith tropospheric delay model which can meet the need of the real-time positioning and navigation of the satellite.The reason is that not only the new model is simple as its input parameters are day of year and location of the station without meteorological parameters but also its accuracy is best compared with several commonly used models.
We analyzed the disturbance anomalyin the global ionosphere VTEC caused by solar storm on August 1,2010,based on the GPS ionosphere detection technology.Affected by the Coronal Mass Ejection toward Earth,the solar storm had some impact on the ionosphere VTEC of North American on August 3,and the maximal VTEC anomaly was 15 TECu.Becasue of the earth rotation,the ionosphere VTEC anomaly area would move to the west along with the sun point-blank point,and the ionosphere VTEC anomaly generally occurred at local time 13:00-17:00.In addition,we analyzed the impact of this solar storm on the quality of GPS observation data and ionosphere higher order term error.The result showed that GPS observations weren’t influenced obviously,but this solar storm made ionosphere higher order term error on L1 and L2 carriers larger,and the error influence reached to centimeter level during this solar storm.
In order to achieve the centimeter-level precise orbit determination for the HY-2 satellite,the methods of orbit determination based on DORIS and SLR are discussed.We have simulated DORIS and SLR observation data for the HY-2 satellite,defined the methods and processes of precise orbit determination.In addition we have discussed the orbit determination accuracy using simulated data given different observation errors,and analyzed the impact of different distribution of ground beacons on orbit determination accuracy and consuming time.In order to obtain more accuracy results,we also analyzed the effect of two observation technique weights on orbit determination accuracy.Experiment results indicate that the optimization of beacon stations can significantly improve the orbit determination precision and save consuming time.Reasonable weights on a variety of technology measurements can make the accuracy of result to achieve the best during the integrated technology orbit determination,if we give DORIS and SLR observations 0.3 mm/s and 10 mm weights respectively,and the obit determination accuracy can achieve centimeter level.
We proposed a refinement method of solar radiation perturbation model in autonomous orbit determination of navigation satellites.With this method,the prediction errors of GPS satellites’ Ω and i could be controlled at ±40 mas level during 180 days.As a result,the tangential and normal direction precision of satellites’ orbit improved in the autonomous orbit determination when constrained the predictions of Ω and i.Finally,user range error became smaller.However,this method has small influences on radial direction of orbit as well as clock bias.
It is pointed out that the vehicle acceleration in the navigation frame affects the accuracy of SINS(strapdown inertial navigation system) directly after analyzing the effect of damp network on the system.Then,the premise and fault of a series of fuzzy damp algorithms in some interrelated documents is discussed and an improved algorithm is proposed.In this algorithm,the three accelerometer outputs are transformed and compensated by the attitude and velocity outputs of the SINS,and are denoised with a fixed point smooth filter.Finally the processed data is input to the fuzzy arbiter to distinguish whether the damp network can be used in the SINS.The simulation and actual test on sea proved that this algorithm can identify the vehicle motion and change the damp state of SINS more suitably.The 46-hour navigation data on sea shows that the system vehicle error decreased to 1 m/s from 2 m/s because the SINS choice it’s damp state properly by means of this algorithm.
Using 5 typical observational curve at sites such as Shihuiyaokou,Hongliuxia,Wozitan,Biandukou and Nanying,in the monitoring area of Gansu Province and its boundaries with Ningxia and Qinghai Provinces,the slow time-varying precursor system is constructed,the abnormity is identified,and the possible relationship between the abnormity and preparation of earthquakes around or over 6.0 magnitude are researched.The results show that at the medium and short-term phase before these earthquakes,the maximum module of latent root exceeded 1.0.This method reduced factitious or experiential identification of abnormity relative to analysis of observational curve,also reflects dynamic variation and adaptive features of precursor system structure for cross-fault deformation.
Coseismic ionospheric disturbances(CID)obviously exist when large earthquake occurs by analysis of Mw7.9 Wenchuan Earthquake and Mw9.0 Japan Earthquake.The duration of CID is short,usually starts 30~40 min before earthquake and reaches the extreme value about 10 min after earthquake.Then it disappears.The duration and amplitude of CID depends on seismic magnitude,specifically,magnitude greater and the duration and amplitude greater.In addition to clear precursory positive anomaly of ionospheric TEC around the earthquake epicenter,it’s mentionable that weak negative anomaly of ionospheric TEC exist far away from the earthquake epicenter.
During downward continuation of airborne gravity,ill-condition causes different effects to different parameters.In order to eliminate or alleviate the effects to an appropriate level,we put forward a new algorithm named regularization by grouping correction.Using the signal-to-noise ratio to assess the ill-condition effects,parameters are grouped.Regularization matrix is constructed by grouping amendment idea.Regularization parameter is selected by minimizing the mean square error.Using the simulative airborne gravity data based on the EGM2008 as true values of the gravity field,the effectiveness of the method is verified.Comparing with three other methods,the new method has higher accuracy.
Three existing methods are compared,and a new ellipsoid expansion method considered ellipsoid misalignment is proposed based on analytic geometry theory.The influence on major semi-axis variation,station latitude variation and Gaussian Plain Coordinate caused by ellipsoid misalignment is examined.According to the experiments,the results from the proposed method and generalized geodesy derivation equation method agree well,and the proposed method is more intuitive.
According to 3 typical vertical flow distribution calculation models,we deduce the calculation models adaptable to ADCP dead zone flow.On the basis of the relating coefficient between the inner coincidence precision of the models and the actual observed vertical flow,we propese the self-adaptive measure determining the ideology and method to calculate dead zone flow under different tense and flow regime.The experiment verifies the correctness of the given dead zone flow calculation models and self-adaptive optimal model determination ideology and method.
The tile access has dynamic features(server peer capability,storage device and hot tiles have dynamic features) and the Hotmap model based on the historical log information can not reflect the real system’s current global information.Reproduction and distribution will produce a huge network flow rate in a large scale distributed nodes environment.A dynamic statistics algorithm for the distribution rule of the spatial data based on P2P is proposed to resolve above-mentioned problems.The service capabilities of the service nodes are calculated in this algorithm.The node agents with good service capabilities are chosen preferentially in the group to fuse dynamic statistical information.The experimental results show that the algorithm can meet the need of dynamic statistics in large scale distributed modes environment with high efficiency.
Aiming at the similarity quantitative calculation of adjacency relations,we combine the feature model and the alignment model of similarity measures based on confirming characteristic quantity of adjacency relations,and build the similarity calculation method of adjacency relations between group objects attended to sequence difference.Finally,the method is verified by an example of usability.
Based on the master/slave model,we present a hybrid MPI+OpenMP parallel implementation for the eigenfunction-base spatial filtering on the multi-core cluster.There are two different implementations of the algorithm: one based on MPI and the other based on a hybrid parallel paradigm with MPI+OpenMP.The experimental results show that MPI+OpenMP method can cut down the process-time effectively and improve the filtering efficiency.
A new method,3D direction relation calculation based on 3D Voronoi diagram is proposed.The 27 direction relations are proposed to represent the qualitative direction relation based on the 2D four direction relations.Shape,distance and other factors play very import role in direction relation calculation.3D Voronoi diagram being an approximation of spatial object plays well in the situation of complex shape and various distances.The principles and methods of 3D Voronoi diagram for the calculation of the 3D direction relations are elaborated.The weighted normal of surfaces in 3D Voronoi diagram is used to calculate the direction relation of the non-intersecting spatial objects.The direction relation in intersection case of the spatial objects is also discussed.
We present a multi-objective land use spatial allocation model(MOLUSA) based on particle swarm optimization algorithm in order to obtain the optimal spatial land use solution.The model considers economic benefit,social benefit,ecological benefit and spatial compactness as objectives and takes the optimal land use structure,current land use conditions and land use transition rules into account as constraints.We employed the model to reshuffle land use spatial pattern in Jiayu County in Hubei Province of China.