2008 Vol. 33, No. 6
Airborne LIDAR technology has a capability to capture the ground information quickly,and has special advantages in many application fields.Filtering is one of the key processes in the processing of airborne LIDAR data.Firstly,an adaptive filtering methodology based on hierarchy model is initiated.Employing the fundament of weighted iteration and combining with the adaptive hierarchy process in the practice,the algorithm can perform the filtering of the cloud points data in the city zone.Then the result of it's practical experiment is given.Finally the qualitative filtering assessment in allusion to the algorithm proposed is discussed in detail.
Auto matching of the large scale aerial imagery over urban areas is one of the most difficult problems in the auto aertial triangulation.Through the analysis of the auto-extracting tie points strategy of VirtuoZo AAT,the vertical line observations in the large scale aerial imagery over urbans have been introduced into the traditional continuous relative orientation;and then a multi-image matching strategy with the correction of geometric distortion in the images due to the height variance has been prompted to improve the success rate of the auto-matching of images between strips.
This paper puts forward a method of generating epipolar images using the elements of exterior orientation on aerial stereo image pairs.Firstly,the strict geometry relation between the epipolar images and original stereo images is proposed;and then the feasibility and accuracy of the geometry relation is proved by the experiments using the actual aerial images at the 1∶3 000 scale.The empirical results indicate that the epipolar images generated using the elements of exterior orientation obtained by bundle block adjustment and calibrated POS data can get the familiar quality compared with the method using elements of relative orientation and the vertical parallax of corresponding image points is within one pixel,which satisfies the demand for stereoscopic observation in photogrammetry,but the ones generated using the POS data uncalibrated has about 70 pixels vertical parallax,which can not be used for stereoscopic observation.
This paper discusses necessity,possibility and technique solution of sharing symbol library in different GIS platforms.The solution is to set up a general data model of symbolic library firstly,then to design an exchange format.A symbolic library model and a draft version of XML schema of symbolic library exchange format are given.
This paper introduces four kinds of error equations about 3D GCPs,plain GCPs,elevation GCPs and pass points based on polynomial ortho-rectification of SAR images;and then a block adjustment model of SAR images is given.According to this model,program has been designed and experiment has been done on airborne SAR images of study site in Chengdu.The experimental results are feasible.
This paper presented a method to evaluate vegetation degradation in semi-arid and semihumid region.Based on eliminating irrigation land,we estimated the potential vegetation index by imitating the upper borderline of scatter plots in NDVI-climate variables feature space and appraised the vegetation degradation through comparing actual vegetation index with potential vegetation index.The degree of vegetation degradation takes on prominent spatial difference.On the whole,the degree of vegetation degradation is the most severe in the middle and the east of Gansu Province and the south of Ningxia autonomous region.NDVI tend to saturation where climate is more humider and so the evaluating result has some uncertainty.
Regularization parameter is a key parameter in nonlinear ill-posed quantitative remote sensing inversion.Taking land surface and atmosphere parameters retrieval for example,and taking Shannon entropy reduce as a quantity to express information utilization in retrieving process,a quantitative method is deduced on reasonable hypothesis.The new method is on choosing regularization parameter(prior information ratio) in iteration process of nonlinear ill-posed quantitative remote sensing inversion.On the condition of the stability of inversion,the new method takes good advantage of information of observations.The result reveals that the precision of the object parameter is improved.
Based on the analysis of the infrared radiation characteristics of different objects,the paper proposes the detection algorithm of nighttime fog on MODIS data,and the algorithm has been tested and analyzed.The result shows that the algorithm can distinguish preferably the nighttime fog from cloud,water and surface.
The essential characteristics with greater information difference of color image are interpreted in detail.A new method to process edge detection of color image is put forward based on the differential information theory of grey system.At first,it analyses to the characteristic of pixels,and confirms differential information serial of color image.Then,it distinguishes between non-edge pixels and edge pixels through computing the differential information measure,the information entropy relative measure and the differential information relative measure.The experimental result indicates that the method of edge detection of color image has certain feasibility and validity.
The paper presents an unsupervised segmentation model based on multiple level sets evolution for high resolution satellite imagery.The model can remove ambiguities occurring to multiple levels sets methods.In the numerical implementation,a novel re-initialization technique for level set functions is suggested to accelerate curves evolution steadily.Experiments show that our method can alleviate "pepper and salt effect" compared with pixel-based method,and to lessen over-and under-segmentation compared to object-based methods.
The paper introduces an extended state cellular automata(CA) model to spatiotemporal data mining(STDM).The core of the model adds numerable and uncountable attribute to the cell and intends to resolve the problem of the sparse data and large attribute information interaction in the spatial and spatiotemporal data mining tasks.The preliminary experiment shows the approach is suited for the nonlinear problems,even in the face of sparse data.They can tackle problems of previously prohibitive complexity and also improve previous approaches.The paper advises the method in combination with domain knowledge and other data mining techniques offer a chance to discover nonlinear spatiotemporal relationships.
On the basis of analysis of existing bend recognition methods,this paper gives a method and its corresponding algorithms for bend recognition with nodes of a curve,segmentation points on a circuitous curve,extreme points,inflection points,and points with unordinary rotation direction change.The experimental results can show the validation of this method.
On the basis of analyzing the influence that linear features' accuracy by simplification algorithms can be divided into geometric accuracy and attribute accuracy,aiming at curves' characteristics on geometric character and position of points change in the process of simplification,considering all the change of linear features' accuracy,position accuracy is the most important and quantitative easily,a series of evaluating figures of geometric accuracy are proposed.Several typical simplification algorithms are evaluated and some credible conclusions are given.
Based on the rules of residential areas changing with scales,a new method of matching the residential features is studied to match two features which present the same entity in the two datasets.The authors implement the method to build up the relationships of blocks and buildings of the residential features in the NFGIS of the same area.
The method to combine the GPS data from different analysis centers by SINEX files is introduced.In order to analyze the systematic error,the Helmert transformation parameters are added into the adjusment model in the data processing.The results show that the accurcy of coordinates with Helmert transformation parameters is higher than that of the coordinates without Helmert transformation parameter so as to indicate that the method can elimilate the systematic errors effectively and the systematic errors can also be estimated.
It is declared that the global geoid with one centimeter-level accuracy is difficult to be realized using Stokes or Molodensky approach.Under the assumption that the density distribution of the shallow mass layer bounded by Ω and a closed reference surface Γ that lies under the geoid G as well as the the global 1°×1° gravitational potential field model is given;an approach for determining the global 1°×1° geoid with centimeter-level accuracy is proposed;and its actual realization procedure is provided.
Based on the loading tide theory,the ground vertical deformation and the geoid undulation caused by loading of neighboring buildings are studied.Further more,the influence on elevation is also considered.The result shows that the ground vertical deformation and the geoid undulation have both reached millimeter magnitude.So it is obvious that the building loading has affected the precise engineering surveying a lot;and it must be seriously considered during the future applying.
The satellite-to-satellite tracking observations include light-of-sight-distance,light-of-sight-velocity and light-of-sight-acceleration,which reflect the various-band information of the Earth gravity field.To extract these information,the relationship between the SST observations and the Earth gravity field coefficients should be establishend.Three types of observation equations are created.And these equation could be realized with numerical method conveniently.
Both forward and inverse formulas between auxiliary and geodetic latitudes are derived with the help of computer algebra system in a comprehensive way;and also some mistakes that once were made in the high orders of previous forward formulas are pointed out and corrected.New symbolic expressions of the inverse expansions for these auxiliary latitudes are systematically given in virtue of Hermite interpolation.Their coefficients are expressed in a power series of the eccentricity e,which makes the expansions more uniform.
The paper analysed the stochastic methods and Functional models which were used to calculated the velocity field and proposed a new interpolation approach integrating two classes of models,for minimizing the residual misfit.It is shown,by comparison between results derived by various models,that the integrated approach is better than the methods based on individual model,either functional or stochastic,for interpolation of velocity field.
Through the wavelet analysis in the image classification and identification,the classification criterion of the seabed terrain complication is given.According to the criterion,the seabed terrain has been divided into 18 types.Through this method,the complication of seabed terrain is changed from qualitative classification to quantitative classification which provides better condition for the hydrographic technique design and the survey accuracy.The applicability of the method was validated with the practical test.
An new analytic algorithm is given for Beidou passive positioning.It not only has high computation efficiency,but also has high positioning accuracy by improving observation configuration.Based on the analysis of the emulation result,it can be shown that the accuracy of new method is higher than that of "3 star plus height" and other analytic solution.
A new idea is put forward that the number of difference point can adjust by a parameter factor.The authors study this method to smooth raw data and detect cycle-slip with wavelet analysis.Taking CHAMP satellite data for example,the authors get some significant conclusions.It is showed that it is valid to detect cycle-slip in GPS phase measurement based on this wavelet function;and it is helpful to improve the precision of GPS data pre-processing and positioning.
A new approach is presented by polynomial-quotient which translates colored observation noises into infinite series,and chooses the finite series,calculates the variance of dummy white observation noises.The stochastic model can be compensated with this method.An ARMA innovation model and the state optimal filter are designed by modern time series analysis method.In order to verify the validity and rationality of this method,a contrast between this method and the approach of observation-expand filter is given.The result shows that the approach can control the influences of the colored observation noises effectively.
The speed of basic fractal encoding is slow.Most of the accelerated algorithm will decrease the decoding image quality.To improve this condition,a novel algorithm is given.First,choosing an accelerated fractal encoding algorithm and the algorithm can get higher compression ration than the basic algorithm,encode the origininal image and get basic code.A residual image will be obtained by calculating between the original image and the decoding image of the basic code.The algorithm divide the residual image into 8×4 block, qualitify it by factor 12,encode the qualitify image,get the additional code. Then the additional code and the basic code constitute the image compression code.The algorithm can speed up the encoding,at the same time,it can adjust the PSNR and the compression ration easily by threshold T,so,it can expand the applied situation.
A novel information disguising method based on empirical mode decomposition is proposed.The pseudorandom sequence controlled by seeds of the SHA-256 one-way hash function is generated;and digital elevation model data for disguising is achieved by decomposing the expanded pseudorandom sequence via empirical mode decomposition(EMD).The high vision fraudulence is obtained for disguised DEM data.Furthermore,the concepts of the histogram for DEM data is also proposed;and the watermarking was reversibly embedded in the disguised DEM data by modifying its histogram.The disguised DEM data can be completely reconstructed without any distortion from the marked data after the watermark has been extracted.The secret DEM data can be recovered via the seed.The proposed algorithm owns high security.
This paper investigates the queue threshold control scheme in shared memory switches.Traditional queue control scheme employs a global threshold,which is derived from current queue length and lacks of the consideration of traffic scenarios.The presented work uses the traffic pressure as a main criterion to establish a per-queue threshold.In this sense,every output port bears the same traffic pressure.The scheme keeps the balance of the common memory in the sense of average,and relieves the demand of buffer space in times of bursty traffic.The result of simulation shows that the proposed scheme outperforms the traditional dynamic threshold in terms of efficiency.