2006 Vol. 31, No. 2
A high-accurate outline detection algorithm based on line diffusion function model(LDFM algorithm) has been proposed in this paper.Then analysis of extraction accuracy and blunder eliminate and noise constrain have been done.Finally this algorithm has been used to extract aircraft outline from image sequences and satisfactory results has been obtained
The digital distortion model(DDM) intends to establish general theory and methods of inspecting the distortions for all kinds of imaging system.In this paper,the theory and methods of establishing DDM have been also expatiated.The ways about getting the exterior and inner elements also have been introduced.In order to validate the theory,image achieved with fisheye digital camera has been tested.
This paper is based on "strict geometric model based on affine transformation for remote sensing image with high resolution",amending the strict geometric model of the calculation for the position and orientation parameters of the remote sensing image with high resolution.According to condition of CCD scanning Earth,amend and get a more exact geometric model,whose computation formula is inferred.Experiments demonstrate the new model is more strict and reliable,though it has the same effect as the(original) model.
In this paper,a new method,aerial images texture classification based on naive Bayes classifiers,is proposed.It is concluded that the new method performs better in overall classification precision than naive Bayes classifiers on the basis of the experiment results.
Based on the analysis to current algorithms,some new approaches are proposed and experimented.Similarity measure based on spectral polygon considers both albedo and wavelength at the same time so its precision is higher than other methods.Spectral similarity measure based on multi-scale wavelet transformation can combine similarity measure with dimensionality reduction,also have high precision.By experiments it proves that these new approaches can be used to spectral feature-based hyperspectral RS image retrieval effectively.
Transects fractal dimensions of urban land surface temperatures image,NDVI image and principal component pixels decomposition analysis image are computed in this paper.The features of three images and their transects fractal dimensions,land surface temperatures relation with vegetation are analyzed.The study shows that the more homogeneous land covered,the lower land surface temperatures fractal dimensions and according to vegetation derived,principal component pixels decomposition analysis model is more effective than NDVI.
In this paper,the land surface temperature(LST) of Beijing area is retrieved from Landsat TM thermal band data utilizing a radiative transfer equation and the urban heat island effect of Beijing and its relationship with land cover and vegetation index are discussed.The result of the LST shows that the urban LST is evidently higher than the suburban one.Thought comparison of LST,the authors find the average urban LST is 4.5℃ and 10℃ higher than the suburban and outer suburban temperature,respectively,which demonstrates the urban heat island effect being prominent.Prominent negative correlation between LST and NDVI is found in the urban area,which suggests the low fractional vegetation cover in the urban area is the main cause of the urban heat island.
The simulated results are used to correct(normalize) the radar backscattering from each data takes.The correlations between normalized radar backscattering and rice biophysical parameters are determined.The results show that ASAR data could be used to estimate rice parameters.
In this paper,the technique of using GNSS-R signals to retrieve ocean surface physical quantities has been discussed in detail.Based on the research work that has been achieved in the world,the theoretical model and the method of retrieving ocean surface variances have been expatiated systemically,and the retrieve flow-chart has been introduced.
The energy edge extraction method for super-spectral images is based on the characters of energy similarity and correlation between super-spectral signals.The energy edge map shows that the main edges between different ground objects can be detected,and these edges are not sensitive to noise.In comparison with other edge extraction methods,energy edge method is much more effective.
A method based on local HSV image and shape of object to recognize object is proposed for robot tracking.After the color segmentation,use the knowledge of shape of objects to recognize objects.The robot tracking result testifies the availability of the method.
In this paper,a blind watermarking algorithm for digital image based on wavelet-packet transform is introduced.This algorithm manipulates the original image by wavelet-packet transform,then adds watermark onto the important wavelet coefficients located in the high frequency subband based on human visual system.The experiment results show that the watermarking is robust and practicability to the common image processing techniques.
A method is developed to determine the location of a pixel in a SAR image in this paper.This technique utilizes the spacecraft ephemeris data and the characteristics of the SAR data collection system to produce an estimate of the latitude and longitude of an arbitrary pixel.This approach has an advantage over previous techniques in that it requires no reference points and is independent of spacecraft attitude knowledge or control.Test is conducted using ERS1 SAR imagery,comparing predicted feature location with the location as determined by high precision area maps.Results indicate an accuracy of 50 m is attainable with this method.
It's necessary to discuss the topology model of spherical surface raster space.Firstly,the basic topological issues(about) spatial object are given in the context of adjacency.Secondly,properties of Jordan curve are discussed in spherical surface raster space.Then the application on topological paradox is analyzed.
According to principles of cartographic name placement on line objects,the name placement rules on city streets are deduced in this paper.And,a novel approach is presented to automatically convert the labels of the annotation layer into the name attributes of the associated spatial objects.The paper presents an automated reasoning strategy and mathematical algorithms on the discrete placement of street names.At last,these algorithms are realized on the Arc/Info platform,and a case study is given.
In this paper,the definition and construction algorithm of layer-structure of speckle data are proposed based on the measure of information and structured selection of speckle-data as well as the spatial distributing and connection,with modeling the selection processing.The subjective influence as well as instability has being released or decreased,with improving the quality and effect of the generalization.
The article proposes the model which can be used to identify the bends of soft polygon,meanwhile,defines the convexity and concavity,systematically constructs morphological index,strictly expresses adjacent relation of the bends,at last,proposes the three kinds of means simplifying the shape of soft polygon.
A new generalization model of road networks based on genetic algorithm is proposed.Experimental results indicate that the principle of the method is correct and it performs well for keeping the practicability and spatial distribution characteristics of road networks.
The concept of invariant moments is given.Then on the basis of the shift invariant moment and rotation invariant moment,the formula of that how to calculate building model parameters is presented,and the coordinates of the feature points of the building are calculated using the parameters of building model.As shown in the examples in the paper,the building model is reconstructed from airborne laser altimetry points cloud data based on invariant moments.Finally,the accuracy of the method is discussed and analyzed.
AMP satellite.The model that resolves the geoid with an accuracy of better than 4 cm at a resolution of 700 km half-wavelength.By using the degree errors of geopotential coefficients difference to compare the model CDS01S with EIGEN3P,EIGEN1S and EGM96,the result indicates that the CDS01S coefficients are most close to EIGEN3P.Comparing the accuracy of geopotential coefficients in these models,the result indicates the accuracy of coefficients in CDS01S is higher than EGM96.The geoid undulations of CDS01S and GGM01C up to 30 degrees are calculated,the result indicates the standard deviation is 4.7 cm between them.
This paper presents a new integrated project,which uses strapdown sensors and fuses the information of the attitude of strapdown star tracker(SST),the position and velocity of the high speed GNSS and the corresponding information of SINS to filter.Meanwhile the design of the simulated real-time GNSS/ST/SINS platform has been realized.The result shows that the system is robust and reliable.
The extended Kalman filter is one of the most widely used methods for tracking and estimation of non-linear systems through linearizing non-linear models.In recent several decades people have realized that there are a lot of constraints in application of the EKF for its hard implementation and intractability.In this paper a new estimation method is proposed,which takes advantage of the Unscented Transformation method thus approximating the true mean and variance more accurately.The new method can be applied to non-linear systems without the linearization process necessary for the EKF,and it does not demand a Gaussian distribution of noise and what's more,its ease of implementation and more accurate estimation features enables it to demonstrate its good performance in numerical experiments of satellite orbit simulation.
The traditional set operator has been extended firstly,the FIS-ES algorithm has been presented on the basis of extended set operator.The detailed analysis about the complexity of algorithm is done in theory and experiment.It shows that the new algorithm has advantages such as more efficient and less space used in the lower minimum support condition.It is a good complementarity for Apriori algorithm.