CN106772379B - Two-parameter CFAR Ship Detection based on Fourier transformation - Google Patents
Two-parameter CFAR Ship Detection based on Fourier transformation Download PDFInfo
- Publication number
- CN106772379B CN106772379B CN201710173725.4A CN201710173725A CN106772379B CN 106772379 B CN106772379 B CN 106772379B CN 201710173725 A CN201710173725 A CN 201710173725A CN 106772379 B CN106772379 B CN 106772379B
- Authority
- CN
- China
- Prior art keywords
- matrix
- clutter
- pixel
- target
- window
- 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.)
- Active
Links
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
- 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
- G01S13/9029—SAR image post-processing techniques specially adapted for moving target detection within a single SAR image or within multiple SAR images taken at the same time
-
- 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
-
- 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
- G01S13/9027—Pattern recognition for feature extraction
-
- 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
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/415—Identification of targets based on measurements of movement associated with the target
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)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a kind of two-parameter CFAR Ship Detection based on Fourier transformation, its thinking are as follows: obtain SAR radar return data, and imaging is carried out to SAR radar return data, SAR radar imagery data matrix I is obtained, and then obtain the bianry image matrix I after threshold processbw;IbwTo be marked with comprising naval vessel classification and the bianry image of pseudo- target;All 1's matrix is setIt is denoted as backdrop window;And it successively calculates the clutter number statistical matrix N in backdrop window and takes down the clutter number statistical matrix in operation rear backdrop windowAnd then calculate the clutter mean value statistical matrix M in background window;Then the clutter variance statistic matrix V in background window is calculated, and calculates naval vessel class target trip current F;Finally according to I and F, detection obtains multiple Ship Targets, each ships target detected is respectively Ship Target interested, so far completes the separation of interested Ship Target and pseudo- target, and then completes Ship Target Detection interested.
Description
Technical field
The invention belongs to Radar Signal Processing Technology field, in particular to a kind of two-parameter CFAR based on Fourier transformation
Ship Detection, i.e., the two-parameter constant false alarm rate (Constant False Alarm Rate, CFAR) based on Fourier transformation
Ship Detection is the object detection method in a kind of synthetic aperture radar, is suitable for synthetic aperture radar under Gaussian Background
Or the radar moving targets detection in Inverse Synthetic Aperture Radar.
Background technique
It is permanent in large scene high-resolution SAR (Synthetic Aperture Radar, SAR) image detection
False alarm rate (Constant False Alarm Rate, CFAR) algorithm is most widely used in SAR image object detection field
A kind of algorithm, including traditional CFAR algorithm and two-parameter CFAR detection algorithm;Wherein tradition CFAR algorithm specific implementation process is:
It is theoretical according to classical statistic mixed-state, under the conditions of given false-alarm probability, the ambient background clutter according to locating for target first
Statistical property adaptively seeks detection threshold value, then distinguishes each of SAR image pixel to be detected and adaptive threshold
It is compared, judges whether it is target point;Two-parameter CFAR detection algorithm is Gaussian Profile it is assumed that suitable based on background clutter
The SAR image detection complicated for local background's variation, and by after setting target window, protecting window and backdrop window with one
Fixed step-length is slided in SAR image;Wherein protecting window be in order to prevent the naval vessel partial compromise in target window to background
In window, the detection threshold of local window then is obtained by calculating the mean value of all pixels and variance in backdrop window, thus
The part for being higher than the thresholding in target window is determined as naval vessel.
Target window, protecting window and the backdrop window size of two-parameter CFAR detection algorithm are respectively necessary for through experience
It practises and obtaining, sliding step is also that experience obtains;For apart from close Ship Target, protecting window does not ensure that naval vessel part
It is actual that the mean value and variance that will not be leaked into backdrop window, therefore pass through clutter in statistics background window can not represent clutter
Value, will appear false-alarm and missing inspection in the detection process.
CFAR detection speed is to determine an important indicator of algorithm of target detection practicability, two-parameter CFAR detection algorithm
It is substantially the detection method of local sliding window, local dynamic station makes each pixel in SAR image repeatedly take part in sliding window fortune respectively
It calculates, leads to that CFAR algorithm operation quantity is big and computational efficiency is lower, and target detection speed is slow;What it is due to detection faces pair is whole picture
SAR image scene, if computational efficiency is low, certainly will affect SAR image target detection place so that the data volume of operation is very big
The performance of whole system reduces the real-time of its system, also increases the requirement of system for implementing hardware.
Summary of the invention
In view of the problems of the existing technology, i.e., two-parameter CFAR detector can not accurately examine short distance Ship Target
It surveys, and there is a problem of that computationally intensive, detection efficiency is low, it is a kind of double based on Fourier transformation it is an object of the invention to propose
Parameter CFAR Ship Detection, this kind can be detected quickly based on the two-parameter CFAR Ship Detection of Fourier transformation
Ship Target in SAR image, and calculation amount is small, detection efficiency is high.
To reach above-mentioned technical purpose, the present invention is realised by adopting the following technical scheme.
A kind of two-parameter CFAR Ship Detection based on Fourier transformation, comprising the following steps:
Step 1, SAR radar return data are obtained, and imaging is carried out to SAR radar return data, obtain SAR radar
Imaging data matrix I;Then threshold process is carried out to SAR radar imagery data matrix I, the bianry image after obtaining threshold process
Matrix Ibw;Bianry image matrix I after the threshold processbwTo be marked with comprising naval vessel classification and the bianry image of pseudo- target;
Step 2, all 1's matrix is setIt is denoted as backdrop window;And according to the bianry image matrix I after threshold processbw, meter
Calculation obtains the clutter number statistical matrix N in backdrop window, and then the clutter taken down in operation rear backdrop window is calculated
Number statistical matrix
Step 3, according to the bianry image matrix I after threshold processbwWith SAR radar imagery data matrix I, and takes and grasp
Make the clutter number statistical matrix in rear backdrop windowThe clutter mean value statistical matrix M in background window is calculated;
Step 4, according to the clutter number statistical matrix taken down in operation rear backdrop windowIt is equal with the clutter in background window
The clutter variance statistic matrix V in background window is calculated in Data-Statistics matrix M;
Step 5, according to the clutter mean value statistical matrix M in background window and the clutter variance statistic matrix V in background window, meter
Calculation obtains naval vessel class target trip current F;
Step 6, according to SAR radar imagery data matrix I and naval vessel class target trip current F, detection obtains multiple naval vessels
Target.
The present invention compared with the conventional method, has the advantage that
The method of the present invention can improve the detection effect of Ship Target in the case where Ship Target accurately detects, and reduce two-parameter
The operand of CFAR detection algorithm improves operation efficiency, and can fast implement the Ship Target Detection in SAR image.
Detailed description of the invention
Invention is further described in detail with reference to the accompanying drawings and detailed description.
Fig. 1 is a kind of two-parameter CFAR Ship Detection flow chart based on Fourier transformation of the invention;
Fig. 2 (a) is schematic diagram of the synthetic aperture radar image-forming in C-band,
Fig. 2 (b) is the result schematic diagram obtained after being detected using the two-parameter CFAR algorithm of tradition to Fig. 2 (a);
Fig. 2 (c) is to obtain after being detected using the two-parameter naval vessel the CFAR detection algorithm of improved SAR image to Fig. 2 (a)
Result schematic diagram;
Fig. 2 (d) is the result schematic diagram obtained after being detected using the method for the present invention to Fig. 2 (a);
Fig. 3 is using the two-parameter CFAR algorithm of tradition, improves two-parameter CFAR algorithm and the method for the present invention to different pixels
Several SAR images carries out the comparison diagram of time used in target detection respectively.
Specific embodiment
It referring to Fig.1, is a kind of two-parameter CFAR Ship Detection flow chart based on Fourier transformation of the invention;Institute
State down the two-parameter CFAR Ship Detection based on Fourier transformation, comprising the following steps:
Step 1, SAR radar return data are obtained, and imaging is carried out to SAR radar return data, obtain SAR radar
It include sea-surface target and non-sea-surface target in imaging data matrix I, the SAR radar imagery data matrix I;Then to SAR thunder
Reach as data matrix I progress threshold process, i.e., by the sea-surface target and non-sea-surface target in SAR radar imagery data matrix I
It separates, and then obtains the bianry image matrix I after threshold processbw;Bianry image matrix I after the threshold processbwFor packet
Classification containing naval vessel is marked with and the bianry image of pseudo- target.
Specifically, since SAR radar imagery data may obey different distributed models, corresponding various distributed models point
Not different, Gaussian Profile, quantum condition entropy, lognormal point with resolution ratio and the imaging parameters variation of SAR radar imagery
Cloth, K distribution can be used to fit object and detects the background clutter of target, and present example uses Gauss Distribution Fitting target
And detect the background clutter of target.
First obtain SAR radar return data, and to SAR radar return data carry out imaging, obtain SAR radar at
As data matrix I;It include sea-surface target and non-sea-surface target in the SAR radar imagery data matrix I;Pass through threshold process reality
The separation of sea-surface target and non-sea-surface target, the bianry image square after obtaining threshold process in existing SAR radar imagery data matrix I
Battle array Ibw;Wherein threshold process uses maximum between-cluster variance algorithm (i.e. Otsu threshold method), Otsu threshold method be it is current a kind of very
Popular threshold selecting algorithm, core concept is: determining an optimal threshold, completes two in SAR radar imagery data matrix I
After value processing, so that the inter-class variance between the background clutter class pixel and target class pixel of SAR radar imagery data matrix I
Maximum achievees the purpose that distinguish background clutter and target in SAR radar imagery data matrix I;Wherein, background clutter is sea
Target, target are non-sea-surface target;And then realize by SAR radar imagery data matrix I sea-surface target and non-sea mesh
Target separation.
The gray level for 1a) setting SAR radar imagery data matrix I is distributed as { 0:K }, and K takes 255 in the present embodiment;Wherein
Gray level isNumber of pixels beThen total number of pixels in SAR radar imagery data matrix I isGray level isPixel occur probability beAndAnd then it calculates
Obtain total average gray of SAR radar imagery data matrix I
1b) it is by gray level in SAR radar imagery data matrix IPixel threshold be denoted as For
Any one gray level in { 0:K } a gray level, and by the pixel grayscale of SAR radar imagery data matrix I
It is divided into two class pixels, is denoted asClass pixel andClass pixel, expression formula are respectively as follows:WhereinClass pixel represents background clutter,Class pixel represents target, target packet
The pseudo- target that target containing ships and strong sea clutter generate;K is the positive integer greater than 1.
Then it calculates separatelyThe probability that class pixel generatesThe probability that class pixel generatesClass picture
The mean value of elementWithThe mean value of class pixelIts expression formula is respectively as follows:
And then it is calculated
The inter-class variance of SAR radar imagery data matrix I isIts expression formula are as follows:
Wherein,For total average gray of SAR radar imagery data matrix I.
1c) the maximum between-cluster variance of SAR radar imagery data matrix I in order to obtain, need to pass through following optimization problem:
Bianry image matrix I after threshold process is calculatedbwOptimal threshold Topt;Arg is indicatedFor
Gray level is in corresponding SAR radar imagery data matrix I when maximum valuePixel thresholdValue, Max expression seek most
Big Value Operations.
1d) according to the bianry image matrix I after the threshold processbwOptimal threshold Topt, to SAR radar imagery data
Matrix I carries out threshold process, the bianry image matrix I after obtaining threshold processbw;Bianry image matrix after the threshold process
IbwBianry image for the pseudo- target for being marked with comprising naval vessel classification and being generated by strong sea clutter, and the two-value after the threshold process
Image array IbwMatrix is tieed up for m' × n';By the bianry image matrix I after threshold processbwIn any one pixel value note
For Ibw(i', j'), expression formula are as follows:
Wherein, SAR radar imagery data matrix I is that m × n ties up matrix, and (i, j) is to appoint in SAR radar imagery data matrix I
It anticipates a pixel coordinate, i ∈ { 1,2 ..., m }, j ∈ { 1,2 ..., n };(i', j') is the bianry image square after threshold process
Battle array IbwIn any one pixel coordinate, i' ∈ 1,2 ..., m'}, j' ∈ 1,2 ..., n'}, m indicate SAR radar imagery data
The line number of matrix I, n indicate that the columns of SAR radar imagery data matrix I, m' indicate the bianry image matrix I after threshold processbw
Line number, n' indicate threshold process after bianry image matrix IbwColumns;Bianry image matrix I after threshold processbwRow
Number, bianry image matrix I threshold process after equal with the line number value of SAR radar imagery data matrix IbwColumns and SAR
The columns value of radar imagery data matrix I is equal, and the bianry image matrix I after threshold processbwIn pixel and SAR radar
Pixel in imaging data matrix I corresponds;M, n, m', n' are respectively the positive integer for being greater than 0.
Step 2, all 1's matrix is tieed up in setting m " × n "It is denoted as backdrop window;And according to the bianry image square after threshold process
Battle array Ibw, the clutter number statistical matrix N in backdrop window is calculated, and then be calculated and take down in operation rear backdrop window
Clutter number statistical matrixWherein, 0≤m "≤m, 0≤n "≤n, m ", n " are respectively positive integer.
Specifically, it after the separation of sea-surface target and non-sea-surface target in SAR radar imagery data matrix I, obtains
The bianry image of pseudo- target for being marked with comprising naval vessel classification and being generated by strong sea clutter;Naval vessel classification is marked in order to reach and
By the separation for the pseudo- target that strong sea clutter generates, backdrop window is set, by the accurate statistics of clutter in the backdrop window, is come
Realize separation, process are as follows:
M " × n " is set and ties up all 1's matrixIt is denoted as backdrop window;And two-dimensional Fourier transform is carried out to two to backdrop window
Frequency domain is tieed up, two-dimensional frequency filtering matrix H is obtained;To the bianry image matrix I after threshold processbwCarry out two-dimensional Fourier transform behaviour
Make, obtains the bianry image matrix operated by two-dimensional Fourier transform;By the binary map by two-dimensional Fourier transform operation
As matrix and two-dimensional frequency filtering matrix H carry out dot product, the matrix S after obtaining dot product;Then two are carried out to the matrix S after dot product
Inverse Fourier transform is tieed up, the clutter number statistical matrix N in backdrop window is obtained;To the clutter number statistical moment in backdrop window
Each of battle array N element is taken respectively to be operated, and obtains taking down the clutter number statistical matrix in operation rear backdrop windowThe clutter number statistical matrix taken down in operation rear backdrop windowFor count in backdrop window the mean value of clutter and
Variance;Wherein, 0≤m "≤m, 0≤n "≤n, m ", n " are respectively positive integer.
2a) all 1's matrix is tieed up in setting m " × n "It is denoted as backdrop window, 0≤m "≤m, 0≤n "≤n, m ", n " is positive respectively
Integer;And two-dimensional Fourier transform is carried out to two-dimensional frequency to backdrop window, obtain two-dimensional frequency filtering matrix H;At threshold value
Bianry image matrix I after reasonbwTwo-dimensional Fourier transform operation is carried out, the binary map operated by two-dimensional Fourier transform is obtained
As matrix;By the bianry image matrix and two-dimensional frequency filtering matrix H progress dot product by two-dimensional Fourier transform operation, obtain
Matrix S after dot product;Then two-dimentional inverse Fourier transform is carried out to the matrix S after dot product, obtains the clutter in backdrop window
Number statistical matrix N, expression formula are as follows:
N=IFFT2 (H × FFT2 (Ibw))
Wherein, the clutter number statistical matrix N in backdrop window isMatrix is tieed up,Indicate the clutter in backdrop window
The line number of number statistical matrix N,Indicate the columns of the clutter number statistical matrix N in background window, the clutter in backdrop window
The line number of number statistical matrix N and the bianry image matrix I after threshold processbwLine number value it is equal, it is miscellaneous in backdrop window
The columns of wave number statistical matrix N and the bianry image matrix I after threshold processbwColumns value it is equal;FFT2 () is two
Fourier transform function is tieed up, IFFT2 () is two-dimentional inverse Fourier transform function.
Each of clutter number statistical matrix N in backdrop window element is taken respectively 2b) and is operated, is obtained
Take down the clutter number statistical matrix in operation rear backdrop windowThe clutter number system taken down in operation rear backdrop window
Count matrixForTie up matrix;The clutter number statistical matrix in operation rear backdrop window will be taken downIn any one
Pixel point value is denoted asIts expression formula are as follows:
Wherein, the clutter number statistical matrix N in backdrop window isMatrix is tieed up,Indicate the clutter in backdrop window
The line number of number statistical matrix N,Indicate the columns of the clutter number statistical matrix N in backdrop window,For backdrop window
Any one pixel point value in interior clutter number statistical matrix N, Expression, which takes, operates rear backdrop
Clutter number statistical matrix in windowLine number,It indicates to take down the clutter number statistical matrix in operation rear backdrop windowColumns;The line number of clutter number statistical matrix N in backdrop window with take down operation rear backdrop window in clutter number
Statistical matrixLine number value it is equal, the columns of the clutter number statistical matrix N in backdrop window and take and operate rear backdrop
Clutter number statistical matrix in windowColumns value it is equal, and in the clutter number statistical matrix N in backdrop window
Pixel with take down operation rear backdrop window in clutter number statistical matrixIn pixel correspond; Point
It Wei not positive integer greater than 0.
Step 3, according to the bianry image matrix I after threshold processbwWith SAR radar imagery data matrix I, and takes and grasp
Make the clutter number statistical matrix in rear backdrop windowThe clutter mean value statistical matrix M in backdrop window is calculated.
Specifically, by the bianry image matrix I after threshold processbwDot product behaviour is carried out with SAR radar imagery data matrix I
Make, obtains the first matrix of consequence of dot product operation;Two-dimensional Fourier transform is carried out to the first matrix of consequence of dot product operation, is obtained
The first matrix of consequence S ' after two-dimensional Fourier transform;By the first matrix of consequence S ' and two-dimensional frequency after two-dimensional Fourier transform
Filtering matrix H carries out dot product operation, the second matrix of consequence after obtaining dot product operation;To the second matrix of consequence after dot product operation
Carry out two-dimentional inverse Fourier transform, the second matrix of consequence after obtaining two-dimentional inverse Fourier transform;By two-dimentional inverse Fourier transform
Rear the second matrix of consequence with take down the clutter number statistical matrix in operation rear backdrop windowDot product is carried out, background window is obtained
Clutter mean value statistical matrix M in mouthful, calculation expression are as follows:
Wherein, FFT2 () is two-dimensional Fourier transform function, and IFFT2 () is two-dimentional inverse Fourier transform function.
Step 4, according to the clutter number statistical matrix taken down in operation rear backdrop windowWith the clutter in backdrop window
The clutter variance statistic matrix V in backdrop window is calculated in mean value statistical matrix M.
Specifically, square operation is carried out to the first matrix of consequence of dot product operation, the matrix of consequence after obtaining square operation;
Two-dimensional Fourier transform is carried out to the matrix of consequence after square operation, obtains the knot after square operation by two-dimensional Fourier transform
Fruit matrix S ";Matrix of consequence S " and two-dimensional frequency filtering matrix H after square operation by two-dimensional Fourier transform is carried out a little
Multiply, the third matrix of consequence after obtaining dot product operation;The inverse Fourier of two dimension is carried out to the third matrix of consequence after dot product operation to become
It changes, the third matrix of consequence after obtaining two-dimentional inverse Fourier transform;Then by the third result square after two-dimentional inverse Fourier transform
Battle array with take down operate after clutter number statistical matrixDot product operation is carried out, the 4th matrix of consequence after obtaining dot product operation;
The 4th matrix of consequence after dot product is operated and the clutter mean value statistical matrix M's in backdrop window square subtracts each other, and obtains dot product
The 5th matrix of consequence after operation;Square root finally is carried out to the 5th matrix of consequence after dot product operation and solves operation, and then is obtained
Clutter variance statistic matrix V in backdrop window, calculation expression are as follows:
Wherein, FFT2 () is two-dimensional Fourier transform function, and IFFT2 () is two-dimentional inverse Fourier transform function.
Step 5, according to the clutter mean value statistical matrix M in backdrop window and the clutter variance statistic matrix in backdrop window
Naval vessel class target trip current F is calculated in V.
5a) the CFAR method based on Gaussian Profile assumes the clutter Gaussian distributed in backdrop window, wherein Gauss point
The probability density function of cloth is f (x), expression formula are as follows:
Wherein, u is the clutter mean value in backdrop window, and σ is the clutter standard deviation in backdrop window, and exp indicates index letter
Number, x is function variable;And then obtain the clutter scores accumulated function F (x) in backdrop window, expression formula are as follows:
F (t) indicates that integral function, t indicate integration variable, and it is {-∞, x } that integration variable t, which obtains integrating range, and-∞ is negative nothing
Thoroughly, dt indicates the differential of t;Make substitution of variableThen obtain the clutter optimization scores accumulated function F' in backdrop window
(x), expression formula are as follows:
Wherein, Φ () indicates standard normal distribution function.
Set false alarm rate Pfa, 0≤Pfa≤ 1, usual false alarm rate PfaValue range be [10-8,10-2];Then according to
Lower formula acquires target detection threshold value Tcfar:
Simplification can obtain:
And then target detection threshold value T is calculatedcfar, calculation expression are as follows:
Tcfar=Φ-1(1-Pfa)×σ+u
Wherein, Φ-1The inverse function of () expression standard normal distribution function.
5b) according to the clutter mean value statistical matrix M in background window and the clutter variance statistic matrix V in background window, and
Target detection threshold value TcfarCalculation expression, naval vessel class target trip current F, the naval vessel class target discrimination square is calculated
Battle array F beTie up matrix, expression formula are as follows:
F=Φ-1(1-Pfa)×V+M
Step 6, according to SAR radar imagery data matrix I and naval vessel class target trip current F, detection obtains multiple naval vessels
Target.
Specifically, by each of SAR radar imagery data matrix I pixel point value respectively with naval vessel class target discrimination square
Corresponding pixel point value carries out pixel value comparison, specific decision process one by one in battle array F are as follows:
If the corresponding pixel point value I (i, j) of any one pixel coordinate (i, j) in SAR radar imagery data matrix I
Greater than any one pixel coordinate in target discrimination matrixCorresponding pixel point valueThen by pixel point value I (i,
J) corresponding pixel coordinate (i, j) is as the Ship Target pixel detected;, whereas if SAR radar imagery data matrix I
In the corresponding pixel point value I (i, j) of any one pixel coordinate (i, j) be less than target discrimination matrix in any one pixel
CoordinateCorresponding pixel point valueThen by the corresponding pixel coordinate (i, j) of pixel point value I (i, j) as by strong
The pseudo- object pixel that sea clutter generates.
Wherein, S AR radar imagery data matrix I is that m × n ties up matrix, and (i, j) is in SAR radar imagery data matrix I
Any one pixel coordinate, I (i, j) are that any one pixel coordinate (i, j) is corresponding in SAR radar imagery data matrix I
Pixel point value, i ∈ { 1,2 ..., m }, j ∈ { 1,2 ..., n };It is sat for any one pixel in target discrimination matrix
Mark,For any one pixel coordinate in target discrimination matrixCorresponding pixel point value, For the line number of naval vessel class target trip current F,For the columns of naval vessel class target trip current F, naval vessel
The line number of class target trip current F is equal with the line number value of SAR radar imagery data matrix I, naval vessel class target trip current F
Columns it is equal with the columns value of SAR radar imagery data matrix I, and the pixel in naval vessel class target trip current F with
Pixel in SAR radar imagery data matrix I corresponds.
Then, each of SAR radar imagery data matrix I and naval vessel class target trip current F pixel are traversed respectively
Point, and each of the SAR radar imagery data matrix I detected Ship Target pixel coordinate is successively recorded, and then
To the respective pixel of multiple Ship Targets, multiple Ship Targets are detected at this time, each ships target detected is respectively
Ship Target interested so far completes sea-surface target and non-sea-surface target (i.e. interested Ship Target and by strong sea
The pseudo- target that clutter generates) separation, and then complete Ship Target Detection interested.
Further verifying explanation is made to effect of the present invention below with reference to emulation experiment.
(1) measured data acquisition system parameter
In Ship Target Detection experiment, the method for the present invention will be with the two-parameter CFAR algorithm of tradition, improved two-parameter CFAR
Detection method is compared respectively, and emulation is detected using synthetic aperture radar image-forming under steady sea situation in the image of C-band,
Its clutter distinguishes Gaussian distributed;Intercept the SAR image that size is 720 × 869 pixels.
In order to remove interference of other targets to detection pixel, and consider Ship Target size and scene point in SAR image
Resolution is sized to 21 × 21 for protect window in experiment, and the size of background data window is 51 × 51;, the void that sets in detection
Alert rate Pfa value is 10-3。
(2) measured data object detection process content and result
Fig. 2 (a) is schematic diagram of the synthetic aperture radar image-forming in C-band, and Fig. 2 (b) is to be calculated using the two-parameter CFAR of tradition
The result schematic diagram that method obtains after detecting to Fig. 2 (a), Fig. 2 (c) are to use the two-parameter naval vessel CFAR of improved SAR image
The result schematic diagram that detection algorithm obtains after detecting to Fig. 2 (a), Fig. 2 (d) are to be carried out using the method for the present invention to Fig. 2 (a)
The result schematic diagram obtained after detection.
Wherein, there is no naval vessel partial compromise into backdrop window in the protecting window of Fig. 2 (c), detection process utilizes sliding window
It realizes;White box is the Ship Target being correctly detecting in Fig. 2 (b)-Fig. 2 (d), is the warship missed in detection in black round frame
Ship target is Ship Target false in detection in black rectangle frame.
In order to measure detection effect, the quality factor FoM of Ship Target Detection is defined,NttFor
The correct ships number of targets for including after detection, NfaFor false-alarm ships number of targets, NgtFor practical ships number of targets;Table 1 lists flat
Quiet sea level synthetic aperture radar imaging is compared in C-band image using the index of 3 kinds of detection methods;Synthetic aperture radar image-forming exists
The number of practical Ship Target is 18 in C-band image.
Table 1
Fig. 3 is using the two-parameter CFAR algorithm of tradition, improves two-parameter CFAR algorithm and the method for the present invention to different pixels
Several SAR images carries out the comparison diagram of time used in target detection respectively;Wherein separately include 256 × 256,512 × 512,
The SAR image of 1024 × 1024,2048 × 2048 pixel numbers uses the two-parameter CFAR algorithm of tradition, improved two-parameter CFAR
Algorithm and the method for the present invention carry out time contrast effect figure used in Ship Target Detection respectively.
(3) Analysis of test results
As can be seen that the method for the present invention and the two-parameter algorithm phase of tradition from Fig. 2 (b), Fig. 2 (c), Fig. 2 (d) and Fig. 3
Than: to the Ship Target of close proximity missing inspection will not occur for the method for the present invention, this is because clutter is to be extracted background when counting
Naval vessel part in window greatly improves detection so that not having naval vessel partial compromise into backdrop window in protecting window
Effect.
The method of the present invention is compared with two-parameter algorithm is improved, and common ground is all using the naval vessel extracted in backdrop window
Part carries out clutter statistics, there is identical detection effect;But the method for the present invention is realized by two-dimensional frequency operation
The process of sliding window in CFAR algorithm, can substantially reduce time complexity, effectively reduce the operand of CFAR detection algorithm, reduce
Detection time;Therefore, the method for the present invention fast implements while the distributed model statistical parameter for being accurately obtained background clutter
Ship Target Detection in SAR image has good practicability for the naval vessel detection of complicated sea situation.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
Mind and range;In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (3)
1. a kind of two-parameter CFAR Ship Detection based on Fourier transform, which comprises the following steps:
Step 1, SAR radar return data are obtained, and imaging is carried out to SAR radar return data, obtain SAR radar imagery
Data matrix I;Then threshold process is carried out to SAR radar imagery data matrix I, the bianry image matrix after obtaining threshold process
Ibw;Bianry image matrix I after the threshold processbwTo be marked with comprising naval vessel classification and the bianry image of pseudo- target;
It is described to obtain the bianry image matrix I after threshold processbw, sub-step are as follows:
The gray level for 1a) setting SAR radar imagery data matrix I is distributed as { 0:K }, and wherein gray level isNumber of pixels be Then total number of pixels in SAR radar imagery data matrix I is Gray level isPixel
The probability of appearance is AndAnd then total ash of SAR radar imagery data matrix I is calculated
Spend average value
1b) it is by gray level in SAR radar imagery data matrix IPixel threshold be denoted as It is a for { 0:K }
Any one gray level in gray level, and the pixel grayscale of SAR radar imagery data matrix I is divided into two class pixels, remember
ForClass pixel andClass pixel, expression formula are respectively as follows:WhereinClass pixel
Background clutter is represented,Class pixel represents target, and target includes ships target and the pseudo- target that strong sea clutter generates;K is big
In 1 positive integer;
Then it calculates separatelyThe probability that class pixel generatesThe probability that class pixel generatesClass pixel
Mean valueWithThe mean value of class pixelIts expression formula is respectively as follows:
And then SAR is calculated
The inter-class variance of radar imagery data matrix I isIts expression formula are as follows:
Bianry image matrix I after threshold process 1c) is calculatedbwOptimal threshold Topt, expression formula are as follows:
Arg is indicatedCorresponding SAR radar imagery data matrix I when for maximum value
Middle gray level isPixel thresholdValue, Max expression seek maxima operation;
1d) according to the bianry image matrix I after the threshold processbwOptimal threshold Topt, to SAR radar imagery data matrix I
Carry out threshold process, the bianry image matrix I after obtaining threshold processbw;Bianry image matrix I after the threshold processbwFor
The bianry image of pseudo- target for being marked with comprising naval vessel classification and being generated by strong sea clutter, and the bianry image after the threshold process
Matrix IbwMatrix is tieed up for m' × n';By the bianry image matrix I after threshold processbwIn the value of any one pixel be denoted as Ibw
(i', j'), expression formula are as follows:
Wherein, SAR radar imagery data matrix I is that m × n ties up matrix, and (i, j) is any one in SAR radar imagery data matrix I
A pixel coordinate, i ∈ { 1,2 ..., m }, j ∈ { 1,2 ..., n };(i', j') is the bianry image matrix I after threshold processbw
In any one pixel coordinate, i' ∈ 1,2 ..., m'}, j' ∈ 1,2 ..., n'}, m indicate SAR radar imagery data matrix
The line number of I, n indicate that the columns of SAR radar imagery data matrix I, m' indicate the bianry image matrix I after threshold processbwRow
Number, n' indicate the bianry image matrix I after threshold processbwColumns;Bianry image matrix I after threshold processbwLine number with
The line number value of SAR radar imagery data matrix I is equal, the bianry image matrix I after threshold processbwColumns and SAR radar
The columns value of imaging data matrix I is equal, and the bianry image matrix I after threshold processbwIn pixel and SAR radar imagery
Pixel in data matrix I corresponds;M, n, m', n' are respectively the positive integer for being greater than 0;
Step 2, all 1's matrix is setIt is denoted as backdrop window;And according to the bianry image matrix I after threshold processbw, calculate
Clutter number statistical matrix N in backdrop window, and then the clutter number system taken down in operation rear backdrop window is calculated
Count matrix
The all 1's matrixSpecially m " × n " ties up all 1's matrix
Clutter number statistical matrix N and the clutter number statistics taken down in operation rear backdrop window in the backdrop window
MatrixIt obtains process are as follows:
2a) m " × n " is set and ties up all 1's matrixIt is denoted as backdrop window, 0≤m "≤m, 0≤n "≤n, m ", n " are positive whole respectively
Number;And two-dimension fourier transform is carried out to two-dimensional frequency to backdrop window, obtain two-dimensional frequency filtering matrix H;To threshold process
Bianry image matrix I afterwardsbwTwo-dimension fourier transform operation is carried out, the bianry image operated by two-dimension fourier transform is obtained
Matrix;By the bianry image matrix and two-dimensional frequency filtering matrix H progress dot product by two-dimension fourier transform operation, obtain a little
Matrix S after multiplying;Then two-dimentional inverse Fourier transform is carried out to the matrix S after dot product, obtains the clutter number in backdrop window
Statistical matrix N, expression formula are as follows:
N=IFFT2 (H × FFT2 (Ibw))
Wherein, the clutter number statistical matrix N in backdrop window isMatrix is tieed up,Indicate the clutter number in backdrop window
The line number of statistical matrix N,Indicate the columns of the clutter number statistical matrix N in background window, the clutter number in backdrop window
The line number of statistical matrix N and the bianry image matrix I after threshold processbwLine number value it is equal, clutter in backdrop window
The columns and the bianry image matrix I after threshold process of number statistical matrix NbwColumns value it is equal;FFT2 () is two-dimentional Fu
Vertical leaf transformation function, IFFT2 () are two-dimentional inverse Fourier transform function;
Each of clutter number statistical matrix N in backdrop window element is taken respectively 2b) and is operated, obtains taking down
Operate the clutter number statistical matrix in rear backdrop windowThe clutter number statistical moment taken down in operation rear backdrop window
Battle arrayForTie up matrix;The clutter number statistical matrix in operation rear backdrop window will be taken downIn any one pixel
Point value is denoted asIts expression formula are as follows:
Wherein, the clutter number statistical matrix N in backdrop window isMatrix is tieed up,Indicate the clutter number in backdrop window
The line number of statistical matrix N,Indicate the columns of the clutter number statistical matrix N in backdrop window,For in backdrop window
Clutter number statistical matrix N in any one pixel point value, Expression takes down operation rear backdrop window
Clutter number statistical matrix in mouthfulLine number,It indicates to take down the clutter number statistical matrix in operation rear backdrop window
Columns;The line number of clutter number statistical matrix N in backdrop window is united with the clutter number in operation rear backdrop window is taken down
Count matrixLine number value it is equal, the columns of the clutter number statistical matrix N in backdrop window and take down operation rear backdrop window
Clutter number statistical matrix in mouthfulColumns value it is equal, and the picture in the clutter number statistical matrix N in backdrop window
Element with take down operation rear backdrop window in clutter number statistical matrixIn pixel correspond; Point
It Wei not positive integer greater than 0;
Step 3, according to the bianry image matrix I after threshold processbwWith SAR radar imagery data matrix I, and take down operation after
Clutter number statistical matrix in backdrop windowThe clutter mean value statistical matrix M in background window is calculated;
Clutter mean value statistical matrix M in the backdrop window, obtains process are as follows:
By the bianry image matrix I after threshold processbwDot product operation is carried out with SAR radar imagery data matrix I, obtains dot product behaviour
The first matrix of consequence made;Two-dimension fourier transform is carried out to the first matrix of consequence of dot product operation, obtains two-dimension fourier change
The first matrix of consequence S ' after changing;By after two-dimension fourier transform the first matrix of consequence S ' and two-dimensional frequency filtering matrix H into
The operation of row dot product, the second matrix of consequence after obtaining dot product operation;It is inverse that two dimension is carried out to the second matrix of consequence after dot product operation
Fourier transform, the second matrix of consequence after obtaining two-dimentional inverse Fourier transform;By the second knot after two-dimentional inverse Fourier transform
Fruit matrix with take down operation rear backdrop window in clutter number statistical matrixDot product is carried out, the clutter in backdrop window is obtained
Mean value statistical matrix M, calculation expression are as follows:
Wherein, FFT2 () is two-dimension fourier transform function, and IFFT2 () is two-dimentional inverse Fourier transform function;
Step 4, according to the clutter number statistical matrix taken down in operation rear backdrop windowWith the equal primary system of clutter in background window
Matrix M is counted, the clutter variance statistic matrix V in background window is calculated;
Clutter variance statistic matrix V in the background window, obtains process are as follows:
Square operation is carried out to the first matrix of consequence of dot product operation, the matrix of consequence after obtaining square operation;To square operation
Rear matrix of consequence carries out two-dimension fourier transform, obtains the matrix of consequence S " after square operation by two-dimension fourier transform;
Dot product will be carried out by the matrix of consequence S " of two-dimension fourier transform and two-dimensional frequency filtering matrix H after square operation, obtained a little
Third matrix of consequence after multiplying operation;Two-dimentional inverse Fourier transform is carried out to the third matrix of consequence after dot product operation, obtains two
Third matrix of consequence after tieing up inverse Fourier transform;Then the third matrix of consequence after two-dimentional inverse Fourier transform is grasped with taking
Clutter number statistical matrix after workDot product operation is carried out, the 4th matrix of consequence after obtaining dot product operation;Dot product is operated
The 4th matrix of consequence afterwards and the clutter mean value statistical matrix M's in backdrop window square subtracts each other, and the after obtaining dot product operation
Five matrixs of consequence;Square root finally is carried out to the 5th matrix of consequence after dot product operation and solves operation, and then obtains backdrop window
Interior clutter variance statistic matrix V, calculation expression are as follows:
Wherein, FFT2 () is two-dimension fourier transform function, and IFFT2 () is two-dimentional inverse Fourier transform function;
Step 5, it according to the clutter mean value statistical matrix M in background window and the clutter variance statistic matrix V in background window, calculates
To naval vessel class target trip current F;
Step 6, according to SAR radar imagery data matrix I and naval vessel class target trip current F, detection obtains multiple Ship Targets.
2. a kind of two-parameter CFAR Ship Detection based on Fourier transform as described in claim 1, which is characterized in that
In steps of 5, the naval vessel class target trip current F, expression formula are as follows:
F=Φ-1(1-Pfa)×V+M
Wherein, Φ-1() indicates the inverse function of standard normal distribution function, PfaFor the false alarm rate of setting, 0≤Pfa≤1。
3. a kind of two-parameter CFAR Ship Detection based on Fourier transform as claimed in claim 2, which is characterized in that
In step 6, the detection obtains multiple Ship Targets, process are as follows:
Each of SAR radar imagery data matrix I pixel point value is corresponding with naval vessel class target trip current F respectively
Pixel point value carries out pixel value comparison, specific decision process one by one are as follows:
If the corresponding pixel point value I (i, j) of any one pixel coordinate (i, j) is greater than in SAR radar imagery data matrix I
Any one pixel coordinate in target discrimination matrixCorresponding pixel point valueIt is then that pixel point value I (i, j) is right
The pixel coordinate (i, j) answered is as the Ship Target pixel detected;, whereas if appointing in SAR radar imagery data matrix I
The corresponding pixel point value I (i, j) of a pixel coordinate (i, j) of anticipating is less than any one pixel coordinate in target discrimination matrixCorresponding pixel point valueThen by the corresponding pixel coordinate (i, j) of pixel point value I (i, j) as pseudo- target picture
Element;
Wherein, SAR radar imagery data matrix I is that m × n ties up matrix, and I (i, j) is any in SAR radar imagery data matrix I
The corresponding pixel point value of one pixel coordinate (i, j), i ∈ { 1,2 ..., m }, j ∈ { 1,2 ..., n };For target discrimination
Any one pixel coordinate in matrix,For any one pixel coordinate in target discrimination matrixCorresponding picture
Vegetarian refreshments value, For the line number of naval vessel class target trip current F,For naval vessel class target
The columns of trip current F, the line number and the line number value phase of SAR radar imagery data matrix I of naval vessel class target trip current F
Deng the columns of naval vessel class target trip current F is equal with the columns value of SAR radar imagery data matrix I, and naval vessel class target
The pixel in pixel and SAR radar imagery data matrix I in trip current F corresponds;
Then, each of SAR radar imagery data matrix I and naval vessel class target trip current F pixel are traversed respectively, and
Each of the SAR radar imagery data matrix I detected Ship Target pixel coordinate is successively recorded, and then is obtained multiple
The respective pixel of Ship Target, detects multiple Ship Targets at this time.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710173725.4A CN106772379B (en) | 2017-03-22 | 2017-03-22 | Two-parameter CFAR Ship Detection based on Fourier transformation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710173725.4A CN106772379B (en) | 2017-03-22 | 2017-03-22 | Two-parameter CFAR Ship Detection based on Fourier transformation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106772379A CN106772379A (en) | 2017-05-31 |
CN106772379B true CN106772379B (en) | 2019-08-02 |
Family
ID=58966489
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710173725.4A Active CN106772379B (en) | 2017-03-22 | 2017-03-22 | Two-parameter CFAR Ship Detection based on Fourier transformation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106772379B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107942329B (en) * | 2017-11-17 | 2021-04-06 | 西安电子科技大学 | Method for detecting sea surface ship target by maneuvering platform single-channel SAR |
CN110992387B (en) * | 2019-11-08 | 2022-06-10 | 浪潮电子信息产业股份有限公司 | Image processing method and device, electronic equipment and storage medium |
CN111458711B (en) * | 2020-04-24 | 2022-03-04 | 北京卫星信息工程研究所 | Satellite-borne dual-band SAR system and detection method of ship target |
CN112649791B (en) * | 2020-12-24 | 2022-08-05 | 北京海兰信数据科技股份有限公司 | Radar echo processing method and device |
CN113256667B (en) * | 2021-04-14 | 2023-07-21 | 大连海事大学 | SAR image ship target detection method based on Otsu and joint distribution |
CN113674308B (en) * | 2021-05-06 | 2024-02-13 | 西安电子科技大学 | SAR image ship target rapid detection method based on image enhancement and multiple detection |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4062012A (en) * | 1974-07-11 | 1977-12-06 | The United States Of America As Represented By The Secretary Of The Navy | Digital radar signal processor |
KR100888674B1 (en) * | 2007-07-02 | 2009-03-16 | (주)인펙비전 | Method for measuring similarity using frenquency range |
CN102768356B (en) * | 2012-08-03 | 2013-08-28 | 南京大学 | Method for detecting ocean stationary targets by multi-temporal radar data |
CN102938150A (en) * | 2012-11-20 | 2013-02-20 | 西安电子科技大学 | Synthetic aperture radar (SAR) image ship detection method based on self-adaptation sea clutter statistics |
CN103996166A (en) * | 2014-06-05 | 2014-08-20 | 上海理工大学 | Holographic anti-counterfeiting algorithm based on shape of screen dot |
CN105974410B (en) * | 2016-04-28 | 2018-07-20 | 西安电子科技大学 | The more Ship Target SAR and ISAR mixing imaging methods of airborne radar |
-
2017
- 2017-03-22 CN CN201710173725.4A patent/CN106772379B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN106772379A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106772379B (en) | Two-parameter CFAR Ship Detection based on Fourier transformation | |
CN105354541B (en) | The SAR image object detection method of view-based access control model attention model and constant false alarm rate | |
Marra et al. | Do gans leave artificial fingerprints? | |
Qin et al. | Effective infrared small target detection utilizing a novel local contrast method | |
CN107229918B (en) | SAR image target detection method based on full convolution neural network | |
CN101727662B (en) | SAR image nonlocal mean value speckle filtering method | |
Gao | A parzen-window-kernel-based CFAR algorithm for ship detection in SAR images | |
Hou et al. | Multilayer CFAR detection of ship targets in very high resolution SAR images | |
CN104408482B (en) | A kind of High Resolution SAR Images object detection method | |
CN106780485A (en) | SAR image change detection based on super-pixel segmentation and feature learning | |
CN110263646A (en) | A kind of sea weak target detection method and system based on convolutional neural networks | |
CN113256667B (en) | SAR image ship target detection method based on Otsu and joint distribution | |
CN107784655A (en) | A kind of visual attention model SAR naval vessels detection algorithm of adaptive threshold | |
CN106842143B (en) | The quick CFAR detection method of radar target based on filtering | |
CN103971364A (en) | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters | |
CN113674308B (en) | SAR image ship target rapid detection method based on image enhancement and multiple detection | |
CN110197185A (en) | A kind of method and system based on Scale invariant features transform algorithm monitoring space under bridge | |
CN108399430A (en) | A kind of SAR image Ship Target Detection method based on super-pixel and random forest | |
CN106291550B (en) | The polarization SAR Ship Detection of core is returned based on local scattering mechanism difference | |
CN104881670B (en) | A kind of fast target extracting method for SAR orientation angular estimation | |
CN109978855A (en) | A kind of method for detecting change of remote sensing image and device | |
Li et al. | Outlier-robust superpixel-level CFAR detector with truncated clutter for single look complex SAR images | |
CN107729903A (en) | SAR image object detection method based on area probability statistics and significance analysis | |
CN107369163A (en) | A kind of quick SAR image object detection method based on best entropy Double Thresholding Segmentation | |
CN108985292A (en) | A kind of SAR image CFAR object detection method and system based on multi-scale division |
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 |