摘要:At high frequencies,radar target can be described as returns from multiple scattering centers and the returns reveal precision geometric shape information of target.With a modest bandwidth of transmitted signal,the spacial distribution along radar radial range axis which is called range profile can be obtained from received signal.By polarization data processing of each scattering center isolated on range axis,the geometric shape and structure of the scattering center can be obtained.And the range-polarization reconstructure is obtained from multiple scattering centers.This paper studies the theory of target rangeploarization reconstucting and presents a novel method of obtaining R-P reconstructure under strong land clutter background.The results are shown by computer simulation.
关键词:target identification;polarization;high range resolution
摘要:In this paper we produce first a pair of the surface orientation maps of the scene using BPS system.Then:1.segmenting the surface orientation maps into the isolated regions corresponding to the objects in the scene.2. matching the disparities between the corresponding regions.3.determining the absolute depth of 3-D surface of objects.Finally,this paper points out emphatically how to recognize and how to analyse some kinds of the main errors:(1) the measurment error of the surface orientation;(2)the reconstructed depth error,and(3)errors caused by epipolar lines being not always parallel to scan lines,and how to correct them.
摘要:First,an analysis wavelet g(t)is proposed.Basing this analysis wavelet,we discussed the features of wavelet transform,studied the orthonormal wavelet transform and inverse wavelet transform of complex signals.Then we discussed the wavelet decomposition and composition of complex signals.Last,we proposed a fast wavelet transform algorithm to implement the signal decomposition.
摘要:Based on the singular value decomposition (SVD) of a data matrix collected by the array sensors,a new approach to space-time two dimensional spectrum estimation is presented in this paper.The noise singular vectors are got under the singular value decomposition of the formed matrix.By using the orthogonal properties of the noise space and the signal space,we can estimate a high resolution space-time two dimensional spectrum.The presented method applies to the fields such as taking bearings and capturing information of the velocities.The results of computer simulation illustrate that the new method is quite effective.
关键词:Space-time two-dimensional signal spectrum estimation;singular value decomposition;array signal processing
摘要:A new technique to identify radar targets,projection mathod,is presented in this paper.It is proved that the projection method is equal to optimal Bayes classifier under some conditions.Some simulation results for identifying two airplanes and three wire airplane models are given.
摘要:In this paper,a class of morphological bandpass filters is constructed by using morphological opening,and its properties are investigated in detail.These filters allow only the image objects with specific shape to pass through them,so that they can be effectively applied to image shape feature extraction.
摘要:The relationship between nonlinearity and correlation immunity of memoryless combining function is addressed in this paper.With the help of Walsh transform,spectral characterization of generalized correlation-immune function is studied,and the generalized correlation immunity order of Bent function is discussed.
摘要:A spectrum space approach to spectral analysis is presented in this paper.We pose some constraints on the spectrum to be estimated besides the autocorrelation function constraints.If these constraints can be related to some closed convex sets,the spectrum can be obtained by the succesive projections.Some theoretical bases underlying the proposed method are given,and its performance is demonstrated by illustrations.
摘要:This paper presents a new method for maximum entropy spectral estimation.The respective AR coefficient estimations minimize an objective function which expresses synthetically both the energy of forward and backward linear prediction errors and the orthogonality relation between the prediction errors and the signal.Simulations verify that not only good spectral-estimation performance and small frequency biases are obtained ,but the spectral line splitting phenomenon is also dissipated.
关键词:Stochastic signal processing;Maximum entropy spectrum estimation;AR model;Prediction error
摘要:In this paper,the communication system is first analysed with Grey System theory.The basic concept and model of grey relational analysis are introduced and the grey relational analysis algorithm of communication system is established.Finally,the analysis example is also given.
摘要:A univeraal switched current(ST) building block with differential output is developed.Systematic methods of constructing biquadratic SI filters by using the proposed building block are presented.From a family of the single,two or three amplifier active RC or SC prototypes,we derive a new series of SI circuits can obtair lowpass,highpass and bandpass filtering functions on a single block.These circuits have low sensitiveties,and are pure transistor structures.Therefore,they are suitable for digital CMOS process.
摘要:This paper studies some basic theories of Fourier analysis for staggered sampling seqences.They include features of Fourier spectrum,properties of Fourier transform,interpolation function in time and freqency domains,Fourier series for periodically staggered seqences and discrete Fourier transform pair for finite length staggered sampling seqences.
摘要:This correspondence states the criterion of cluster validity problem which injects validity measurement functionals into the final judgement i such as fuzzy partition coefficient and entropy.It involves the optimizing of validity functionals through the implementation of FCM algorithm to evaluate optimal clustering.In this way,a unique global minimum can be guaranteed for validity functional over fuzzy c-partition space.A clustering experiment for recognition of an ambiguous photograph of MIG- 29 fighter taken by a spy satellite is successfully developed by computer simulation.
摘要:Inference is the embodiment of human intelligence.Neural inference is a distributed inference form based on distributed rules.Typically,it is weak and closed-loop inference,and its base is interelation.The paper presents and describes some neural circuits to realize neural inference,and explains its principle with classificatory problems,
摘要:Synchronization is a key technique in burst communication systems.Its rapidity,reliability and bigh antijamming are required.A new method of SAWTDL rapid synchronizations is introduced in this paper,thus satisfying the requirement.
摘要:A new approach to parameter estimation for Gaussian random field images is presented.It seeks to determine the minimum relative entropy estimates of the parameters using simulated annealing technique.Examples on the application of this algorithm are also given for estimating the parameters of the mixtures of M(M is a posivive integer) univariate Gaussian distributions.
摘要:Two new methods are proposed for determining the order of a Gaussian MA model from sample autocorrelations.Simulation examples show the efficiency of the new methods.
关键词:MA model;Order determination;singular value decomposition
摘要:A new spatial-temporal adaptive processing,called Generalized Power Inversion(GPI),is proposed to suppress ground clutter for airborne radars based on a side phased array antenna.First,a multi-channel Doppler filtering operation is carried out over the output data of each column subarray,then Generalized Power Inversion is applied to process the three Doppler output channels.Theoretical analysis and computer simulation results show that the GPI approach has tolerant ability in the presence of random amplitude and phase errors existing in the array elements.
摘要:The variance of maximum likelihood function of DOA(Direction of Arrival)is derived.According to the relation,the SNR resolution threashold formula of maximum likelihood estimation of DOA is deduced.from which the SNR resolution threshold is analyzed for two sources of different correlativity with the application of uniformly spaced linear array.