40 results about "Doppler bandwidth" patented technology

The invention discloses a DCS algorithm-based satellite-borne video SAR (synthetic aperture radar) imaging processing method. The method includes the following steps that: 1, satellite-borne SAR beam bunching mode echo simulation data and corresponding parameters are read; 2, Doppler bandwidth and synthetic aperture time required by frame images are determined; 3, a data partitioning mode is determined, and data are partitioned; 4, de-sloping processing is performed on frame segment data in an azimuth direction; 5, signals obtained after the de-sloping processing are subjected to azimuth-direction Fourier transform, multiplied by an CS factor, subjected to range-direction Fourier transform, and multiplied by a range-direction compensation factor sequentially; 6, the phase difference between the center slope distance of each frame of data and the center slope distance of the entire echo data is compensated; 7, range-direction Fourier transform, azimuth-direction compensation processing, azimuth-direction inverse Fourier transform are performed, so that an imaging result can be obtained; 8, step 4 to step 7 are repeated, so that all frame images can be obtained; and a video SAR product can be obtained. With the DCS algorithm-based satellite-borne video SAR imaging processing method of the invention adopted, the resolutions of the frame images can be consistent.

The invention discloses a high-resolution imaging method for radar based on multi-beam scanning, and is applied in the field of radar imaging. The method comprises the steps: forming a plurality of received subbeams through the digital beam forming technology, and enabling the received beams to employ a folding-fan working mode for the scanning of a front view region, thereby reducing the front view Doppler fuzzy region. Meanwhile, the bigger Doppler bandwidth is obtained through the echo overlapping of multiple beams, thereby improving the target azimuth resolution of a big front view region of an airborne platform, and achieving the large-range high-resolution imaging of the big front view region of a radar. The method proposed by the invention can be used for solving a problem of inherent conflicts between the resolution and the front view imaging blind region in the conventional Doppler beam sharpening technique.

The invention provides an SAR motion target detection method based on Doppler spectral analyses. The SAR motion target detection method based on Doppler spectral analyses is used for detecting ground motion targets in an SAR picture, and mainly solves the problem that an existing detection method is not high in detection probability so as to improve detection probability. According to the method, rectangular neighborhoods of scattering points in the SAR picture are chosen, clutter lock is conducted on Doppler frequency spectrums in each neighborhood to acquire Doppler center frequency, and then standard deviations of frequency spectrums after clutter lock are calculated to represent Doppler bandwidths. Traversal is conducted on all scattering points in the SAR picture, the scattering points are processed in the same way, statistics is then carried out, a measure is constructed through obtained statistics to ensure that the detection boundary of the measure is an oval, and a detection threshold is determined according to constant false alarm rate (CFAR). Because the detection boundary is an oval, on the premise that the constant false alarm rates are the same, oval detection boundary detection probability is larger than rectangle detection boundary detection probability, and target missing detection does not occur easily.

The invention discloses a subaperture wave number domain imaging method for a squint sliding spotlight SAR (Synthetic Aperture Radar). A traditional sliding spotlight SAR subaperture method is improved aiming at squint conditions. The subaperture wave number domain imaging method comprises the following steps: firstly, overlapping a whole aperture and dividing subapertures; taking an extended wave number domain method used for the squint conditions as a subaperture basic imaging method; and finishing azimuth processing in a distance Doppler domain by adopting an improved BAS method to splice the subapertures, so as to obtain a whole aperture image under the squint conditions. With the adoption of the subaperture wave number domain imaging method, the problems of the squint sliding spotlight SAR of large data size, non-zero azimuth Doppler center and oversized Doppler bandwidth are solved, and the practical value is larger.

The invention discloses a spaceborne sliding bunching MIMO-SAR imaging method based on multi-frequency subband concurrency. The difficulty and the cost of a radar system are reduced, and at the same time two-dimensional high resolution, wide mapping band and high signal to noise ratio are acquired. According to the realizing process, a channel sampling position error correction method is proposed; a multi-subband orientation de-skewing method is proposed; the Doppler bandwidth of a signal is reduced to be less than the equivalent PRF of a system; after orientation de-skewing, the distance course difference among subband signals can be easily compensated; and a traditional frequency-domain subband splicing method is improved, so that the method avoids distance-to-interpolation, has no special requirements on system parameters, and has higher adaptability.

The invention provides an azimuth parameter obtaining method for a satellite-borne synthetic aperture radar in a sliding bunching mode. An azimuth initial imaging belt length, an antenna beamwidth, an initial squint angle, a terminate squint angle and a target slant distance are utilized to establish a reference slant distance model corresponding to a beam rotation center; reference slant distance data are obtained, and an effective irradiation time model corresponding to a point target is established according to the reference slant distance data; effective initial irradiation time and effective terminate irradiation time of the point target are obtained, and an effective Doppler frequency and Doppler bandwidth model corresponding to the point target is established according to the effective initial irradiation time and effective terminate irradiation time of the point target; initial effective Doppler frequency, terminate effective Doppler frequency, the Doppler bandwidth and the resolution declining coefficient corresponding to the point target are obtained, an azimuth resolution model corresponding to the point target is established according to the initial effective Doppler frequency, terminate effective Doppler frequency, the Doppler bandwidth and the resolution declining coefficient corresponding to the point target, and the azimuth resolution corresponding to the point target is obtained; when the azimuth resolution is larger than a resolution index requirement, the maximum azimuth imaging belt length datum is output.

The invention relates to the technical field of radar, and belongs to a synthetic aperture radar signal processing method, in particular to an azimuth signal preprocessing method. The azimuth spectrum of each sub-beam echo of the single-phase central multi-beam synthetic aperture radar system is sampled by the pulse repetition frequency (PRF) and partially overlaps with each other on the frequency axis, which can be processed in the frequency domain, using FFT frequency shifting and other means, Synthesize a broad Doppler spectrum. The azimuth echo data with a wider Doppler bandwidth is synthesized after being processed by the signal processing method, so that the azimuth resolution can be improved after azimuth compression processing.

The invention belongs to the field of radar technology, and discloses an azimuth multichannel SAR signal adaptive reconstruction method. The method comprises the steps of acquiring a two-dimensional spectrum of echo data which are received by each channel; obtaining a corresponding multichannel data matrix by means of the two-dimensional spectrum and obtaining a corresponding covariance matrix for each base band frequency in a base band frequency range; constructing a corresponding first guiding vector at each Doppler frequency in a Doppler frequency range, and furthermore calculating N Doppler spectrums at the Doppler frequency by means of the first guiding vector and a covariance matrix which corresponds with each base band frequency, determining a base band Doppler center and an effective Doppler bandwidth by means of N Doppler spectrums at each Doppler frequency, obtaining a fuzzy component number according to the base band Doppler center and the effective Doppler bandwidth at each base band frequency, and a signal matrix at the base band frequency is reconstructed by means of the fuzzy component number. The azimuth multichannel SAR signal adaptive reconstruction method can keep integrity of the Doppler spectrum and improves signal-to-noise ratio of the reconstructed signal.

The invention discloses a land-sea clutter classification method based on time domain and frequency domain multi-features. The method comprises the following steps: 1, calculating normalized power variance, Doppler bandwidth, Doppler spectrum entropy and Doppler spectrum peak power ratio of echoes of a single radar spatial resolution unit; 2, generating feature data and category data of ground clutter and sea clutter based on the step 1; 3, constructing a BP neural network; 4, calculating normalized power variance, Doppler bandwidth, Doppler spectrum entropy and Doppler spectrum peak power ratio of each spatial resolution unit in a radar detection scene; and 5, performing judgment according to an output result. According to the land-sea clutter classification method based on time domain and frequency domain multi-features, land and sea clutter classification of different motion platform radars can be realized.

The invention discloses a method and device for determining the effective Doppler bandwidth of a fuze receiver. The method comprises the steps of dividing a target into multiple target scattering surface elements after geometric modeling is conducted on the target, calculating the Doppler shift of each target scattering surface element and the echo power generated by each target scattering surface element under the missile target encounter condition, drawing a Doppler frequency spectrum diagram of the target according to the calculated Doppler shift and echo power of each target scattering surface element, and cutting out a spectral line range within which echo power is larger than set threshold power from the drawn Doppler frequency spectrum diagram and determining the effective Doppler bandwidth of the fuze receiver according to the cut-out spectral line range. By the adoption of the method and device, estimation of the effective Doppler bandwidth of the fuze receiver can be achieved, so that the fuze receiver can be designed based on the estimated effective Doppler bandwidth of the fuze receiver during fuze slideway testing, and then fuze slideway testing is facilitated.

The invention relates to a ship target interference detection method based on a satellite-borne Ka-band SAR (Synthetic Aperture Radar) system, which comprises the following steps of: scanning orthogonal signals of a pulse repetition period in an imaging area by using a Ka-band transmitting antenna to realize coverage of large width and large Doppler bandwidth; before pulse emission, a Ka wave band receiving antenna is used for carrying out parallel receiving and digital beam forming recovery on echo signals with large breadth and large Doppler bandwidth, and direction sources of interference signals are obtained; receiving data of a Ka-band receiving antenna is subjected to weighting processing to control a direction function, a nulling technology is used in the direction of an interference signal, the interference signal is suppressed, and interference caused by a sidelobe strong target of the Ka-band receiving antenna is suppressed at the same time; and obtaining an interference SAR image of the imaging area by using the Ka-band SAR cross-rail baseline, extracting elevation winding distribution information, and detecting and identifying the ship target according to the elevation winding distribution characteristics of the ship target. The ship target detection and recognition performance can be improved.

The invention provides a geosynchronous orbit SAR satellite burst imaging work method which comprises the following steps: in a first step, an orbit of a geosynchronous orbit SAR satellite is designed according to system task requirements; in a second step, based on orbit parameters, a geosynchronous orbit SAR satellite wave beam ground speed change law and an SAR satellite Doppler bandwidth change law are obtained via simulation analysis, and a geosynchronous orbit SAR satellite orientation resolution change law can also be obtained; in a third step, based on the geosynchronous orbit SAR satellite wave beam ground speed change law and a geosynchronous orbit SAR satellite burst imaging work mode, a work subaperture timing sequence is designed, a geosynchronous orbit SAR satellite full orbit work timing sequence can be obtained, and continuous observation of a full plotting band can be fulfilled. Application of the method disclosed in the invention can help improve application effectiveness of a geosynchronous orbit SAR satellite, and application prospects of the geosynchronous orbit SAR satellite can be widened.

The invention discloses a synthetic aperture radar (SAR) azimuth signal reconstruction method. The method comprises the following steps of according to a first channel transfer function matrix, determining a signal to noise ratio (SNR) amplification factor of the first channel transfer function matrix, and according to the SNR amplification factor, determining an equivalent Doppler bandwidth; according to the equivalent Doppler bandwidth and the first channel transfer function matrix, determining a second channel transfer function matrix corresponding to each Doppler frequency point; according to the second channel transfer function matrix corresponding to each Doppler frequency point, a pre-measured aliasing frequency spectrum and an original frequency spectrum corresponding to the aliasing frequency spectrum, determining a frequency spectrum filter extracting a Doppler frequency spectrum. The frequency spectrum filter is used for extracting the Doppler frequency spectrum in an azimuth signal. The invention simultaneously discloses a SAR azimuth signal reconstruction apparatus.

The invention discloses a segmental variable repeat frequency model sampling uniformity digital processing method and a segmental variable repeat frequency model sampling uniformity digital processingdevice. The method includes steps of step S1, confirming every sampling position after uniformity sampling according to the working time and sampling rate under every sampling state; step 2, performing short-time cutting on echo data according to the sampling rate and an instant Doppler bandwidth to obtain every sub-section echo data; step S3, resampling every sub-section echo data after short-time cutting to the uniform equivalent sampling position through a special interpolation treatment; step S4, performing distance and time alignment on the data after orientation resampling. According tothe feature of the segmental variable repeat frequency signal, a signal uniformity pretreatment plan is designed, so as to transform non-uniform sampling to uniform equivalent sampling signal, thus the observing target property is acquired by means of analysis algorithm of subsequent uniform sampling signal.

The invention discloses a sparse representation-based doppler beam sharpened imaging method and device. The method comprises the following steps of carrying out matched filtering processing on an echo signal received in a range direction, carrying out range migration correction on the processed data; carrying out related function method processing on the data subjected to range migration correction and determining a mean value to obtain a doppler center value; carrying out doppler center compensation on the data subjected to range migration correction to obtain doppler center compensation data; carrying out DBS imaging on the doppler center compensation data and calculating the noise power level; constructing a hypercomplete sparse dictionary according to a doppler frequency resolution and carrying out sparse recovery on the doppler center compensation data; and extracting data corresponding to a doppler bandwidth from the sparse data in a range doppler domain by adopting a linear interpolation method, converting the extracted data into a geodetic coordinate by taking a first group of data as a criterion and sequentially splicing the extracted data until obtaining a wide-area monitoring image. The method has a relatively high transverse resolution.

The invention discloses a method and device for determining the effective Doppler bandwidth of a fuze receiver. The method comprises the steps of dividing a target into multiple target scattering surface elements after geometric modeling is conducted on the target, calculating the Doppler shift of each target scattering surface element and the echo power generated by each target scattering surface element under the missile target encounter condition, drawing a Doppler frequency spectrum diagram of the target according to the calculated Doppler shift and echo power of each target scattering surface element, and cutting out a spectral line range within which echo power is larger than set threshold power from the drawn Doppler frequency spectrum diagram and determining the effective Doppler bandwidth of the fuze receiver according to the cut-out spectral line range. By the adoption of the method and device, estimation of the effective Doppler bandwidth of the fuze receiver can be achieved, so that the fuze receiver can be designed based on the estimated effective Doppler bandwidth of the fuze receiver during fuze slideway testing, and then fuze slideway testing is facilitated.

The invention provides an SAR motion target detection method based on Doppler spectral analyses. The SAR motion target detection method based on Doppler spectral analyses is used for detecting ground motion targets in an SAR picture, and mainly solves the problem that an existing detection method is not high in detection probability so as to improve detection probability. According to the method, rectangular neighborhoods of scattering points in the SAR picture are chosen, clutter lock is conducted on Doppler frequency spectrums in each neighborhood to acquire Doppler center frequency, and then standard deviations of frequency spectrums after clutter lock are calculated to represent Doppler bandwidths. Traversal is conducted on all scattering points in the SAR picture, the scattering points are processed in the same way, statistics is then carried out, a measure is constructed through obtained statistics to ensure that the detection boundary of the measure is an oval, and a detection threshold is determined according to constant false alarm rate (CFAR). Because the detection boundary is an oval, on the premise that the constant false alarm rates are the same, oval detection boundary detection probability is larger than rectangle detection boundary detection probability, and target missing detection does not occur easily.

The invention provides a phase-locked loop circuit for implementing weak carrier signal acquisition. The phase-locked loop circuit comprises a first phase discriminator, a loop filter, a voltage-controlled oscillator (VCO), an adder and a scanning voltage generation circuit, wherein the first phase discriminator, the loop filter, the adder and the VCO form a loop to acquire and track weak carrier signals, the adder is used for adding loop integral voltage and scanning voltage, and then inputting the sum into the VCO; the scanning voltage generation circuit analyzes and generates scanning voltage by utilizing the input weak carrier signals and signals outputted by the VCO. According to the phase-locked loop circuit, acquisition bandwidth is increased by utilizing the externally added scanning voltage in an acquisition assisting method, so that complete acquisition within Doppler bandwidth is implemented, signal Doppler bandwidth requirements are met, and high-sensitivity broadband acquisition and tracking are achieved with the assistance of digital-control scanning voltage.

The invention discloses an orientation multichannel synthetic aperture radar imaging method based on Doppler estimation. The method comprises the following steps of determining that a synthetic aperture radar includes M channels, wherein the M channels of the synthetic aperture radar receive target echo signals in a detection scope respectively; then calculating equivalent phase center distances of the M channels respectively so as to calculate a two-dimensional frequency domain target echo signal S(fr, fb), and according to the respective equivalent phase center distance of the M channels, calculating and acquiring an effective Doppler bandwidth Be of the synthetic aperture radar; according to the S(fr, fb) and the Be, calculating an orientation direction channel error gamma a of the synthetic aperture radar and an orientation direction channel error estimation value of the synthetic aperture radar; and using the orientation direction channel error estimation value and the S(fr, fb) to carry out error correction, calculating a target echo signal after the error correction is performed on the channel and carrying out ambiguity resolution processing; calculating the target echo signal after the ambiguity resolution processing and then carrying out imaging processing; and calculating synthetic aperture radar imaging of the target echo signal.

The invention relates to a forward-looking sonar Doppler beam sharpening method. Multiple targets are distinguished by using the difference of Doppler frequencies of different azimuth targets in a same wave beam. In range migration correction, the phase compensation technology is used, center frequency zero setting is realized, and Doppler center frequency ambiguity is eliminated. A pulse repetition frequency is improved by reducing a coherence time and emitting orthogonal waveforms in turns, and Doppler bandwidth ambiguity is eliminated. Through the beam widening technology, the length of a synthetic aperture is expanded, a coherent accumulation time is increased, and a sharpening ratio is increased. According to the forward-looking sonar Doppler beam sharpening method, under a given condition, the forward-looking sonar imaging resolution can be effectively improved, under a condition of a given resolution, the size of a receiving array and the number of channels can be reduced, and the complexity of the system is reduced.

The invention relates to an airborne synthetic aperture radar self-focusing method. According to the method, a channel cross correlation model is established in a noise-ignoring multichannel convolution model, an underdetermined homogeneous linear equation set about the ideal data is acquired, unknown parts exceeding doppler bandwidth partially realize zero setting, a non-zero solution is acquired, frequency spectrum leakage and additive noise are considered, singular value decomposition for a coefficient matrix is carried out, a unique solution under the Euclidean norm is acquired, the ideal data of zero is partially combined with the unique solution, the integral ideal data after recovery is acquired, and azimuth compression for the ideal data is carried out to acquire a focusing image. Through the method, an SAR imaging scene has not any hypothesis, good robustness is realized, the method can be applicable to any SAR imaging scenes, derivation under a theory framework of linear algebra is directly carried out, phase error estimation and compensation without iteration can be carried out to realize focusing of an SAR image, and thereby estimation and compensation of motion errors of the airborne synthetic aperture radar echo data can be effectively realized.

The invention relates to the technical field of radar testing, in particular to a target micro-motion feature extraction method based on a double-view-angle Doppler image. The method comprises the following steps: establishing an angle model in which an observation angle between a radar sight line and a conical target axis changes along with a precession angle, and establishing a Doppler bandwidth change model; at the position where the precession angle is 0, Taylor expansion is conducted on the angle model, and after a quadratic term and a high-order term are ignored, an approximate angle model is obtained; combining the Doppler bandwidth change model with the approximate angle model to obtain an approximate Doppler bandwidth model; two radars with different observation angles are set, each radar obtains an average value of Doppler bandwidth through measurement, and sine curve fitting or curve detection is performed on the two average values and the approximate Doppler bandwidth model to obtain two groups of constants, amplitudes, periods and initial phases; and according to the two groups of constants, amplitudes, periods and initial phases, two observation angles, spin linear velocities and precession angles are calculated. According to the scheme, the problem of low micro-motion feature robustness is solved.

The invention discloses a STAP method for high-speed radar in an accelerated motion state. The method includes 1) acquiring the covariance gap matrix of the clutter of an arbitrary range gate; 2) analyzing the factors causing the clutter to diffuse along the Doppler domain; 3) Use time-domain covariance matrix tapering to describe the Doppler spread caused by acceleration; 4) Taper the time-domain covariance matrix Rt for clutter suppression, so that the space-time processor notch widens to suppress Doppler spread clutter ; 5) Obtain the tapering matrix T; 6) Obtain the relative space-time adaptive weight based on the tapering of the covariance matrix. This method determines the taper coefficient according to the Doppler bandwidth caused by acceleration, and performs the Hadamard product with the estimated covariance matrix, so that the notch of the space-time processor is widened along the Doppler direction, and the clutter caused by acceleration is suppressed along the Doppler direction. System performance loss caused by dimension widening.

The invention discloses a sparse representation-based doppler beam sharpened imaging method and device. The method comprises the following steps of carrying out matched filtering processing on an echo signal received in a range direction, carrying out range migration correction on the processed data; carrying out related function method processing on the data subjected to range migration correction and determining a mean value to obtain a doppler center value; carrying out doppler center compensation on the data subjected to range migration correction to obtain doppler center compensation data; carrying out DBS imaging on the doppler center compensation data and calculating the noise power level; constructing a hypercomplete sparse dictionary according to a doppler frequency resolution and carrying out sparse recovery on the doppler center compensation data; and extracting data corresponding to a doppler bandwidth from the sparse data in a range doppler domain by adopting a linear interpolation method, converting the extracted data into a geodetic coordinate by taking a first group of data as a criterion and sequentially splicing the extracted data until obtaining a wide-area monitoring image. The method has a relatively high transverse resolution.

The invention discloses a method for selecting a field of view of a moving platform scanning radar super-resolution imaging, which is applied to the technical field of radar detection and imaging. Aiming at the problem of selecting the applicable range of a moving platform scanning radar super-resolution imaging field of view, the invention analyzes the radar antenna pattern by analyzing the field of view of the moving platform scanning radar. The spatial sampling bandwidth and the Doppler bandwidth of the target in the beam are quantitatively derived, and the applicable boundary of the super-resolution imaging field of view of the airborne scanning radar is quantitatively derived, which is used to select the super-resolution imaging field of view of the moving platform scanning radar; the method of the present invention helps Based on the working mode design and imaging field selection of the airborne scanning radar, it provides help for the effective field of view selection of the super-resolution imaging of the airborne scanning radar.

The invention discloses a ground-sea clutter classification method based on time-domain and frequency-domain multi-features, comprising the following steps: Step 1, calculating the normalized power variance, Doppler bandwidth, and Doppler bandwidth of a single radar spatial resolution unit echo Le spectrum entropy and Doppler spectrum peak power ratio: step 2, generate feature data and category data of ground clutter and sea clutter based on step 1: step 3, construct BP neural network, step 4, calculate each The normalized power variance, Doppler bandwidth, Doppler spectrum entropy and Doppler spectrum peak power ratio of a spatial resolution unit, step 5, judge according to the output result. The ground-sea clutter classification method based on time-domain and frequency-domain multi-features disclosed by the invention can realize the ground-sea clutter classification of different moving platform radars.

An embodiment of the invention provides a determination method for a pulse repetition rate lower limiting value of synthetic aperture radar. The determination method can break through restriction of a radar working platform and working modes to obtain the accurate pulse repetition rate lower limiting value. The method comprises the following steps: determining the Doppler bandwidth of a target echo signal of the synthetic aperture radar; determining the Doppler off-centering bandwidth of the target echo signal of the synthetic aperture radar; and according to the sum of the Doppler bandwidth and the Doppler off-centering bandwidth, determining the pulse repetition rate lower limiting value of the synthetic aperture radar.

The invention provides an azimuth parameter obtaining method for a satellite-borne synthetic aperture radar in a sliding bunching mode. An azimuth initial imaging belt length, an antenna beamwidth, an initial squint angle, a terminate squint angle and a target slant distance are utilized to establish a reference slant distance model corresponding to a beam rotation center; reference slant distance data are obtained, and an effective irradiation time model corresponding to a point target is established according to the reference slant distance data; effective initial irradiation time and effective terminate irradiation time of the point target are obtained, and an effective Doppler frequency and Doppler bandwidth model corresponding to the point target is established according to the effective initial irradiation time and effective terminate irradiation time of the point target; initial effective Doppler frequency, terminate effective Doppler frequency, the Doppler bandwidth and the resolution declining coefficient corresponding to the point target are obtained, an azimuth resolution model corresponding to the point target is established according to the initial effective Doppler frequency, terminate effective Doppler frequency, the Doppler bandwidth and the resolution declining coefficient corresponding to the point target, and the azimuth resolution corresponding to the point target is obtained; when the azimuth resolution is larger than a resolution index requirement, the maximum azimuth imaging belt length datum is output.

An azimuth signal reconstruction method for a Synthetic Aperture Radar (SAR) is provided. A first Signal Noise Ratio (SNR) amplification factor of a first channel transfer function matrix is determined according to the first channel transfer function matrix, and an equivalent doppler bandwidth is determined according to the SNR amplification factor; a second channel transfer function matrix corresponding to each doppler frequency point is determined according to the equivalent doppler bandwidth and the first channel transfer function matrix; and a spectrum filter for extracting a doppler spectrum is determined according to the second channel transfer function matrix corresponding to each doppler frequency point, a premeasured aliasing spectrum and an original spectrum corresponding to the aliasing spectrum, the spectrum filter being configured to extract the doppler spectrum in an azimuth signal. An azimuth signal reconstruction device for the SAR is further provided.

The SVD algorithm separates a clutter subspace, a target subspace and a noise subspace in radar echoes through singular value decomposition, and the purpose of clutter suppression is achieved. However, due to aliasing of the target and the sea clutter, the clutter suppression effect can be affected by clutter subspace division. In order to solve the problem that the clutter subspace is difficult to determine after the radar echo signal is subjected to SVD decomposition, the invention provides a method for adaptively selecting the clutter subspace by using an improved K-means clustering method. Aiming at the problems that K-means is sensitive to the initial centroid and different characteristic parameters are different in the aspects of statistical stability and anti-noise performance, the initial centroid is fixed, and the sample distance is calculated in a weighted mode. According to the method, after four characteristics of singular value distribution, echo component correlation, Doppler bandwidth and relative Doppler variable coefficient are used as K-means input, clutter subspaces are selected in a self-adaptive mode, and the purpose of clutter suppression is achieved by mapping signals into orthogonal spaces of the clutter subspaces.