CN107274401A - A kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism - Google Patents
A kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism Download PDFInfo
- Publication number
- CN107274401A CN107274401A CN201710482797.7A CN201710482797A CN107274401A CN 107274401 A CN107274401 A CN 107274401A CN 201710482797 A CN201710482797 A CN 201710482797A CN 107274401 A CN107274401 A CN 107274401A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mfrac
- detection
- probability
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10044—Radar image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
Abstract
The invention discloses a kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism, the technology belongs to radar image object detection field.The intelligent demand and requirement of real-time of detection can not be met mainly for existing High Resolution SAR Images target detection, the notable model of vision is extracted in the spectral residuum part that target information is contained in image by the Fourier transformation based on frequency domain；Then notable figure binary conversion treatment and region of interesting extraction are carried out to notable figure, design a local maxima posterior probability grader from classification angle analysis carries out target detection to potential target region, realizes and detects through parameter Estimation, decision rule.This method can improve the real-time and accuracy of High Resolution SAR Images Ship Target Detection, be prevented effectively from false-alarm problem higher in detection.
Description
Technical field
The present invention is under the jurisdiction of diameter radar image object detection field, and being related to one kind, to meet radar image target real-time
Property detection demand detection method, it is adaptable to include warship under the complicated sea situation such as a large amount of uneven sea clutter regions and speckle noise
The high resolution synthetic aperture radar image of ship detection monitoring.
Background technology
Synthetic aperture radar (Synthetic aperture radar, SAR) has round-the-clock, round-the-clock, a wide range of etc.
Feature, is the important component of marine monitoring and monitoring, and wherein Ship Target Detection is increasingly becoming study hotspot.SAR image
Ship Target Detection is premise and the basis of its classification and identification, is the importance of SAR image application.
Run with the transmitting of a new generation such as Radarsat-2, TerraSAR-X and high score three SAR sensors, SAR
Gradually develop to high-resolution, big breadth, multipolarization direction.Become larger however as High Resolution SAR Images size, tradition
Ground is slow using the node-by-node algorithm processing speed based on image, the SAR image information and limited computer disposal of big data quantity
Contradiction between ability, it is difficult to reach the requirement handled in real time.Secondly, traditional low resolution SAR image detection method at this stage
Detection accuracy is not high during applied to High Resolution SAR Images, is improving by speckle noise and uneven sea clutter background pair
Deficiency is still had in terms of the false-alarm that testing result is brought, it is difficult to meet the accurate Intelligent Measurement demand of target in image.
The content of the invention
The present invention proposes a kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism, it is intended to solve
Many false-alarm problems in existing SAR image target detection, meet the intelligent demand of detection and requirement of real-time.
A kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism of the present invention, specific bag
Include following technical measures：The acquisition of the notable model of vision asks for algorithm, base using a kind of global marking area of spectral residuum method
Realized in the Fast Fourier Transform (FFT) of frequency domain, target is contained into the part that same shape is different from the frequency spectrum of piece image
Extracted the spectral residuum part of information；Two step operations are carried out to obtained notable figure according to the image property of notable figure,
That is the extraction of the binary conversion treatment of notable figure and area-of-interest.Complete this step need to be by carrying out twice threshold point to notable figure
Cut processing to realize, first by first threshold value by potential naval vessel region segmentation in visual saliency map, second of Threshold segmentation leads to
Cross two threshold values the pixel in notable figure is classified, it is follow-up to be accurately performed by being screened to sorted part
Target area and the approximate fits of background area grey level histogram；According to the theory of machine learning, analyzed with the angle of classification
Ship Target Detection problem, the maximum a posteriori probability grader for designing a part further carries out warship to marking area in image
Ship detection, target detection problems is converted into the binary hypothesis test problem of each pixel of potential target region；According to
The decision rule of grader, belongs to priori of all categories to the conditional probability and pixel to be measured for giving pixel to be measured under classification
Probability carries out parameter Estimation and asked for, the classifier parameters obtained with reference to estimation, in the naval vessel potential target region of marking area
Pixel realizes secondary detection.
A kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism proposed by the present invention, can improve
The real-time and accuracy of High Resolution SAR Images Ship Target Detection, use for reference vision attention theory and machine Learning Theory is entered
Row image procossing, is prevented effectively from false-alarm problem higher in detection.
Brief description of the drawings
Fig. 1 is a kind of High Resolution SAR Images naval vessel overhaul flow chart of view-based access control model attention mechanism；
Fig. 2 is a kind of High Resolution SAR Images naval vessel detects schematic diagram of view-based access control model attention mechanism.
Embodiment
A kind of technical side of the High Resolution SAR Images Ship Detection of view-based access control model attention mechanism proposed by the present invention
Case comprises the following steps：
Step 1.1：Computation model asks for algorithm using the marking area of global search mode, is that one kind is based on image frequency domain
Vision significance method for extracting region, realized using Fast Fourier Transform (FFT), it is generally the case that have in the frequency spectrum of piece image
It is not the spectral residuum part for containing target information in the part of same shape, the notable model of vision will be given this residual fraction
To extract；Assuming that I (x) is piece image, the frequency spectrum FFT [I (x)] of image is decomposed into amplitude spectrum A (f) and phase spectrum P (f) two
Part；
P (f)=FFT [I (x)]/| { FFT [I (x)] } | (1)
R (f)=BP (f) P (f) (2)
S (f)=FFT-1[R(f)] (3)
Wherein, FFT and FFT-1Fast Fourier Transform (FFT) and its inverse transformation of image are represented, P (f) is original image phase spectrum, R
(f) spectral residuum is represented, BP (f) is bandpass filter, uses centre frequency for f in model0, cut-off frequency be △ f Gauss
Wave filter；S (f) is notable figure；The High Resolution SAR Images Ship Detection intermediate frequency of view-based access control model attention mechanism in the present invention
Compose residual error conspicuousness computation model and mainly include two step computings：One is the normalized of original signal spectrum, and two be the filter of frequency domain band logical
Ripple；
Step 1.2：Visual saliency map is obtained using previous action intermediate frequency spectrum residual error method, two are carried out to obtained notable figure
Step operation：One is the binary conversion treatment of notable figure；Two be the extraction of area-of-interest.We are split using twice threshold, first
Marking area in visual saliency map is split, with realize from visual attention computation model filter out image in conspicuousness
Region is potential naval vessel region.Second of Threshold segmentation is divided the pixel in notable figure by setting two threshold values
Cut, more accurately complete the approximate of follow-up target area and background area grey level histogram.
Step 2.1：The thought classified from machine learning analyzes Ship Target Detection problem, and Ship Target Detection is
One two class classification problem, one local maxima posterior probability grader of design is to potential after the Threshold segmentation of visual salient region
Ship Target region carries out secondary detection：According to bayesian theory, High Resolution SAR Images Ship Target Detection problem is converted
For the binary hypothesis test problem to data vector x；Data sample is divided into two classes, sample class ωiRespectively ω1And ω0, mark
The sample class for being designated as target is ω1, the sample class labeled as background is ω0；If P (ωi) represent that input pixel belongs to ωi
Prior probability, this dualism hypothesis detection bayesian criterion be：
Wherein, P (ω1| x) with P (ω0| x) refer to respectively detected pixel be target and background posterior probability, P (x | ωi)
It is in given classification ωiUnder conditional probability, P (x) refer to obtain pixel probability；
According to bayesian criterion and maximum a-posteriori estimation criterion, grader is defined as：
The condition met in the presence of target is：
Conversely, the condition met when target is not present is：
The decision rule of local maxima posterior probability grader is：
Step 2.2：Ask for pixel to be measured under given classification conditional probability P (x | ωi)：
Local classifiers are carried out with parameter Estimation, the first step need to ask for conditional probability P (x | ωi), as target and the back of the body
The probability density function of scape, sets two threshold value T1、T2, wherein T2>T1；Enter row threshold division and corresponding original image to notable figure
Middle approximate target and background area grey level histogram fit operation, this process are all modeled to background and target：
A) notable figure is split using two threshold values, extracts in image to be detected and be more than threshold value T in correspondence notable figure1Portion
The all pixels divided, obtain grey level histogram, are distributed using Gamma, Weibull is distributed, Log-Normal distribution equal distribution moulds
Type carries out approximate gray-scale histogram-fitting, to ask for the probability distribution of Ship Target；
B) such as above-mentioned sub-step a) method, extract in original image and be less than threshold value T in correspondence notable figure2Region in institute
Have pixel, as the approximate of background area, carry out grey level histogram fitting, as background probability be distributed it is approximate；
C) probability distribution according to obtained by histogram-fitting asks for Ship Target and the probability density function of clutter background, obtains
Need to ask for into grader target and background conditional probability P (x | ωi)。
Step 2.3：Ask for the prior probability P (ω that pixel to be measured belongs to of all categoriesi)：
Pixel to be measured is asked for by a sliding window and belongs to ωiPrior probability, prior probability P (ωi) be defined as：
Wherein, xtRepresent current grey scale pixel value to be detected, x1、x2,...,xNIt is all in sliding window to belong to background area
The pixel in domain.G is SAR image gray level maximum, and a is an empirical parameter for adjusting prior probability, a ∈ (0,1]；This step
Carry out the calculating of prior probability in operation to the potential region of target using sliding window, parameters need foundation in sliding window design
Depending on elemental area, size and its distribution situation in SAR image shared by Ship Target；
Step 2.4：With reference to obtained prior probability and conditional probability, to picture in the naval vessel potential target region of marking area
Vegetarian refreshments realizes secondary detection, using the local maxima posterior probability grader of design to all pixels point in marking area in image
Ship Target Detection is carried out, last testing result is obtained.
Claims (3)
1. a kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism, it is characterised in that including following step
Suddenly：
Step 1：A kind of High Resolution SAR Images marking area detection algorithm of global search is designed, passes through the normalization of frequency spectrum
Processing and frequency domain bandpass filtering obtain the notable computation model of spectral residuum vision, the area-of-interest of quick obtaining vision；
Step 2：With reference to the binary hypothesis test thought in bayesian theory, a local maxima posterior probability grader is designed,
Pixel two in marking area is completed through parameter Estimation, decision rule to classify to realize target detection.
2. a kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism as claimed in claim 1, it is special
Levy and be, the step 1 specifically includes following sub-step：
Step 1.1：The marking area that model is searched for using a kind of global scope asks for algorithm, and computation model is handled based on frequency domain,
Realized using Fast Fourier Transform (FFT), target information is contained into the part that same shape is different from the frequency spectrum of piece image
Spectral residuum part extracted；Assuming that I (x) is piece image, the frequency spectrum FFT [I (x)] of image is resolved into amplitude spectrum A
(f) with phase spectrum P (f) two parts,
P (f)=FFT [I (x)]/| { FFT [I (x)] } | (1)
R (f)=BP (f) P (f) (2)
S (f)=FFT-1[R(f)] (3)
Wherein, FFT and FFT-1Fast Fourier Transform (FFT) and its inverse transformation of image are represented, P (f) refers to the phase spectrum of original image, R
(f) spectral residuum is represented, BP (f) is bandpass filter, uses a centre frequency for f in model0, cut-off frequency is △ f
Gaussian filter, S (f) is notable figure, and the notable model of spectral residuum mainly includes two step computings：The normalized and frequency of frequency spectrum
Domain bandpass filtering；
Step 1.2：Visual saliency map is obtained by previous step spectral residuum method, two step operations are carried out to obtained notable figure：
That is the extraction of the binary conversion treatment of notable figure and area-of-interest, binary conversion treatment and acquisition interested are by using two subthresholds
Value carries out notable figure segmentation and obtained, and extracts the marking area in visual saliency map split first, mould is calculated according to vision attention
Type filters out potential Ship Target region in image；Second of Threshold segmentation is realized by two threshold values, by notable figure
Pixel is classified, and more accurately completes the grey level histogram approximate fits of follow-up target area and background area.
3. a kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism as claimed in claim 1, it is special
Levy and be, the step 2 specifically includes following sub-step：
Step 2.1：Ship Target Detection problem, the maximum a posteriori probability classification of one part of design are analyzed from the angle of classification
Device further carries out Ship Target Detection to marking area in image：According to bayesian theory, target detection problems are converted into
To data vector x binary hypothesis test problem；Data sample is divided into two classes, sample class is respectively ω1And ω0If, P
(ωi) represent that input pixel belongs to ωiPrior probability, this dualism hypothesis detection bayesian criterion be：
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>1</mn>
</msub>
<mo>|</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<msub>
<mi>&omega;</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>&CenterDot;</mo>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>0</mn>
</msub>
<mo>|</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<msub>
<mi>&omega;</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>&CenterDot;</mo>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, P (ω1| x) with P (ω0| x) refer to respectively detected pixel be target and background posterior probability, P (x | ωi) be
Given classification ω0Under conditional probability, P (x) refer to obtain pixel probability；
According to bayesian criterion and maximum posteriori criterion, grader may be defined as：
<mrow>
<mi>Y</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mi>argmax</mi>
<mi>i</mi>
</munder>
<mrow>
<mo>(</mo>
<mi>P</mi>
<mo>(</mo>
<mrow>
<msub>
<mi>&omega;</mi>
<mi>i</mi>
</msub>
<mo>|</mo>
<mi>x</mi>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
</mrow>
1
The condition met in the presence of target is：
<mrow>
<mfrac>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>1</mn>
</msub>
<mo>|</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mn>0</mn>
</msub>
<mo>|</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>></mo>
<mn>1</mn>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
The decision rule that maximum a posteriori probability grader is used for：
<mrow>
<mfrac>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<msub>
<mi>w</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<msub>
<mi>w</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<munderover>
<mtable>
<mtr>
<mtd>
<mo>></mo>
</mtd>
</mtr>
<mtr>
<mtd>
<mo><</mo>
</mtd>
</mtr>
</mtable>
<mn>0</mn>
<mn>1</mn>
</munderover>
<mfrac>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>w</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>w</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
</mrow>
Step 2.2：Ask for pixel to be measured under given classification conditional probability P (x | ωi)：
Conditional probability P (x | ωi) be target and background probability density function, set two threshold value T1、T2, wherein T2>T1；
A) notable figure is split using two threshold values, extracts in image to be detected and be more than threshold value T in correspondence notable figure1Partial
All pixels, obtain grey level histogram；It is distributed using Gamma, Weibull is distributed, Log-Normal distribution approximate gray-scale Nogatas
Figure fitting, to ask for the probability distribution of Ship Target；
B) such as above-mentioned sub-step a) method, extract correspondence notable figure in artwork and be less than threshold value T2Region in all pixels, enter
Row grey level histogram be fitted, as background probability be distributed it is approximate；
C) probability distribution according to obtained by histogram-fitting asks for Ship Target and the probability density function of clutter background.
Step 2.3：Ask for the prior probability P (ω that pixel to be measured belongs to of all categoriesi)：
Pixel to be measured is asked for by a sliding window and belongs to ωiPrior probability, prior probability P (ωi) be defined as：
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&omega;</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mfrac>
<mi>a</mi>
<mi>G</mi>
</mfrac>
<mo>&CenterDot;</mo>
<mo>|</mo>
<msub>
<mi>x</mi>
<mi>t</mi>
</msub>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<munderover>
<mi>&Sigma;</mi>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mfrac>
<mi>a</mi>
<mi>G</mi>
</mfrac>
<mo>&CenterDot;</mo>
<mo>|</mo>
<msub>
<mi>x</mi>
<mi>t</mi>
</msub>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mi>N</mi>
</mfrac>
<munderover>
<mi>&Sigma;</mi>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>|</mo>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, xtRepresent current grey scale pixel value to be detected, x1、x2,...,xNIt is all in sliding window to belong to background area
Pixel；G is SAR image gray level maximum, and a is an empirical parameter for adjusting prior probability, a ∈ (0,1]；In this step
In, the calculating of prior probability is carried out to the potential region of target using sliding window, parameters are needed according to SAR in sliding window design
Depending on elemental area, size and its distribution situation in image shared by Ship Target；
Step 2.4：With reference to obtained prior probability and conditional probability, to pixel in the naval vessel potential target region of marking area
Secondary detection is realized, using the local maxima posterior probability grader of design to all pixels in marking area in image in detection
Point carries out Ship Target Detection, and last testing result is obtained by decision rule.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710482797.7A CN107274401B (en) | 2017-06-22 | 2017-06-22 | High-resolution SAR image ship detection method based on visual attention mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710482797.7A CN107274401B (en) | 2017-06-22 | 2017-06-22 | High-resolution SAR image ship detection method based on visual attention mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107274401A true CN107274401A (en) | 2017-10-20 |
| CN107274401B CN107274401B (en) | 2020-09-04 |
Family
ID=60069020
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710482797.7A Active CN107274401B (en) | 2017-06-22 | 2017-06-22 | High-resolution SAR image ship detection method based on visual attention mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107274401B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446584A (en) * | 2018-01-30 | 2018-08-24 | 中国航天电子技术研究院 | A kind of unmanned plane scouting video image target automatic testing method |
| CN108599884A (en) * | 2018-03-19 | 2018-09-28 | 重庆大学 | A kind of Noise enhancement minimizes the signal detecting method of error probability |
| CN108694714A (en) * | 2018-05-14 | 2018-10-23 | 浙江大学 | Ship seakeeping system in a kind of adaptive colony intelligence optimization SAR Radar Seas |
| CN109165660A (en) * | 2018-06-20 | 2019-01-08 | 扬州大学 | A kind of obvious object detection method based on convolutional neural networks |
| CN110084210A (en) * | 2019-04-30 | 2019-08-02 | 电子科技大学 | The multiple dimensioned Ship Detection of SAR image based on attention pyramid network |
| CN110853050A (en) * | 2019-10-21 | 2020-02-28 | 中国电子科技集团公司第二十九研究所 | SAR image river segmentation method, device and medium |
| CN111008585A (en) * | 2019-11-29 | 2020-04-14 | 西安电子科技大学 | Ship target detection method based on self-adaptive layered high-resolution SAR image |
| CN111046871A (en) * | 2019-12-11 | 2020-04-21 | 厦门大学 | Region-of-interest extraction method and system |
| CN111666854A (en) * | 2020-05-29 | 2020-09-15 | 武汉大学 | High-resolution SAR image vehicle target detection method fusing statistical significance |
| CN111767856A (en) * | 2020-06-29 | 2020-10-13 | 哈工程先进技术研究院(招远)有限公司 | Infrared small target detection algorithm based on gray value statistical distribution model |
| CN112836571A (en) * | 2020-12-18 | 2021-05-25 | 华中科技大学 | Ship target detection and identification method, system and terminal in remote sensing SAR image |
| CN112862748A (en) * | 2020-12-25 | 2021-05-28 | 重庆大学 | Multidimensional domain feature combined SAR (synthetic aperture radar) ship intelligent detection method |
| CN113111758A (en) * | 2021-04-06 | 2021-07-13 | 中山大学 | SAR image ship target identification method based on pulse neural network |
| CN113203991A (en) * | 2021-04-29 | 2021-08-03 | 电子科技大学 | Anti-deception jamming method of multi-base SAR (synthetic aperture radar) in multi-jammer environment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090067730A1 (en) * | 2004-10-22 | 2009-03-12 | Henry Schneiderman | Object Recognizer and Detector for Two-Dimensional Images Using Bayesian Network Based Classifier |
| CN102122352A (en) * | 2011-03-01 | 2011-07-13 | 西安电子科技大学 | Characteristic value distribution statistical property-based polarized SAR image classification method |
| US20150235073A1 (en) * | 2014-01-28 | 2015-08-20 | The Trustees Of The Stevens Institute Of Technology | Flexible part-based representation for real-world face recognition apparatus and methods |
| CN105354541A (en) * | 2015-10-23 | 2016-02-24 | 西安电子科技大学 | SAR (Synthetic Aperture Radar) image target detection method based on visual attention model and constant false alarm rate |
| CN105427314A (en) * | 2015-11-23 | 2016-03-23 | 西安电子科技大学 | Bayesian saliency based SAR image target detection method |
-
2017
- 2017-06-22 CN CN201710482797.7A patent/CN107274401B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090067730A1 (en) * | 2004-10-22 | 2009-03-12 | Henry Schneiderman | Object Recognizer and Detector for Two-Dimensional Images Using Bayesian Network Based Classifier |
| CN102122352A (en) * | 2011-03-01 | 2011-07-13 | 西安电子科技大学 | Characteristic value distribution statistical property-based polarized SAR image classification method |
| US20150235073A1 (en) * | 2014-01-28 | 2015-08-20 | The Trustees Of The Stevens Institute Of Technology | Flexible part-based representation for real-world face recognition apparatus and methods |
| CN105354541A (en) * | 2015-10-23 | 2016-02-24 | 西安电子科技大学 | SAR (Synthetic Aperture Radar) image target detection method based on visual attention model and constant false alarm rate |
| CN105427314A (en) * | 2015-11-23 | 2016-03-23 | 西安电子科技大学 | Bayesian saliency based SAR image target detection method |
Non-Patent Citations (1)
| Title |
|---|
| 刘赵强: "高分辨SAR图像机动目标检测与识别技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑（月刊 ）》 * |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446584A (en) * | 2018-01-30 | 2018-08-24 | 中国航天电子技术研究院 | A kind of unmanned plane scouting video image target automatic testing method |
| CN108599884B (en) * | 2018-03-19 | 2020-11-03 | 重庆大学 | Signal detection method for noise enhancement minimizing error probability |
| CN108599884A (en) * | 2018-03-19 | 2018-09-28 | 重庆大学 | A kind of Noise enhancement minimizes the signal detecting method of error probability |
| CN108694714A (en) * | 2018-05-14 | 2018-10-23 | 浙江大学 | Ship seakeeping system in a kind of adaptive colony intelligence optimization SAR Radar Seas |
| CN109165660B (en) * | 2018-06-20 | 2021-11-09 | 扬州大学 | Significant object detection method based on convolutional neural network |
| CN109165660A (en) * | 2018-06-20 | 2019-01-08 | 扬州大学 | A kind of obvious object detection method based on convolutional neural networks |
| CN110084210A (en) * | 2019-04-30 | 2019-08-02 | 电子科技大学 | The multiple dimensioned Ship Detection of SAR image based on attention pyramid network |
| CN110084210B (en) * | 2019-04-30 | 2022-03-29 | 电子科技大学 | SAR image multi-scale ship detection method based on attention pyramid network |
| CN110853050A (en) * | 2019-10-21 | 2020-02-28 | 中国电子科技集团公司第二十九研究所 | SAR image river segmentation method, device and medium |
| CN110853050B (en) * | 2019-10-21 | 2023-05-26 | 中国电子科技集团公司第二十九研究所 | SAR image river segmentation method, device and medium |
| CN111008585A (en) * | 2019-11-29 | 2020-04-14 | 西安电子科技大学 | Ship target detection method based on self-adaptive layered high-resolution SAR image |
| CN111008585B (en) * | 2019-11-29 | 2023-04-07 | 西安电子科技大学 | Ship target detection method based on self-adaptive layered high-resolution SAR image |
| CN111046871A (en) * | 2019-12-11 | 2020-04-21 | 厦门大学 | Region-of-interest extraction method and system |
| CN111046871B (en) * | 2019-12-11 | 2023-07-11 | 厦门大学 | Region of interest extraction method and system |
| CN111666854A (en) * | 2020-05-29 | 2020-09-15 | 武汉大学 | High-resolution SAR image vehicle target detection method fusing statistical significance |
| CN111767856A (en) * | 2020-06-29 | 2020-10-13 | 哈工程先进技术研究院(招远)有限公司 | Infrared small target detection algorithm based on gray value statistical distribution model |
| CN111767856B (en) * | 2020-06-29 | 2023-11-10 | 烟台哈尔滨工程大学研究院 | Infrared small target detection algorithm based on gray value statistical distribution model |
| CN112836571A (en) * | 2020-12-18 | 2021-05-25 | 华中科技大学 | Ship target detection and identification method, system and terminal in remote sensing SAR image |
| CN112862748A (en) * | 2020-12-25 | 2021-05-28 | 重庆大学 | Multidimensional domain feature combined SAR (synthetic aperture radar) ship intelligent detection method |
| CN112862748B (en) * | 2020-12-25 | 2023-05-30 | 重庆大学 | Multi-dimensional domain feature combined SAR ship intelligent detection method |
| CN113111758A (en) * | 2021-04-06 | 2021-07-13 | 中山大学 | SAR image ship target identification method based on pulse neural network |
| CN113111758B (en) * | 2021-04-06 | 2024-01-12 | 中山大学 | SAR image ship target recognition method based on impulse neural network |
| CN113203991A (en) * | 2021-04-29 | 2021-08-03 | 电子科技大学 | Anti-deception jamming method of multi-base SAR (synthetic aperture radar) in multi-jammer environment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107274401B (en) | 2020-09-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107274401A (en) | A kind of High Resolution SAR Images Ship Detection of view-based access control model attention mechanism | |
| CN107229918B (en) | SAR image target detection method based on full convolution neural network | |
| CN108280460B (en) | SAR vehicle target identification method based on improved convolutional neural network | |
| CN107808138B (en) | Communication signal identification method based on FasterR-CNN | |
| CN105320764B (en) | A kind of 3D model retrieval method and its retrieval device based on the slow feature of increment | |
| CN108171119B (en) | SAR image change detection method based on residual error network | |
| CN104657717A (en) | Pedestrian detection method based on layered kernel sparse representation | |
| CN106557740A (en) | The recognition methods of oil depot target in a kind of remote sensing images | |
| CN102999908A (en) | Synthetic aperture radar (SAR) airport segmentation method based on improved visual attention model | |
| Ouyang et al. | The research of the strawberry disease identification based on image processing and pattern recognition | |
| Tahseen et al. | Binarization Methods in Multimedia Systems when Recognizing License Plates of Cars | |
| CN106447686A (en) | Method for detecting image edges based on fast finite shearlet transformation | |
| CN112784757B (en) | Marine SAR ship target significance detection and identification method | |
| CN106548195A (en) | A kind of object detection method based on modified model HOG ULBP feature operators | |
| Ju et al. | A novel fully convolutional network based on marker-controlled watershed segmentation algorithm for industrial soot robot target segmentation | |
| CN109191436A (en) | The low-dose CT Lung neoplasm detection algorithm of view-based access control model conspicuousness spectrum residual error method | |
| CN112036419B (en) | SAR image component interpretation method based on VGG-Attention model | |
| CN110223295B (en) | Significance prediction method and device based on deep neural network color perception | |
| Gautam et al. | A GUI for automatic extraction of signature from image document | |
| Yuankui et al. | Automatic target recognition of ISAR images based on Hausdorff distance | |
| CN106548180A (en) | A kind of method for obtaining the Feature Descriptor for obscuring constant image | |
| Chaiyakhan et al. | Traffic sign classification using support vector machine and image segmentation | |
| CN111898531A (en) | Satellite communication signal identification method and device and electronic equipment | |
| Xu et al. | A new method based on two-stage detection mechanism for detecting ships in high-resolution sar images | |
| Zhao et al. | Feature extraction and classification of ocean oil spill based on SAR image |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200807 Address after: 264001 Research and Academic Department, 188 Erma Road, Zhifu District, Yantai City, Shandong Province Applicant after: NAVAL AERONAUTICAL University Address before: 264001 Yantai City, Zhifu Province, No. two road, No. 188, Department of research, Applicant before: NAVAL AERONAUTICAL AND ASTRONAUTICAL University PLA |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |