CN114488133A - Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship - Google Patents

Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship Download PDF

Info

Publication number
CN114488133A
CN114488133A CN202210219273.XA CN202210219273A CN114488133A CN 114488133 A CN114488133 A CN 114488133A CN 202210219273 A CN202210219273 A CN 202210219273A CN 114488133 A CN114488133 A CN 114488133A
Authority
CN
China
Prior art keywords
signals
gnss
ship target
frequency band
ship
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202210219273.XA
Other languages
Chinese (zh)
Other versions
CN114488133B (en
Inventor
夏正欢
张涛
刘新
赵志龙
张瑶
张可佳
易春宏
张庆君
金世超
岳富占
彭涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Institute of Satellite Information Engineering
Original Assignee
Beijing Institute of Satellite Information Engineering
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Institute of Satellite Information Engineering filed Critical Beijing Institute of Satellite Information Engineering
Priority to CN202210219273.XA priority Critical patent/CN114488133B/en
Publication of CN114488133A publication Critical patent/CN114488133A/en
Application granted granted Critical
Publication of CN114488133B publication Critical patent/CN114488133B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
    • G01S13/9004—SAR image acquisition techniques
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
    • G01S13/9021—SAR image post-processing techniques
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/13—Receivers
    • G01S19/14—Receivers specially adapted for specific applications
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/01—Satellite radio beacon positioning systems transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/13—Receivers
    • G01S19/35—Constructional details or hardware or software details of the signal processing chain
    • G01S19/37—Hardware or software details of the signal processing chain

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Signal Processing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention relates to a method for extracting and classifying multidimensional scattering characteristics of a satellite-borne GNSS-S radar ship, which comprises the following steps: a. receiving ship target scattering signals of a plurality of navigation satellite signals by using a satellite-borne GNSS-S radar (10), and performing double-station SAR imaging on the signals to obtain a multi-dimensional SAR image of the ship target; b. carrying out non-coherent fusion processing on the multi-dimensional SAR image, and carrying out ship target detection on the fused image to obtain position information of a ship target; c. extracting the length direction and the bow direction of the ship target, and calculating the multidimensional scattering coefficient of the ship target; d. and constructing a vectorized multi-dimensional electromagnetic scattering set, and classifying the types of the ship targets by using a target classification network. The invention fully utilizes the multidimensional electromagnetic scattering information of the ship target, utilizes the convolutional neural network model to realize the highly reliable and intelligent classification of the ship target, has higher classification accuracy and can solve the problem of ship target classification under the condition of medium and high sea.

Description

Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship
Technical Field
The invention relates to a method for extracting and classifying multidimensional scattering characteristics of a satellite-borne GNSS-S radar ship.
Background
The detection and classification of sea surface ship targets are related to ocean safety and marine transportation safety, and the main modes of the existing sea surface ship target detection include an optical satellite, a Synthetic Aperture Radar (SAR) satellite and the like. The optical satellite has the advantages of high-resolution imaging, strong interpretation capability, high classification accuracy and the like, so the optical satellite is widely applied to earth and sea observation, but is influenced by weather such as sea cloud, fog and rain, and the like, and is difficult to continuously exert the advantages of high-resolution imaging observation and target classification. The SAR satellite realizes high-resolution imaging by transmitting a high-power electromagnetic signal and carrying out long-time synthetic aperture processing, and the electromagnetic signal can penetrate through clouds and fog, so that the SAR satellite has the advantages of all-weather observation all-weather, but the SAR satellite has high power consumption, short working time of each orbit, difficulty in continuously carrying out imaging detection on the sea, high transmitted signal power and easiness in interception and interference, and the imaging detection quality is limited.
With the improvement and the comprehensive open utilization of the satellite navigation system technology, the mode of detecting and classifying the sea surface ship targets by means of the multi-dimensional electromagnetic signals of a plurality of navigation satellites gradually becomes an important way for solving various defects of the existing sea surface ship target detection mode, and particularly, the intelligent classification of the scattering signals (GNSS-S) of the global navigation satellite system with a plurality of angles, a plurality of frequency bands and a plurality of polarizations by combining the same ship target is performed. The satellite-borne GNSS-S radar can simultaneously receive multi-dimensional signals of a plurality of navigation satellites and provide rich information basis for sea surface ship target detection and classification, but the signal power of the navigation satellite signals is low, the signal bandwidth is narrow, the imaging resolution is low, and high-performance detection and classification are difficult to be performed by directly utilizing low-resolution images, so that a new ship target detection and classification method needs to be explored, a new system radar technology is developed to realize the sea surface ship target detection, and all-day, all-weather and full-duty-ratio detection is realized.
Disclosure of Invention
The invention aims to provide a method for extracting and classifying multidimensional scattering characteristics of a satellite-borne GNSS-S radar ship.
In order to achieve the above object, the present invention provides a method for extracting and classifying multidimensional scattering characteristics of a satellite-borne GNSS-S radar ship, comprising the following steps:
a. receiving ship target scattering signals of a plurality of navigation satellite signals by using a satellite-borne GNSS-S radar, and performing double-station SAR imaging on the signals to obtain a multi-dimensional SAR image of the ship target;
b. carrying out non-coherent fusion processing on the multi-dimensional SAR image, and carrying out ship target detection on the fused image to obtain position information of a ship target;
c. extracting the length direction and the bow direction of the ship target, and calculating the multidimensional scattering coefficient of the ship target;
d. and constructing a vectorized multi-dimensional electromagnetic scattering set, and classifying the types of the ship targets by using a target classification network.
According to one aspect of the invention, in said step (a), receiving a ship target scatter signal to obtain multi-dimensional observation information, including multi-angle, multi-band, multi-polarization of the ship target;
in the step (c), estimating the length and width of the ship target to obtain the length direction of the ship target;
calculating scattering coefficients of the ship target at different angles, different frequency bands and different polarizations by using an electromagnetic scattering characteristic calculation unit to obtain a multidimensional scattering coefficient of the ship target;
in the step (d), the target classification network is a convolutional neural network model, the vectorized multi-dimensional electromagnetic scattering set is input as the ship target, and the output is the type of the ship target.
According to one aspect of the invention, a satellite-borne GNSS-S radar comprises:
the azimuth multi-channel antenna is used for receiving multi-dimensional GNSS-S signals of a detection area, and has three frequency band signal receiving capacity, the frequencies of the three frequency bands are fc1, fc2 and fc3 respectively, and the corresponding working bandwidths are Bw1, Bw2 and Bw3 respectively; the antenna has left-hand circular polarization and right-hand circular polarization signal receiving capacity, and the polarization isolation degree is greater than 20 dB; the whole antenna is divided into M receiving antennas along the azimuth direction, wherein M is 4-8, each receiving antenna consists of Na multiplied by Nr antenna units, Na is 6-12, and Nr is 12-24;
the M six-channel receivers are used for performing low-noise amplification, band-pass filtering, down-conversion, intermediate-frequency low-noise amplification, intermediate-frequency analog-to-digital converter (ADC) sampling and digital domain IQ demodulation on the GNSS-S signals of three frequency bands and two polarizations output by each receiving channel to obtain six-baseband GNSS-S complex signals;
the digital beam forming processing unit is used for carrying out digital beam forming on GNSS-S signals of three frequency bands and two polarizations and forming P digital sub-beams in the azimuth direction;
the P matched filter groups are used for performing matched filtering on the GNSS-S signals output by each digital sub-beam by using signals emitted by N navigation satellites as reference signals to obtain N independent GNSS-S signals with high signal-to-noise ratio and outputting P frame signal sets, and each frame signal set comprises 6N GNSS-S signals with high signal-to-noise ratio;
after signals of the M receiving antennas are subjected to digital beam forming, a plurality of high-gain narrow beams are formed, and the gain of each narrow beam is larger than 30 dB.
According to an aspect of the invention, a digital beam forming processing unit includes:
first of allP sets of complex multiply-add operation units for left-handed circularly polarized M GNSS-S signals SA in frequency band fc1m(t) performing Digital Beamforming (DBF) processing along the azimuth direction to form P digital sub-beams, and outputting P high signal-to-noise ratio GNSS-S signals as:
Figure BDA0003536346030000031
a second P sets of complex multiply-add units for performing right-hand circular polarization M GNSS-S signals SB at frequency band fc1m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure BDA0003536346030000041
a third P sets of complex multiply-add units for performing left-handed circularly polarized M GNSS-S signals SC at frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure BDA0003536346030000042
a fourth P sets of complex multiply-add units for performing right-hand circularly polarized M GNSS-S signals SD on the frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure BDA0003536346030000043
a fifth P sets of complex multiply-add units for performing left-handed circularly polarized M GNSS-S signals SE at frequency band fc3m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure BDA0003536346030000044
a sixth P sets of complex multiply-add units for performing right-hand circularly polarized M GNSS-S signals SF at frequency band fc3m(t) DBF processing is carried out along the azimuth direction to form P digital sub-beams, and P high signal-to-noise ratio GNSS-S signals are output as follows:
Figure BDA0003536346030000045
a DBF weight calculation unit for calculating 6 sets of complex weights wa required by DBF processingp,m、wbp,m、wcp,m、wdp,m、wep,m、wfp,mCalculating;
wherein t is a fast time variable; m is a serial number, and the value is M1, 2. P is a serial number, and the value is P1, 2. SAm(t) is frequency band fc1, left hand circularly polarized M GNSS-S signals; RAp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the left-handed circularly polarized signals; wa (a)p,mThe complex weight is the complex weight when the frequency band fc1 and the left-handed circularly polarized signal are subjected to DBF processing; SB (service bus)m(t) is frequency band fc1, right hand circularly polarized M GNSS-S signals; RB (radio B)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the right-hand circularly polarized signals; wbp,mThe complex weight value is the complex weight value when the frequency band fc1 and the right-hand circularly polarized signal are subjected to DBF processing; SC (Single chip computer)m(t) is frequency band fc2, left hand circularly polarized M GNSS-S signals; RC (resistor-capacitor) capacitorp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc2 and the left-handed circularly polarized signals; wc cp,mThe complex weight is the complex weight when the frequency band fc2 and the left-handed circularly polarized signal are subjected to DBF processing; SDm(t) is frequency band fc2, right hand circularly polarized M GNSS-S signals; RDp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc2 and the right-hand circularly polarized signals; wdp,mThe complex weight value is the complex weight value when the frequency band fc2 and the right-hand circularly polarized signal are subjected to DBF processing; SEm(t) is frequency band fc3, left hand circularly polarized M GNSS-S signals; REp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the left-handed circularly polarized signals; wep,mThe complex weight is the complex weight when the frequency band fc3 and the left-handed circularly polarized signal are subjected to DBF processing; SFm(t) is frequency band fc3, right hand circularly polarized M GNSS-S signals; RF (radio frequency)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the right-hand circularly polarized signals; wfp,mThe complex weights are used for DBF processing of the frequency band fc3 and the right-hand circularly polarized signal. According to one aspect of the invention, bi-site SAR imaging includes:
a1, constructing a double-station SAR imaging geometric scene according to the position and speed information of a navigation satellite output by a satellite navigation positioning system and the position and speed information of a GNSS-S radar system, wherein the size of the imaging scene is X multiplied by Y, and the sizes of imaging grids are delta X multiplied by delta Y;
a2, carrying out parallelization time domain BP double-station imaging processing on the GNSS-S signal by utilizing P parallelization Back Projection (BP) imaging processing units to obtain 6N double-station SAR images;
the detection area is irradiated by N navigation satellite signals to form N double-station radars, each double-station radar comprises three frequency bands and two kinds of polarization information, and the parallelization BP imaging processing unit totally comprises 6N time domain BP double-station imaging processing submodules.
According to one aspect of the invention, the multi-dimensional SAR image fusion processing comprises the following steps:
b1, compensating the power path loss of the double-station SAR image at different observation angles;
b2, performing weight superposition on the 6N double-station SAR images by using a non-coherent processing method to obtain a high signal-to-noise ratio SAR image.
According to an aspect of the invention, the length direction extraction includes:
c11, carrying out slice extraction on the detected ship target, and removing the sea clutter background;
c12, calculating the length-width ratio of the ship target;
c13, selecting the direction with the largest length-width ratio as the long axis direction of the ship target;
the extraction of the ship bow direction comprises the following steps:
c21, calculating the ship target position in each frame of the fused two-station SAR image;
c22, correlating the P frame ship target position information to obtain the motion track and the motion direction of the ship target;
c23, obtaining the ship bow direction according to the long axis direction and the motion direction of the ship target, and calculating the azimuth included angle between the qth ship and the GNSS-S radar receiving antenna beam center
Figure BDA0003536346030000061
The multi-dimensional scattering property calculation includes:
c31, respectively carrying out azimuth multi-view processing on the 6N double-station SAR images before fusion, wherein the multi-view times K are 3-10;
c32, calculating scattering coefficients of the ship target in the 6N multi-view processed double-station SAR images respectively to obtain scattering coefficients sigma 1 of the ship target for N double-station angles, three frequency bands and two polarizationsq,n、σ2q,n、σ3q,n、σ4q,n、σ5q,n、σ6q,nThe frequency band fc1 and the left-hand circularly polarized scattering coefficient, the frequency band fc1 and the right-hand circularly polarized scattering coefficient, the frequency band fc2 and the left-hand circularly polarized scattering coefficient, the frequency band fc2 and the right-hand circularly polarized scattering coefficient, the frequency band fc3 and the left-hand circularly polarized scattering coefficient, and the frequency band fc3 and the right-hand circularly polarized scattering coefficient, which are formed by the irradiation of the nth navigation satellite on the qth ship target, are respectively shown.
According to an aspect of the present invention, in step (d), a multidimensional scattering characteristic set of the ship target is constructed according to an observation angle, a working frequency band, a polarization mode, a position and a speed of a navigation satellite, a position and a speed of a satellite-borne GNSS-S radar, and a ship heading direction of the ship target, where the multidimensional scattering characteristic set Ω q of a qth ship target is:
Figure BDA0003536346030000071
Figure BDA0003536346030000072
wherein, Delta theta sq,n、
Figure BDA0003536346030000073
Respectively the incident angle and the azimuth angle of the nth navigation satellite signal irradiated on the qth ship target, wherein the azimuth angle is in reference to the ship bow direction of the ship target, and the delta theta sq,n＝θsn，
Figure BDA0003536346030000074
Δθrq、
Figure BDA0003536346030000075
Respectively receiving an incident angle and an azimuth angle of the qth ship target by the satellite-borne GNSS-S radar, wherein the azimuth angle is based on the ship bow direction of the ship target, and the delta theta rq＝θrin，
Figure BDA0003536346030000076
Llp、LrpLeft hand circular polarization and right hand circular polarization, respectively.
According to an aspect of the present invention, in step (d), the received triple-frequency dual-polarization information of each navigation transmitting satellite and the corresponding scattering coefficients are normalized and then input into the target classification network together;
the object classification network includes:
the multilayer convolution is used for sequentially performing convolution, activation and pooling on input, extracting characteristics including three-frequency point dual polarization and corresponding scattering coefficients, and sharing multilayer convolution parameters corresponding to a plurality of navigation transmitting satellites;
the multilayer perceptron is used for carrying out multilayer perception on the characteristic information obtained by multilayer convolution in a parameter sharing mode;
realizing the feature fusion of a plurality of launching satellites;
and the classifier is used for judging the class of the ship according to the fusion characteristics.
According to one aspect of the invention, the track height H of the satellite-borne GNSS-S radar is 200-800km, and a ship multi-dimensional scattering signal of a plurality of navigation satellite signals is received by a three-frequency-band, dual-polarization and azimuth multi-channel array antenna;
the antenna receives multi-dimensional GNSS-S signals of a ship target in a front side view mode, the multi-dimensional GNSS-S signals consist of M receiving channels in the azimuth direction, the wave beam in the azimuth direction of the antenna of each receiving channel is a wide wave beam, and the wave beam angle theta is larger than the wave beam angle thetaaGreater than 5 DEG, the total synthetic aperture length being LS；
Obtaining P sub-beams after digital beam forming processing is carried out on the M receiving channels, wherein each sub-beam corresponds to multiple frames of different position information of a ship target, and multiple frames of SAR images are obtained after double-station SAR imaging;
angle of incidence thetar at the center of the antenna beamin25-55 degrees, and the azimuth angle of the center of the antenna beam is 0;
the detection area is simultaneously irradiated by N navigation satellite signals, N double-station radar systems are formed by the detection area and the satellite-borne GNSS-S radar, and the incident angle of the nth navigation satellite signal is theta SnIn an azimuth of
Figure BDA0003536346030000081
According to the concept of the invention, a ship multidimensional scattering characteristic extraction and intelligent classification method based on a satellite-borne GNSS-S radar is provided. The satellite-borne GNSS-S radar can simultaneously receive ship target scattering signals of a plurality of navigation satellite signals and obtain multi-dimensional observation information of a sea surface ship target, such as multi-angle, multi-frequency band, multi-polarization and the like. Firstly, carrying out parallelization double-station SAR imaging on a multi-dimensional GNSS-S signal of a ship target to obtain a multi-dimensional SAR image of the ship target; secondly, carrying out non-coherent fusion processing on the multi-dimensional SAR image so as to improve the contour information of the ship target, realize the detection of the ship target and determine the length direction of the ship target; thirdly, judging the motion direction of the ship target by using the continuous multi-frame images, determining the ship bow direction of the ship target, calculating the multi-dimensional equivalent scattering coefficient of the ship target, and establishing a multi-dimensional scattering characteristic set of the ship target; and finally, intelligently classifying the ship target by using a convolutional neural network, inputting the SAR image slice and the multi-dimensional scattering characteristic set of the ship target, and outputting the SAR image slice and the multi-dimensional scattering characteristic set as the type of the ship target. Therefore, compared with the existing ship target classification method based on the SAR image, the ship target classification method based on the SAR image fully utilizes multi-dimensional electromagnetic scattering information of the ship target, such as multi-angle, multi-frequency band, multi-polarization and the like, acquired by the satellite-borne GNSS-S radar, establishes the multi-dimensional scattering characteristic set of the ship target, utilizes the convolutional neural network model to realize the ship target elevation reliable intelligent classification, has higher classification accuracy, can solve the problem of ship target classification under the condition of medium and high sea, does not need to actively transmit high-power signals, can realize the detection and classification of the ship target only by receiving and processing ship target scattering signals of a plurality of navigation satellite signals, and has the advantages of low cost, low power consumption, light weight and the like.
According to one scheme of the invention, the observation time of a ship target on the sea surface is prolonged by utilizing a satellite-borne GNSS-S radar azimuth multi-channel technology, a multi-frame image of the ship target is obtained by adopting an azimuth digital beam forming and double-station SAR imaging method, the motion direction and the ship bow direction of the ship target are extracted, and a reference azimuth angle is provided for the construction of a multi-dimensional scattering characteristic set of the ship target.
Drawings
FIG. 1 is a flow chart of a process of a method for intelligently classifying ship targets according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating multi-dimensional scattering information acquisition of a satellite-borne GNSS-S radar according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a satellite-borne GNSS-S radar system according to an embodiment of the present invention;
FIG. 4 is a flow diagram schematically illustrating an azimuth multi-channel digital beamforming process according to an embodiment of the present invention;
FIG. 5 is a flow chart of a process for fusion of two-station SAR imaging and a multi-dimensional SAR image according to an embodiment of the present invention;
FIG. 6 is a flow chart of the bow direction extraction and multi-dimensional scattering property set construction of a ship target according to an embodiment of the present invention;
fig. 7 is a flow chart for implementing ship target classification using a convolutional neural network according to an embodiment of the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
The present invention is described in detail below with reference to the drawings and the specific embodiments, which are not repeated herein, but the embodiments of the present invention are not limited to the following embodiments.
Referring to fig. 1, the ship target multidimensional scattering characteristic extraction and intelligent classification method based on the satellite-borne GNSS-S radar of the invention utilizes navigation satellite signals to realize detection and classification of ship targets under complex sea conditions, and is suitable for technical research of space-based distributed high-resolution wide-range SAR imaging systems.
The method comprises the steps of firstly, simultaneously receiving ship target scattering signals of a plurality of navigation satellite signals by using a satellite-borne GNSS-S radar 10, and thus obtaining multi-dimensional observation information of the ship target, such as multi-angle, multi-frequency band, multi-polarization and the like. Subsequently, a parallelization double-station SAR imaging 20 is carried out on the multi-dimensional GNSS-S signal of the ship target by adopting a BP imaging method, and a multi-dimensional SAR image of the sea surface ship target is obtained. Then, multi-dimensional SAR image fusion processing 30 is carried out, namely, non-coherent fusion processing is carried out on the multi-dimensional SAR image of the ship target so as to obtain a high signal-to-noise ratio SAR image, and therefore the contour information of the ship target is enhanced. And then carrying out ship target detection 40 on the SAR image after the fusion processing to obtain the position information of the ship target. Then, carrying out length direction extraction 50 and warship bow direction extraction 60 on the ship target, namely, roughly estimating the length and width of the ship target to obtain the length direction of the ship target; and extracting the ship bow direction of the ship target to obtain a scattering characteristic calculation reference azimuth of the ship target. And the electromagnetic scattering property calculation unit 70 is utilized to calculate the scattering coefficients of the ship target at different angles, different frequency bands and different polarizations, so as to obtain the multidimensional scattering coefficient of the ship target. And then, constructing 80 a multi-dimensional electromagnetic scattering set, namely constructing a vectorized multi-dimensional electromagnetic scattering set of the ship target according to the position and the speed of the navigation satellite, the position and the speed of the satellite-borne GNSS-S radar, the ship bow direction of the ship target and the like. And finally, intelligently classifying the type of the ship target by using a target classification network, wherein the target classification network is a convolutional neural network model 90, the input (layer) of the target classification network is a vectorized multidimensional electromagnetic scattering set of the ship target, and the output (layer) of the target classification network is the type of the ship target.
Referring to FIG. 2, the orbit height of the GNSS-S radar 10 is H, 200 + 800km, and one triple frequency band (fc1, fc2, fc3) and dual polarization (L-polarization left-handed)lpAnd right hand circular polarization Lrp) And the azimuth multi-channel array antenna simultaneously receives multi-dimensional scattering signals of the ships of the navigation satellite signals. The antenna receives multi-dimensional GNSS-S signals in a front side view mode, the antenna consists of M receiving channels in the azimuth direction, the antenna azimuth beam of each receiving channel is a wide beam, and the beam angle thetaaMore than 5 degrees, the observation time of the ship target is increased, and the total synthetic aperture length is LS. After digital beam forming processing is carried out on the M receiving channels, P sub-beams are obtained, each sub-beam corresponds to multiple frames of different position information of a ship target, and multiple frames of SAR images are obtained after double-station SAR imaging. Recording the incident angle of the center of the antenna beam as thetarinTaking 25-55 degrees, the azimuth angle of the antenna beam center is 0, and the azimuth included angle between the qth ship target and the antenna beam center is
Figure BDA0003536346030000111
The detection area is simultaneously irradiated by N navigation satellite signals, N double-station radar systems are formed by the detection area and the satellite-borne GNSS-S radar 10, and the incident angle of the nth navigation satellite signal is recorded as theta SnIn an azimuth of
Figure BDA0003536346030000112
Δθsq,n、
Figure BDA0003536346030000113
Respectively the incident angle and the azimuth angle of the nth navigation satellite signal irradiated on the qth ship target, wherein the azimuth angle is in reference to the ship bow direction of the ship target, and the delta theta sq,n＝θsn，
Figure BDA0003536346030000121
Δθrq、
Figure BDA0003536346030000122
The incident angle and the azimuth angle of the q-th ship target received by the satellite-borne GNSS-S radar 10 are respectively delta theta rq＝θrin，
Figure BDA0003536346030000123
Referring to fig. 3, the GNSS-S radar 10 on board includes: the azimuth multi-channel antenna 101 is used for receiving multi-dimensional GNSS-S signals of a detection area, and has three frequency band signal receiving capacity, the frequencies of the three frequency bands are fc1, fc2 and fc3 respectively, the frequency of fc1 is 1.191GHz, the frequency of fc2 is 1.268GHz, the frequency of fc3 is 1.575GHz, the corresponding working bandwidths are Bw1, Bw2 and Bw3 respectively, the frequency of Bw1 and Bw2 is 2.046MHz, and the frequency of Bw3 is 20.46 MHz; the antenna has left-hand circular polarization and right-hand circular polarization signal receiving capacity, and the polarization isolation degree is greater than 20 dB; the whole antenna is divided into M receiving antennas along the azimuth direction, wherein M is 4-8, each receiving antenna consists of Na multiplied by Nr antenna units, and Na is 6-12, so that each receiving antenna has a wide beam angle in the azimuth direction, the observation time of a ship target is increased, and multi-frame sub-aperture imaging processing is facilitated; nr takes 12-24, so that each receiving antenna has the characteristics of narrow beam angle, high gain and the like in the distance direction; after signals of M receiving antennas are subjected to digital beam forming, a plurality of high-gain narrow beams are formed, and the gain of each narrow beam is larger than 30 dB; the M six-channel receivers 102 are configured to perform low-noise amplification, band-pass filtering, down-conversion, intermediate-frequency low-noise amplification, intermediate-frequency ADC sampling, digital domain IQ demodulation, and the like on GNSS-S signals of three frequency bands and two polarizations output by each receiving channel, so as to obtain six baseband GNSS-S complex signals; the 6 digital beam forming processing units 103 are configured to perform digital beam forming on GNSS-S signals of three frequency bands and two polarizations, and form P digital sub-beams in an azimuth direction; and the P matched filter banks 104 are configured to perform matched filtering on the GNSS-S signals output by each digital sub-beam by using signals emitted by N navigation satellites as reference signals to obtain N independent GNSS-S signals with high signal-to-noise ratio, and output P frame signal sets, where each frame signal set includes 6N GNSS-S signals with high signal-to-noise ratio.
Referring to fig. 4, the digital beam forming processing unit 103 includes: a first P sets of complex multiply-add units 1031 for left-handed circularly polarized M GNSS-S signals SA in frequency band fc1m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000131
a second P complex multiply-add unit 1032 for performing right-hand circular polarization M GNSS-S signals SB in the frequency band fc1m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000132
a third P sets of complex multiply-add units 1033 for performing left-handed circularly polarized M GNSS-S signals SC at frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000133
a fourth P sets of complex multiply-add units 1034 for generating M GNSS-S signals SD with right-hand circular polarization and frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000134
a fifth P sets of complex multiply-add units 1035 for left-handed circularly polarized M GNSS-S signals SE at frequency band fc3m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000135
a sixth P sets of complex multiply-add units 1036 for right-hand circularly polarized M GNSS-S signals SF at fc3m(t) performing DBF processing along the azimuth direction to form P digital sub-beams and output P GNSS-S signals with high signal-to-noise ratio, namely,
Figure BDA0003536346030000141
a DBF weight calculation unit 1037 for calculating 6 sets of complex weights wa required for DBF processingp,m、wbp,m、wcp,m、wdp,m、wep,m、wfp,mAnd (6) performing calculation. Wherein t is a fast time variable; m is a serial number, and the value is M1, 2. P is a serial number, and the value is P1, 2. SAm(t) is frequency band fc1, left hand circularly polarized M GNSS-S signals; RAp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the left-handed circularly polarized signals; wa (a)p,mThe complex weight is the complex weight when the frequency band fc1 and the left-handed circularly polarized signal are subjected to DBF processing; SB (bus bar)m(t) is frequency band fc1, right hand circularly polarized M GNSS-S signals; RB (radio B)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the right-hand circularly polarized signals; wbp,mThe complex weights are the complex weights when DBF processing is carried out on the frequency band fc1 and the right-hand circularly polarized signal; SC (Single chip computer)m(t) is frequency band fc2, left hand circularly polarized M GNSS-S signals; RC (resistor-capacitor) capacitorp(t) is a left-hand circularly polarized signal corresponding to the frequency band fc2P GNSS-S signals with high signal-to-noise ratio after DBF processing; wc cp,mThe complex weight is the complex weight when the frequency band fc2 and the left-handed circularly polarized signal are subjected to DBF processing; SDm(t) is frequency band fc2, right hand circularly polarized M GNSS-S signals; RDp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc2 and the right-hand circularly polarized signals; wdp,mThe complex weight value is the complex weight value when the frequency band fc2 and the right-hand circularly polarized signal are subjected to DBF processing; SEm(t) is frequency band fc3, left hand circularly polarized M GNSS-S signals; REp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the left-handed circularly polarized signals; wep,mThe complex weight is the complex weight when the frequency band fc3 and the left-handed circularly polarized signal are subjected to DBF processing; SFm(t) is frequency band fc3, right hand circularly polarized M GNSS-S signals; RF (radio frequency)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the right-hand circularly polarized signals; wfp,mThe complex weights are used for DBF processing of the frequency band fc3 and the right-hand circularly polarized signal.
Referring to fig. 5, the two-station SAR imaging 20 comprises: constructing 201 a double-station SAR imaging geometric scene, namely constructing a double-station SAR imaging geometric coordinate system and an imaging scene size for a detection area, specifically constructing the double-station SAR imaging geometric scene according to navigation satellite position and speed information output by a satellite navigation positioning system and position and speed information of a GNSS-S radar system, wherein the imaging scene size is X multiplied by Y, and the imaging grid size is delta X multiplied by delta Y; performing parallelization time domain BP double-station SAR imaging processing on three frequency bands and two polarization GNSS-S signals of N double-station radars by using P parallelization BP imaging processing units 202 to obtain 6N double-station SAR images; the detection area (imaging scene) is irradiated by N navigation satellite signals to form N double-station radars, each double-station radar comprises three frequency bands and two kinds of polarization information, and the parallelization BP imaging processing unit 202 comprises 6N time domain BP double-station imaging processing sub-modules in total.
The multi-dimensional SAR image fusion process 30 includes: image power compensation 301, namely, the path loss of the power of the double-station SAR image at different observation angles is compensated, so as to reduce the power error of the ship target caused by the attenuation of different navigation satellite routes; and (2) performing non-coherent fusion processing 302, namely performing fusion processing on the 6N bistatic SAR images to obtain a high signal-to-noise ratio bistatic SAR image, specifically performing weight superposition on the 6N bistatic SAR images by using a non-coherent processing method to improve the timeliness of the fusion processing, improve the contour information of a ship target, realize ship target detection, and further facilitate extraction of the length direction of the ship target.
Referring to fig. 6, the length direction extraction 50 includes: ship slice extraction 501, namely, ship target detection is carried out on the high signal-to-noise ratio double-station SAR image (namely the fused double-station SAR image), the position of the ship target is determined, slice extraction is carried out on the detected ship target, and the sea clutter background is removed; ship aspect ratio calculation 502, i.e., calculating the aspect ratio of the ship target, as for civil ships, 4:1-7: 1; and (3) calculating the length direction 503, namely determining the long axis direction of the ship target, and specifically selecting the direction with the largest length-width ratio as the long axis direction of the ship target.
The warship bow direction extraction 60 is to determine the motion direction of the ship target by using continuous multi-frame sub-aperture images, and certainly, beam synthesis needs to be performed on M receiving antennas of the GNSS-S radar by using a digital beam forming method, P sub-beams are formed in the azimuth direction, and the GNSS-S signal of each sub-beam is subjected to the above processing to obtain P-frame SAR images of the ship target, so as to obtain P positions of the ship target. Specifically, the method comprises the following steps: extracting 601 a ship target position, namely calculating the ship target position in each frame of the fused two-station SAR image; extracting 602 a ship target motion track, namely, correlating the position information of the P frame ship target to obtain the motion track and the motion direction of the ship target; calculating the ship bow direction 603, namely determining the ship bow direction of the ship target, specifically obtaining the ship bow direction according to the long axis direction and the motion direction of the ship target, and calculating the azimuth included angle between the qth ship and the GNSS-S radar receiving antenna beam center
Figure BDA0003536346030000161
The multidimensional scattering characteristic calculation by the electromagnetic scattering characteristic calculation unit 70 includes: direction of orientationMulti-view processing 701, namely, respectively performing azimuth multi-view processing on the 6N double-station SAR images before fusion to reduce speckle noise, wherein the multi-view times K are 3-10; calculating the ship scattering coefficient 702, namely calculating the scattering coefficients of the ship target in the 6N multi-view processed double-station SAR images respectively, specifically calculating the radar cross-sectional area of the ship target of the 6N double-station SAR images by using a radar equation, and calculating the electromagnetic scattering coefficients of the ship target under the conditions of different observation angles, different frequency bands and different polarizations according to the imaging resolution of the double-station SAR images, namely obtaining the scattering coefficient sigma 1 of the ship target on multidimensional electromagnetic signals of N double-station angles, three frequency bands, two polarizations and the likeq,n、σ2q,n、σ3q,n、σ4q,n、σ5q,n、σ6q,nThe frequency band fc1 and the left-hand circularly polarized scattering coefficient, the frequency band fc1 and the right-hand circularly polarized scattering coefficient, the frequency band fc2 and the left-hand circularly polarized scattering coefficient, the frequency band fc2 and the right-hand circularly polarized scattering coefficient, the frequency band fc3 and the left-hand circularly polarized scattering coefficient, and the frequency band fc3 and the right-hand circularly polarized scattering coefficient, which are formed by the irradiation of the nth navigation satellite on the qth ship target, are respectively shown.
In the multidimensional electromagnetic scattering set construction 80, a multidimensional scattering characteristic set of a ship target is constructed according to information such as an observation angle, a working frequency band, a polarization mode, a position and a speed of a navigation satellite, a position and a speed of the satellite-borne GNSS-S radar 10, a ship bow direction of the ship target and the like of 6N double-station SAR images, and then the multidimensional scattering characteristic set of a qth ship target is as follows:
Figure BDA0003536346030000171
Figure BDA0003536346030000172
wherein, Delta theta sq,n、
Figure BDA0003536346030000173
The incident angle and the incident angle of the nth navigation satellite signal to the qth ship target respectivelyAn azimuth angle; delta theta rq、
Figure BDA0003536346030000174
Receiving an incident angle and an azimuth angle of a qth ship target by the satellite-borne GNSS-S radar 10 respectively, wherein the azimuth angle is based on the ship bow direction of the ship target as reference; l islp、LrpLeft hand circular polarization and right hand circular polarization, respectively.
Referring to fig. 7, received triple-frequency point dual-polarization information of each navigation transmitting satellite and a corresponding scattering coefficient are normalized and then input into (a multilayer convolution of) a ship target classification network together. Wherein the target classification network comprises: the multilayer convolution 901 is used for sequentially performing convolution, activation and pooling on input, extracting characteristics including three-frequency point dual polarization and corresponding scattering coefficients, sharing parameters of the multilayer convolution 901 corresponding to a plurality of navigation transmitting satellites and realizing a network structure with variable transmitting satellite number; the multilayer perceptron 902 is used for performing multilayer perception on the characteristic information obtained by the multilayer convolution 901 in a parameter sharing mode so as to further reduce the dimension of high-dimensional information, and the parameter sharing mode can realize a perception structure with variable transmitting satellite number; feature matching fusion 903, namely, realizing feature fusion of a plurality of transmitting satellites so as to comprehensively utilize three-frequency point and dual-polarization scattering features obtained by observing a ship target from a plurality of angles; the classifiers 904 of a plurality of ship classes are used for judging the classes of ships according to the fusion characteristics, and the process is mainly realized by multilayer perception.
In conclusion, the satellite-borne GNSS-S radar acquires and utilizes multi-dimensional information of a ship target, such as multi-angle, multi-frequency band, multi-polarization and the like, establishes a multi-dimensional scattering characteristic set of the ship target, and utilizes a convolutional neural network to realize highly reliable intelligent classification of the ship target. Compared with the existing image-based ship classification method, the ship target multi-dimensional electromagnetic scattering information acquired by the satellite-borne GNSS-S radar can be fully utilized to construct a multi-dimensional scattering characteristic set of the ship target, the ship target can be classified with better accuracy by means of the convolutional neural network model, the ship target detection and classification can be realized only by receiving and processing ship target scattering signals of a plurality of navigation satellite signals without actively transmitting high-power signals, and the ship target classification method has the advantages of low cost, low power consumption, light weight and the like.
The above description is only one embodiment of the present invention, and is not intended to limit the present invention, and it is apparent to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A method for extracting and classifying multidimensional scattering characteristics of a satellite-borne GNSS-S radar ship comprises the following steps:
a. receiving ship target scattering signals of a plurality of navigation satellite signals by using a satellite-borne GNSS-S radar (10), and performing double-station SAR imaging on the signals to obtain a multi-dimensional SAR image of the ship target;
b. carrying out non-coherent fusion processing on the multi-dimensional SAR image, and carrying out ship target detection on the fused image to obtain position information of a ship target;
c. extracting the length direction and the bow direction of the ship target, and calculating the multidimensional scattering coefficient of the ship target;
d. and constructing a vectorized multi-dimensional electromagnetic scattering set, and classifying the types of the ship targets by using a target classification network.
2. The method of claim 1, wherein in step (a), ship target scatter signals are received to obtain multi-dimensional observation information, including multi-angle, multi-band, multi-polarization of ship targets;
in the step (c), estimating the length and the width of the ship target to obtain the length direction of the ship target;
calculating scattering coefficients of the ship target at different angles, different frequency bands and different polarizations by using an electromagnetic scattering characteristic calculation unit (70) to obtain a multidimensional scattering coefficient of the ship target;
in the step (d), the target classification network is a convolutional neural network model, the vectorized multi-dimensional electromagnetic scattering set is input as the ship target, and the output is the type of the ship target.
3. The method according to claim 1, characterized in that the on-board GNSS-S radar (10) comprises:
the multi-channel azimuth antenna (101) is used for receiving multi-dimensional GNSS-S signals of a detection area, and has three frequency band signal receiving capabilities, wherein the frequencies of the three frequency bands are fc1, fc2 and fc3 respectively, and the corresponding working bandwidths are Bw1, Bw2 and Bw3 respectively; the antenna has left-hand circular polarization and right-hand circular polarization signal receiving capacity, and the polarization isolation degree is greater than 20 dB; the whole antenna is divided into M receiving antennas along the azimuth direction, wherein M is 4-8, each receiving antenna consists of Na multiplied by Nr antenna units, Na is 6-12, and Nr is 12-24;
the M six-channel receivers (102) are used for performing low-noise amplification, band-pass filtering, down-conversion, intermediate-frequency low-noise amplification, intermediate-frequency ADC (analog-to-digital converter) sampling and digital domain IQ (in-phase quadrature) demodulation on the GNSS-S signals of three frequency bands and two polarizations output by each receiving channel to obtain six-baseband GNSS-S complex signals;
the digital beam forming processing unit (103) is used for carrying out digital beam forming on GNSS-S signals of three frequency bands and two polarizations and forming P digital sub-beams in the azimuth direction;
the P matched filter banks (104) are used for performing matched filtering on the GNSS-S signals output by each digital sub-beam by using signals emitted by N navigation satellites as reference signals to obtain N independent GNSS-S signals with high signal-to-noise ratio and outputting a P frame signal set, wherein each frame signal set comprises 6N GNSS-S signals with high signal-to-noise ratio;
after signals of the M receiving antennas are subjected to digital beam forming, a plurality of high-gain narrow beams are formed, and the gain of each narrow beam is larger than 30 dB.
4. The method according to claim 1, characterized in that the digital beam forming processing unit (103) comprises:
a first P groups of complex multiplication and addition arithmetic units (1031) for left-handed circularly polarized M GNSS signals of frequency band fc1-S Signal SAm(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000031
a second P sets of complex multiply-add units (1032) for right-hand circularly polarized M GNSS-S signals SB in the frequency band fc1m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000032
a third P sets of complex multiply-add units (1033) for left-handed circularly polarized M GNSS-S signals SC in the frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000033
a fourth P sets of complex multiply-add units (1034) for right-hand circularly polarized M GNSS-S signals SD in the frequency band fc2m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000034
a fifth P sets of complex multiply-add units (1035) for left-handed circularly polarized M GNSS-S signals SE at frequency band fc3m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000035
a sixth P sets of complex multiply add units (1036) for right-hand circularly polarized M GNSS-S signals SF at frequency band fc3m(t) performing DBF processing along the azimuth direction to form P digital sub-beams, and outputting P GNSS-S signals with high signal-to-noise ratio as follows:
Figure FDA0003536346020000036
a DBF weight calculation unit (1037) for processing 6 sets of complex weights wa required for DBFp,m、wbp,m、wcp,m、wdp,m、wep,m、wfp,mCalculating;
wherein t is a fast time variable; m is a serial number, and the value is M1, 2. P is a serial number, and the value is P1, 2. SAm(t) is frequency band fc1, left hand circularly polarized M GNSS-S signals; RAp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the left-handed circularly polarized signals; wa (a)p,mThe complex weights are complex weights when DBF processing is carried out on the frequency band fc1 and the left-hand circularly polarized signals; SB (bus bar)m(t) is frequency band fc1, right hand circularly polarized M GNSS-S signals; RB (radio B)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc1 and the right-hand circularly polarized signals; wbp,mThe complex weight value is the complex weight value when the frequency band fc1 and the right-hand circularly polarized signal are subjected to DBF processing; SC (Single chip computer)m(t) M GNSS-S signals of left-hand circular polarization at frequency band fc 2; RC (resistor-capacitor) capacitorp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc2 and the left-handed circularly polarized signals; wc cp,mThe complex weight is the complex weight when the frequency band fc2 and the left-handed circularly polarized signal are subjected to DBF processing; SDm(t) M right-hand circularly polarized GNSS-S signals at frequency band fc 2; RDp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc2 and the right-hand circularly polarized signals; wdp,mThe complex weight value is the complex weight value when the frequency band fc2 and the right-hand circularly polarized signal are subjected to DBF processing; SEm(t) is frequency band fc3, left hand circular polarizationM GNSS-S signals; REp(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the left-handed circularly polarized signals; wep,mThe complex weight is the complex weight when the frequency band fc3 and the left-handed circularly polarized signal are subjected to DBF processing; SFm(t) M right-hand circularly polarized GNSS-S signals at frequency band fc 3; RF (radio frequency)p(t) P high signal-to-noise ratio GNSS-S signals after DBF processing is carried out on the frequency band fc3 and the right-hand circularly polarized signals; wfp,mThe complex weights are used for DBF processing of the frequency band fc3 and the right-hand circularly polarized signal.
5. The method of claim 1, wherein dual-station SAR imaging comprises:
a1, constructing a double-station SAR imaging geometric scene according to the position and speed information of a navigation satellite output by a satellite navigation positioning system and the position and speed information of a GNSS-S radar system, wherein the size of the imaging scene is X multiplied by Y, and the sizes of imaging grids are delta X multiplied by delta Y;
a2, performing parallelization time domain BP double-station imaging processing on the GNSS-S signal by utilizing P parallelization BP imaging processing units (202) to obtain 6N double-station SAR images;
the detection area is irradiated by N navigation satellite signals to form N double-station radars, each double-station radar comprises three frequency bands and two kinds of polarization information, and the parallelization BP imaging processing unit (202) comprises 6N time domain BP double-station imaging processing sub-modules in total.
6. The method of claim 1, wherein the multi-dimensional SAR image fusion process comprises:
b1, compensating the power path loss of the double-station SAR image at different observation angles;
b2, performing weight superposition on the 6N double-station SAR images by using a non-coherent processing method to obtain a high signal-to-noise ratio SAR image.
7. The method of claim 1, wherein the lengthwise extraction comprises:
c11, carrying out slice extraction on the detected ship target, and removing the sea clutter background;
c12, calculating the length-width ratio of the ship target;
c13, selecting the direction with the largest length-width ratio as the long axis direction of the ship target;
the extraction of the ship bow direction comprises the following steps:
c21, calculating the ship target position in each frame of the fused two-station SAR image;
c22, correlating the P frame ship target position information to obtain the motion track and the motion direction of the ship target;
c23, obtaining the ship bow direction according to the long axis direction and the motion direction of the ship target, and calculating the azimuth included angle between the qth ship and the GNSS-S radar receiving antenna beam center
Figure FDA0003536346020000051
The multi-dimensional scattering property calculation includes:
c31, respectively carrying out azimuth multi-view processing on the 6N double-station SAR images before fusion, wherein the multi-view times K are 3-10;
c32, calculating scattering coefficients of the ship target in the 6N multi-view processed double-station SAR images respectively to obtain scattering coefficients sigma 1 of the ship target for N double-station angles, three frequency bands and two polarizationsq,n、σ2q,n、σ3q,n、σ4q,n、σ5q,n、σ6q,nThe frequency band fc1 and the left-hand circularly polarized scattering coefficient, the frequency band fc1 and the right-hand circularly polarized scattering coefficient, the frequency band fc2 and the left-hand circularly polarized scattering coefficient, the frequency band fc2 and the right-hand circularly polarized scattering coefficient, the frequency band fc3 and the left-hand circularly polarized scattering coefficient, and the frequency band fc3 and the right-hand circularly polarized scattering coefficient, which are formed by the irradiation of the nth navigation satellite on the qth ship target, are respectively shown.
8. The method of claim 1, wherein in step (d), the multidimensional scattering property set of the ship target is constructed according to observation angle, working frequency band, polarization mode, navigation satellite position and speed, position and speed of the satellite-borne GNSS-S radar (10), ship target bow direction, and thenMultidimensional scattering characteristic set omega of q ship targetsqComprises the following steps:
Figure FDA0003536346020000061
Figure FDA0003536346020000062
wherein, Delta theta sq,n、
Figure FDA0003536346020000063
Respectively the incident angle and the azimuth angle of the nth navigation satellite signal irradiated on the qth ship target, wherein the azimuth angle is in reference to the ship bow direction of the ship target, and the delta theta sq,n＝θsn，
Figure FDA0003536346020000064
Δθrq、
Figure FDA0003536346020000065
Respectively receiving the incident angle and the azimuth angle of the qth ship target by the satellite-borne GNSS-S radar (10), wherein the azimuth angle is respectively based on the ship bow direction of the ship target, and the delta theta rq＝θrin，
Figure FDA0003536346020000066
Llp、LrpLeft hand circular polarization and right hand circular polarization, respectively.
9. The method according to claim 1, wherein in step (d), the received triple-frequency dual-polarization information and corresponding scattering coefficients of each navigational transmitting satellite are normalized and then input into the target classification network together;
the object classification network includes:
the multilayer convolution (901) is used for sequentially performing convolution, activation and pooling on input, extracting the characteristics including three-frequency point dual polarization and corresponding scattering coefficients, and sharing the parameters of the multilayer convolution (901) corresponding to the plurality of navigation transmitting satellites;
the multilayer perceptron (902) is used for carrying out multilayer perception on the characteristic information obtained by the multilayer convolution (901) in a parameter sharing mode;
realizing the feature fusion of a plurality of launching satellites;
and the classifier (904) is used for judging the class of the ship according to the fusion characteristics.
10. The method as claimed in claim 1, wherein the orbit height H of the satellite-borne GNSS-S radar (10) is 200-;
the antenna receives multi-dimensional GNSS-S signals of a ship target in a front side view mode, the multi-dimensional GNSS-S signals consist of M receiving channels in the azimuth direction, the wave beam in the azimuth direction of the antenna of each receiving channel is a wide wave beam, and the wave beam angle theta is larger than the wave beam angle thetaaGreater than 5 DEG, the total synthetic aperture length being LS；
Obtaining P sub-beams after digital beam forming processing is carried out on the M receiving channels, wherein each sub-beam corresponds to multiple frames of different position information of a ship target, and multiple frames of SAR images are obtained after double-station SAR imaging;
angle of incidence thetar at the center of the antenna beamin25-55 degrees, and the azimuth angle of the center of the antenna beam is 0;
the detection area is simultaneously irradiated by N navigation satellite signals, N double-station radar systems are formed by the detection area and the satellite-borne GNSS-S radar (10), and the incident angle of the nth navigation satellite signal is theta SnIn an azimuth of
Figure FDA0003536346020000071
CN202210219273.XA 2022-03-08 2022-03-08 Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship Active CN114488133B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210219273.XA CN114488133B (en) 2022-03-08 2022-03-08 Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210219273.XA CN114488133B (en) 2022-03-08 2022-03-08 Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship

Publications (2)

Publication Number Publication Date
CN114488133A true CN114488133A (en) 2022-05-13
CN114488133B CN114488133B (en) 2023-03-07

Family

ID=81486528

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210219273.XA Active CN114488133B (en) 2022-03-08 2022-03-08 Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship

Country Status (1)

Country Link
CN (1) CN114488133B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115728760A (en) * 2022-11-22 2023-03-03 北京卫星信息工程研究所 Tensor scattering information-based passive detection method for sea surface wind wave stream satellite
CN115755051A (en) * 2022-11-18 2023-03-07 北京卫星信息工程研究所 Distributed on-orbit processing method and system for double-satellite high-resolution wide-amplitude SAR (synthetic aperture radar) signals
CN115856967A (en) * 2022-11-14 2023-03-28 北京卫星信息工程研究所 Sea surface ship multi-station radar RCS measuring method and system based on GNSS signals
CN116125423A (en) * 2023-01-13 2023-05-16 东莘电磁科技(成都)有限公司 Scattered field characterization method of electromagnetic target

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120056780A1 (en) * 2006-04-28 2012-03-08 Paul Antonik Method and apparatus for simultaneous synthetic aperture radar and moving target indication
CN109444887A (en) * 2018-12-31 2019-03-08 成都汇蓉国科微***技术有限公司 Recognition methods and system are perceived based on star ship forward sight double-base SAR image-region
CN109507668A (en) * 2018-12-21 2019-03-22 曲卫 A kind of biradical imaging method based on navigation satellite signal
US20190101639A1 (en) * 2017-09-29 2019-04-04 United States Of America As Represented By The Administrator Of Nasa Spaceborne synthetic aperture radar system and method
CN112731398A (en) * 2021-01-11 2021-04-30 北京空间飞行器总体设计部 Multi-dimensional information detection SAR satellite detection method
CN113608216A (en) * 2021-06-25 2021-11-05 航天恒星科技有限公司 Satellite-borne multi-band common-caliber SAR and target combined on-orbit detection system and method
CN114002673A (en) * 2021-08-20 2022-02-01 航天恒星科技有限公司 Satellite-borne passive SAR non-cooperative signal sensing system and multi-dimensional parameter estimation method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120056780A1 (en) * 2006-04-28 2012-03-08 Paul Antonik Method and apparatus for simultaneous synthetic aperture radar and moving target indication
US20190101639A1 (en) * 2017-09-29 2019-04-04 United States Of America As Represented By The Administrator Of Nasa Spaceborne synthetic aperture radar system and method
CN109507668A (en) * 2018-12-21 2019-03-22 曲卫 A kind of biradical imaging method based on navigation satellite signal
CN109444887A (en) * 2018-12-31 2019-03-08 成都汇蓉国科微***技术有限公司 Recognition methods and system are perceived based on star ship forward sight double-base SAR image-region
CN112731398A (en) * 2021-01-11 2021-04-30 北京空间飞行器总体设计部 Multi-dimensional information detection SAR satellite detection method
CN113608216A (en) * 2021-06-25 2021-11-05 航天恒星科技有限公司 Satellite-borne multi-band common-caliber SAR and target combined on-orbit detection system and method
CN114002673A (en) * 2021-08-20 2022-02-01 航天恒星科技有限公司 Satellite-borne passive SAR non-cooperative signal sensing system and multi-dimensional parameter estimation method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ZHILONG ZHAO 等: "A novel method of ship detection by combining space-borne SAR and GNSS-R", 《IET INTERNATIONAL RADAR CONFERENCE (IET IRC 2020)》 *
夏正欢 等: "一种新型无人机高分SAR信号非均匀混合采样技术", 《现代雷达》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115856967A (en) * 2022-11-14 2023-03-28 北京卫星信息工程研究所 Sea surface ship multi-station radar RCS measuring method and system based on GNSS signals
CN115856967B (en) * 2022-11-14 2023-12-19 北京卫星信息工程研究所 GNSS signal-based sea surface ship multi-station radar RCS measurement method and system
CN115755051A (en) * 2022-11-18 2023-03-07 北京卫星信息工程研究所 Distributed on-orbit processing method and system for double-satellite high-resolution wide-amplitude SAR (synthetic aperture radar) signals
CN115728760A (en) * 2022-11-22 2023-03-03 北京卫星信息工程研究所 Tensor scattering information-based passive detection method for sea surface wind wave stream satellite
CN115728760B (en) * 2022-11-22 2023-08-11 北京卫星信息工程研究所 Sea surface storm satellite-borne passive detection method based on tensor scattering information
CN116125423A (en) * 2023-01-13 2023-05-16 东莘电磁科技(成都)有限公司 Scattered field characterization method of electromagnetic target
CN116125423B (en) * 2023-01-13 2023-09-01 东莘电磁科技(成都)有限公司 Scattered field characterization method of electromagnetic target

Also Published As

Publication number Publication date
CN114488133B (en) 2023-03-07

Similar Documents

Publication Publication Date Title
CN114488133B (en) Method for extracting and classifying multidimensional scattering characteristics of satellite-borne GNSS-S radar ship
CN111458711B (en) Satellite-borne dual-band SAR system and detection method of ship target
US20200341152A1 (en) Symmetrical Multistatic Radar Constellation for Earth Observation
CA3064739C (en) Apparatus and methods for a synthetic aperture radar with self-cueing
CN114895338B (en) Large-range sea surface wind field inversion system and method for satellite-borne GNSS-S radar multi-dimensional information
US20230176209A1 (en) Synthetic aperture radar imaging apparatus and methods for moving targets
EP2762912B1 (en) Device and method for collecting data for locating a source of interference
CN114910934B (en) Sea surface vector wind field inversion system and method based on satellite-borne GNSS-R/S integrated receiving
CN114660552B (en) Satellite-borne GNSS-S radar ship target signal receiving and direct interference suppression method
CN114488135B (en) Low-orbit small satellite distributed GNSS-S radar system and in-orbit processing method
FR2712755A1 (en) Radio Frequency Receiver and its control process.
CN113608216B (en) Satellite-borne multiband common-caliber SAR and target joint on-orbit detection system and method
CN114509754B (en) Satellite-borne multi-channel GNSS-S radar mass data on-orbit processing system and method
CN114637004B (en) Satellite-borne GNSS-S multi-station radar on-orbit processing and ship information fusion system and method
CN114488134A (en) Satellite-borne multi-channel GNSS-S radar video imaging system and ship track extraction method
CN115840226B (en) Directional multichannel ScanSAR rapid target detection method
Gogineni et al. Sounding and imaging of fast flowing glaciers and ice-sheet margins
FR2986334A1 (en) Radar i.e. X band synthetic aperture radar instrument, for use on carrier satellite for maritime surveillance mission, has selection unit to select instrument between small swath mode and large swath mode to allow ambiguous zone observation
Zhao et al. Wide-swath SAR based on networking of multiple small satellites for maritime applications
CN115728760B (en) Sea surface storm satellite-borne passive detection method based on tensor scattering information
CN114355402B (en) Satellite-borne multidimensional GNSS-S radar system and ship target detection method
Bai et al. Comparison of modeled and measured ocean surface DDMs from aircraft measurements of s-band satellite digital audio radio service transmissions
CN115825960A (en) Sea surface wind field inversion method based on satellite-borne GNSS-S radar
Wang MILLIMETER WAVE SYNTHETIC APERTURE RADAR TECHNOLOGY.
CN116299209A (en) Large-scale domain ship signal interference suppression method for airborne GNSS-S radar

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant