CN110008785A - A kind of target identification method and device - Google Patents

A kind of target identification method and device Download PDF

Info

Publication number
CN110008785A
CN110008785A CN201810007812.7A CN201810007812A CN110008785A CN 110008785 A CN110008785 A CN 110008785A CN 201810007812 A CN201810007812 A CN 201810007812A CN 110008785 A CN110008785 A CN 110008785A
Authority
CN
China
Prior art keywords
block
target
pixel
cluster centre
coding vector
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201810007812.7A
Other languages
Chinese (zh)
Other versions
CN110008785B (en
Inventor
曾亮
杨健
包君梁
陈杭
金侃
林慧平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN201810007812.7A priority Critical patent/CN110008785B/en
Publication of CN110008785A publication Critical patent/CN110008785A/en
Application granted granted Critical
Publication of CN110008785B publication Critical patent/CN110008785B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • 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
    • G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
    • 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
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection

Abstract

This application discloses a kind of target identification method and devices, comprising: carries out super-pixel segmentation to polarimetric synthetic aperture radar SAR image；Target belonging to each block after determining super-pixel segmentation, to realize target identification.The application is by carrying out super-pixel segmentation to polarimetric SAR image, then pass through target described in each block after determining super-pixel segmentation, to finally determine target, it realizes and identifies target from polarimetric SAR image, and this method is calculated by unit of block, rather than calculated by unit of pixel, while considering the correlation between block, to improve arithmetic speed, the accuracy of target identification is improved.

Description

A kind of target identification method and device
Technical field
This application involves image processing techniques, espespecially a kind of target identification method and device.
Background technique
Remote sensing refers to all contactless long-range detections, is one of the important means of perception ambient enviroment.Wherein polarize Synthetic aperture radar (SAR, Synthetic Aperture Radar) remote sensing is one of most important active remote sensing mode.SAR It is a kind of active earth observation systems, mountable on the flying platforms such as aircraft, satellite, spaceship, round-the-clock, whole day It waits and implements to observe over the ground and there is certain ground penetrating ability.After the polarimetric SAR image obtained by SAR, need to polarization Target in SAR image is identified, and not yet provides effective solution scheme in the related technology.
Summary of the invention
This application provides a kind of target identification method and devices, and target can be identified from polarimetric SAR image.
This application provides a kind of target identification methods, comprising:
Super-pixel segmentation is carried out to polarimetric synthetic aperture radar SAR image；
Target belonging to each block after determining super-pixel segmentation, to realize target identification.
Optionally, described to include: to polarimetric SAR image progress super-pixel segmentation
K cluster centre is initialized in the polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
Each corresponding cluster centre, calculates each pixel in the predeterminable area centered on the cluster centre With the distance metric of corresponding cluster centre, using with the smallest pixel of the distance metric of the cluster centre as new cluster Center；Wherein, the predeterminable area is the square region that side length is 2S,N is the pixel of the polarimetric SAR image The number of point；
Continue to calculate each pixel in the predeterminable area centered on the new cluster centre with it is corresponding new Cluster centre distance metric, until the cluster centre is no longer changed；
It is the same block by the pixel merger with the cluster centre with connectivity.
Optionally, each pixel in the predeterminable area of the calculating centered on the cluster centre with it is corresponding The distance metric of cluster centre includes:
According to formulaCalculate the ith pixel point in the predeterminable area and j-th The distance metric of cluster centre；
Wherein,Alternatively, For the vector with pixel One-to-one vector, Tr are the mark of matrix.
Optionally, target belonging to each block after the determining super-pixel segmentation includes:
Each described block is encoded to obtain the coding vector of block, and each described target is encoded Obtain the coding vector of target；
When each block after super-pixel segmentation in the polarization SAR belongs to corresponding target, according to the block Coding vector and the target coding vector computation energy function value；
Change target belonging at least one described block, again according to the coding vector of the block and the target Coding vector calculates the energy function value；
Determine that corresponding target is target belonging to corresponding block when the energy function value minimum.
Optionally, described to include: according to the coding vector of block and the coding vector computation energy function value of target
According to formulaCalculate the energy function value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on it is pre- If the coding vector of j-th of block in region, yiFor the coding vector of target belonging to i-th of block, h is block xiItself Probability distribution, joint probability distribution of the β between block, η be block xiBelong to target yiConditional probability distribution.
This body provides a kind of Target Identification Unit, comprising:
Divide module, for carrying out super-pixel segmentation to polarimetric synthetic aperture radar SAR image；
Determining module, for determining target belonging to each block after super-pixel segmentation, to realize target identification.
Optionally, the segmentation module is specifically used for:
K cluster centre is initialized in the polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
Each corresponding cluster centre, calculates each pixel in the predeterminable area centered on the cluster centre With the distance metric of corresponding cluster centre, using with the smallest pixel of the distance metric of the cluster centre as new cluster Center；Wherein, the predeterminable area is the square region that side length is 2S,N is the pixel of the polarimetric SAR image The number of point；
Continue to calculate each pixel in the predeterminable area centered on the new cluster centre with it is corresponding new Cluster centre distance metric, until the cluster centre is no longer changed；
It is the same block by the pixel merger with the cluster centre with connectivity.
Optionally, during the segmentation module is specifically used for realizing that the calculating with the cluster centre is in the following ways Each pixel in the predeterminable area of the heart and the distance metric of corresponding cluster centre:
According to formulaCalculate the ith pixel point in the predeterminable area and j-th The distance metric of cluster centre；
Wherein,Alternatively, For the vector one with pixel One corresponding vector, Tr are the mark of matrix.
Optionally, the determining module is specifically used for:
Each described block is encoded to obtain the coding vector of block, and each described target is encoded Obtain the coding vector of target；
When each block after super-pixel segmentation in the polarization SAR belongs to corresponding target, according to the block Coding vector and the target coding vector computation energy function value；
Change target belonging at least one described block, again according to the coding vector of the block and the target Coding vector calculates the energy function value；
Determine that corresponding target is target belonging to corresponding block when the energy function value minimum.
Optionally, the determining module is specifically used for realizing the coding vector and mesh according to block in the following ways Target coding vector computation energy function value:
According to formulaCalculate the energy function value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on it is pre- If the coding vector of j-th of block in region, yiFor the coding vector of target belonging to i-th of block, h is block xiItself Probability distribution, joint probability distribution of the β between block, η be block xiBelong to target yiConditional probability distribution.
It is described computer-readable to deposit this application provides a kind of terminal, including processor and computer readable storage medium It is stored with instruction in storage media, when described instruction is executed by the processor, realizes any of the above-described kind of target identification method.
This application provides a kind of computer readable storage mediums, are stored thereon with computer program, the computer journey The step of any of the above-described kind of target identification method is realized when sequence is executed by processor.
Compared with the relevant technologies, the application includes: to carry out super-pixel segmentation to polarimetric synthetic aperture radar SAR image；Really Target belonging to each block after determining super-pixel segmentation, to realize target identification.The application by polarimetric SAR image into Row super-pixel segmentation, then by target described in each block after determining super-pixel segmentation, thus finally determine target, It realizes and identifies target from polarimetric SAR image, and this method is calculated by unit of block, rather than be single with pixel Member is calculated, to improve arithmetic speed, improves the accuracy of target identification.
Other features and advantage will illustrate in the following description, also, partly become from specification It obtains it is clear that being understood and implementing the application.The purpose of the application and other advantages can be by specifications, right Specifically noted structure is achieved and obtained in claim and attached drawing.
Detailed description of the invention
Attached drawing is used to provide to further understand technical scheme, and constitutes part of specification, with this The embodiment of application is used to explain the technical solution of the application together, does not constitute the limitation to technical scheme.
Fig. 1 is the flow chart of target identification method of the embodiment of the present invention；
Fig. 2 is the schematic diagram of polarimetric SAR image of the embodiment of the present invention；
Fig. 3 is the schematic diagram of the polarimetric SAR image after super-pixel segmentation of the embodiment of the present invention；
Fig. 4 is the schematic diagram after target identification of the embodiment of the present invention；
Fig. 5 is the structure composition schematic diagram of Target Identification Unit of the embodiment of the present invention；
Fig. 6 is the structure composition schematic diagram of the terminal of that embodiment of the invention.
Specific embodiment
For the purposes, technical schemes and advantages of the application are more clearly understood, below in conjunction with attached drawing to the application Embodiment be described in detail.It should be noted that in the absence of conflict, in the embodiment and embodiment in the application Feature can mutual any combination.
Step shown in the flowchart of the accompanying drawings can be in a computer system such as a set of computer executable instructions It executes.Also, although logical order is shown in flow charts, and it in some cases, can be to be different from herein suitable Sequence executes shown or described step.
Referring to Fig. 1, the embodiment of the present invention proposes a kind of target identification method, comprising:
Step 100 carries out super-pixel segmentation to polarimetric SAR image.
Optionally, before step 100 further include: pre-processed to polarimetric SAR image, then after step 100 pair pretreatment Polarimetric SAR image carry out super-pixel segmentation.
Wherein, pretreated process includes: and carries out drop spot, polarization Contrast enhanced, polarization SAR to polarimetric SAR image to decompose Deng.
In this step, super-pixel segmentation can be carried out to polarimetric SAR image using following methods.
K cluster centre is initialized in polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
For each cluster centre, calculate each pixel in the predeterminable area centered on cluster centre with it is right The distance metric for the cluster centre answered, using with the smallest pixel of the distance metric of cluster centre as new cluster centre；After Each pixel in the continuous predeterminable area calculated centered on new cluster centre and corresponding new cluster centre away from From measurement, until cluster centre is no longer changed；Wherein, predeterminable area is the square region that side length is 2S,N For the number of the pixel of polarimetric SAR image；
It is the same block by the pixel merger with cluster centre with connectivity.
Wherein, k cluster centre is initialized in polarimetric SAR image to refer to and arbitrarily choose k picture in polarimetric SAR image Vegetarian refreshments is as k cluster centre.
Wherein, according to formulaCalculate the ith pixel point in predeterminable area and j-th The distance metric of cluster centre；
Wherein,Alternatively, For the vector one with pixel One corresponding vector, Tr are the mark of matrix.
Fig. 2 is the schematic diagram of polarimetric SAR image of the embodiment of the present invention.As shown in Fig. 2, marine, there are two targets.Fig. 3 is this The schematic diagram of polarimetric SAR image after inventive embodiments super-pixel segmentation.As shown in figure 3, the polarization SAR figure after super-pixel segmentation As including multiple blocks, two marine targets are also divided into multiple blocks, and subsequent to do is to determine each block institute The target of category can come out the target identification in polarimetric SAR image.
Step 101 determines target belonging to each block after super-pixel segmentation, to realize target identification.
It, can be based on the side Hidden Markov random field (HRMF, Hidden Markov Random Fields) in this step Method determines target belonging to each block after super-pixel segmentation, it may be assumed that
Each block is encoded to obtain the coding vector of block, and each target is encoded to obtain target Coding vector；When each block after super-pixel segmentation in polarization SAR belongs to corresponding target, according to the coding of block The coding vector computation energy function value of vector sum target；Change target belonging at least one block, again according to block The coding vector computation energy function value of coding vector and target determines that corresponding target is corresponding area when energy function value minimum Target belonging to block.
Wherein, it can specifically be encoded using one-hot coding (One-Hot Encoding) mode, and according to initial Probability distribution is encoded.For example, it is assumed that the quantity of target is M, the then coding vector of block and target in polarimetric SAR image Coding vector is the vector of M dimension；And only one equal element in the coding vector of block and the coding vector of target It is 1, other elements 0.When being encoded according to initial probability distribution, which indicates mesh belonging to each block Target probability, the coding vector of the target when coding vector of block is the maximum probability of target belonging to the block.
Wherein, according to formulaCalculate the energy function Value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on it is pre- If the coding vector of j-th of block in region, yiFor the coding vector of target belonging to i-th of block, h is block xiItself Probability distribution, joint probability distribution of the β between block, η be block xiBelong to target yiConditional probability distribution.
Wherein, as shown in figure 3, with xiCentered on predeterminable area in block can refer to and xiAdjacent block, or With xiDistance be less than preset threshold block, etc..Specific predeterminable area take it is much can preset, the embodiment of the present invention This is not construed as limiting.
Above-mentioned formulaMiddle Section 2 considers block Between correlation, further improve the accuracy of target identification.
For example, it is assumed that super-pixel segmentation after-polarization SAR image includes two blocks, respectively x1And x2, it is understood that there may be mesh Mark is also two, respectively y1And y2, then the coding vector of block is (0,1) or (1,0), the coding vector of target be (0, Or (1,0) 1).
Assuming that x1=(0,1), x2=(1,0), y1=(0,1), y2=(1,0), then,
As block x1Affiliated target is y1, block x2Affiliated target is y1When, E1(X, Y)=h (| x1|+|x2|)-β (x1x2+x2x1)-η(x1y1+x2y1)；
As block x1Affiliated target is y1, block x2Affiliated target is y2When, E2(X, Y)=h (| x1|+|x2|)-β (x1x2+x2x1)-η(x1y1+x2y2)；
As block x1Affiliated target is y2, block x2Affiliated target is y1When, E3(X, Y)=h (| x1|+|x2|)-β (x1x2+x2x1)-η(x1y2+x2y1)；
As block x1Affiliated target is y2, block x2Affiliated target is y2When, E4(X, Y)=h (| x1|+|x2|)-β (x1x2+x2x1)-η(x1y2+x2y2)；
Four E values that above-mentioned four kinds of situations obtain are calculated, the situation of E value minimum is taken.For example, determining area when E1 minimum Block x1Belong to target y1, block x2Belong to target y2.Other situations and so on, which is not described herein again.
Fig. 4 is the schematic diagram after target identification of the embodiment of the present invention.As shown in figure 4, including three mesh in polarimetric SAR image Mark, respectively sea-surface target and two naval targets.The application is then led to by carrying out super-pixel segmentation to polarimetric SAR image Target described in each block after determining super-pixel segmentation is crossed, to finally determine target, is realized from polarimetric SAR image In identify target, and this method is calculated by unit of block, rather than is calculated by unit of pixel, to improve Arithmetic speed, improves the accuracy of target identification.
Referring to Fig. 5, the embodiment of the present invention proposes a kind of Target Identification Unit, comprising:
Divide module, for carrying out super-pixel segmentation to polarimetric synthetic aperture radar SAR image；
Determining module, for determining target belonging to each block after super-pixel segmentation, to realize target identification.
Optionally, the segmentation module is specifically used for:
K cluster centre is initialized in the polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
Each corresponding cluster centre, calculates each pixel in the predeterminable area centered on the cluster centre With the distance metric of corresponding cluster centre, using with the smallest pixel of the distance metric of the cluster centre as new cluster Center；Wherein, the predeterminable area is the square region that side length is 2S,N is the pixel of the polarimetric SAR image The number of point；
Continue to calculate each pixel in the predeterminable area centered on the new cluster centre with it is corresponding new Cluster centre distance metric, until the cluster centre is no longer changed；
It is the same block by the pixel merger with the cluster centre with connectivity.
Optionally, during the segmentation module is specifically used for realizing that the calculating with the cluster centre is in the following ways Each pixel in the predeterminable area of the heart and the distance metric of corresponding cluster centre:
According to formulaIt calculates the ith pixel point in the predeterminable area and j-th is gathered The distance metric at class center；
Wherein,Alternatively, For the vector one with pixel One corresponding vector, Tr are the mark of matrix.
Optionally, the determining module is specifically used for:
Each described block is encoded to obtain the coding vector of block, and each described target is encoded Obtain the coding vector of target；
When each block after super-pixel segmentation in the polarization SAR belongs to corresponding target, according to the block Coding vector and the target coding vector computation energy function value；
Change target belonging at least one described block, again according to the coding vector of the block and the target Coding vector calculates the energy function value；
Determine that corresponding target is target belonging to corresponding block when the energy function value minimum.
Optionally, the determining module is specifically used for realizing the coding vector and mesh according to block in the following ways Target coding vector computation energy function value:
According to formulaCalculate the energy function value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on it is pre- If the coding vector of j-th of block in region, yiFor the coding vector of target belonging to i-th of block, h is block xiItself Probability distribution, joint probability distribution of the β between block, η be block xiBelong to target yiConditional probability distribution.
Referring to Fig. 6, the embodiment of the present invention proposes a kind of terminal, including processor and computer readable storage medium, institute It states and is stored with instruction in computer readable storage medium, when described instruction is executed by the processor, realize any of the above-described kind Target identification method.
The embodiment of the present invention proposes a kind of computer readable storage medium, is stored thereon with computer program, the meter The step of calculation machine program realizes any of the above-described kind of target identification method when being executed by processor.
Although embodiment disclosed by the application is as above, the content only for ease of understanding the application and use Embodiment is not limited to the application.Technical staff in any the application fields, is taken off not departing from the application Under the premise of the spirit and scope of dew, any modification and variation, but the application can be carried out in the form and details of implementation Scope of patent protection, still should be subject to the scope of the claims as defined in the appended claims.

Claims (12)

1. a kind of target identification method characterized by comprising
Super-pixel segmentation is carried out to polarimetric synthetic aperture radar SAR image；
Target belonging to each block after determining super-pixel segmentation, to realize target identification.
2. target identification side's method according to claim 1, which is characterized in that described to carry out super picture to polarimetric SAR image Element is divided
K cluster centre is initialized in the polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
Each corresponding cluster centre, calculate each pixel in the predeterminable area centered on the cluster centre with it is right The distance metric for the cluster centre answered, using with the smallest pixel of the distance metric of the cluster centre as in new cluster The heart；Wherein, the predeterminable area is the square region that side length is 2S,N is the pixel of the polarimetric SAR image Number；
Continue to calculate each pixel in the predeterminable area centered on the new cluster centre to gather with corresponding new The distance metric at class center, until the cluster centre is no longer changed；
It is the same block by the pixel merger with the cluster centre with connectivity.
3. target identification method according to claim 2, which is characterized in that the calculating is centered on the cluster centre Predeterminable area in each pixel include: with the distance metric of corresponding cluster centre
According to formulaIt calculates in the ith pixel point and j-th of cluster in the predeterminable area The distance metric of the heart；
Wherein,Alternatively,One by one for the vector with pixel Corresponding vector, Tr are the mark of matrix.
4. target identification method according to claim 1, which is characterized in that each after the determining super-pixel segmentation Target belonging to block includes:
Each described block is encoded to obtain the coding vector of block, and each described target is encoded to obtain The coding vector of target；
When each block after super-pixel segmentation in the polarization SAR belongs to corresponding target, according to the volume of the block The coding vector computation energy function value of code vector and the target；
Change target belonging at least one described block, again according to the coding of the coding vector of the block and the target Vector calculates the energy function value；
Determine that corresponding target is target belonging to corresponding block when the energy function value minimum.
5. target identification method according to claim 4, which is characterized in that the coding vector and target according to block Coding vector computation energy function value include:
According to formulaCalculate the energy function value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on preset areas The coding vector of j-th of block in domain, yiFor the coding vector of target belonging to i-th of block, h is block xiItself general Rate distribution, joint probability distribution of the β between block, η are block xiBelong to target yiConditional probability distribution.
6. a kind of Target Identification Unit characterized by comprising
Divide module, for carrying out super-pixel segmentation to polarimetric synthetic aperture radar SAR image；
Determining module, for determining target belonging to each block after super-pixel segmentation, to realize target identification.
7. Target Identification Unit according to claim 6, which is characterized in that the segmentation module is specifically used for:
K cluster centre is initialized in the polarimetric SAR image；Wherein, k is the integer more than or equal to 1；
Each corresponding cluster centre, calculate each pixel in the predeterminable area centered on the cluster centre with it is right The distance metric for the cluster centre answered, using with the smallest pixel of the distance metric of the cluster centre as in new cluster The heart；Wherein, the predeterminable area is the square region that side length is 2S,N is the pixel of the polarimetric SAR image Number；
Continue to calculate each pixel in the predeterminable area centered on the new cluster centre to gather with corresponding new The distance metric at class center, until the cluster centre is no longer changed；
It is the same block by the pixel merger with the cluster centre with connectivity.
8. Target Identification Unit according to claim 7, which is characterized in that the segmentation module is specifically used for using following Mode realize each pixel in the predeterminable area of the calculating centered on the cluster centre in corresponding cluster The distance metric of the heart:
According to formulaIt calculates in the ith pixel point and j-th of cluster in the predeterminable area The distance metric of the heart；
Wherein,Alternatively,One by one for the vector with pixel Corresponding vector, Tr are the mark of matrix.
9. Target Identification Unit according to claim 6, which is characterized in that the determining module is specifically used for:
Each described block is encoded to obtain the coding vector of block, and each described target is encoded to obtain The coding vector of target；
When each block after super-pixel segmentation in the polarization SAR belongs to corresponding target, according to the volume of the block The coding vector computation energy function value of code vector and the target；
Change target belonging at least one described block, again according to the coding of the coding vector of the block and the target Vector calculates the energy function value；
Determine that corresponding target is target belonging to corresponding block when the energy function value minimum.
10. Target Identification Unit according to claim 9, which is characterized in that the determining module be specifically used for use with Under type realizes the coding vector computation energy function value of the coding vector according to block and target:
According to formulaCalculate the energy function value；
Wherein, E (X, Y) is the energy function value, xiFor the coding vector of i-th of block, xjFor with xiCentered on preset areas The coding vector of j-th of block in domain, yiFor the coding vector of target belonging to i-th of block, h is block xiItself general Rate distribution, joint probability distribution of the β between block, η are block xiBelong to target yiConditional probability distribution.
11. a kind of terminal, including processor and computer readable storage medium, it is stored in the computer readable storage medium Instruction, which is characterized in that when described instruction is executed by the processor, realize mesh as claimed in any one of claims 1 to 5 Mark recognition methods.
12. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program The step of target identification method as claimed in any one of claims 1 to 5 is realized when being executed by processor.
CN201810007812.7A 2018-01-04 2018-01-04 Target identification method and device Active CN110008785B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810007812.7A CN110008785B (en) 2018-01-04 2018-01-04 Target identification method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810007812.7A CN110008785B (en) 2018-01-04 2018-01-04 Target identification method and device

Publications (2)

Publication Number Publication Date
CN110008785A true CN110008785A (en) 2019-07-12
CN110008785B CN110008785B (en) 2022-09-02

Family

ID=67164308

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810007812.7A Active CN110008785B (en) 2018-01-04 2018-01-04 Target identification method and device

Country Status (1)

Country Link
CN (1) CN110008785B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104376330A (en) * 2014-11-19 2015-02-25 西安电子科技大学 Polarization SAR image ship target detection method based on superpixel scattering mechanism
CN105354599A (en) * 2015-11-25 2016-02-24 安徽创世科技有限公司 Color recognition method based on improved SLIC super-pixel segmentation algorithm
CN105844292A (en) * 2016-03-18 2016-08-10 南京邮电大学 Image scene labeling method based on conditional random field and secondary dictionary study
CN106056159A (en) * 2016-06-03 2016-10-26 西安电子科技大学 Image fine classification method based on Fisher Vector
CN106651884A (en) * 2016-12-30 2017-05-10 西安电子科技大学 Sketch structure-based mean field variational Bayes synthetic aperture radar (SAR) image segmentation method
CN106778821A (en) * 2016-11-25 2017-05-31 西安电子科技大学 Classification of Polarimetric SAR Image method based on SLIC and improved CNN
CN106815842A (en) * 2017-01-23 2017-06-09 河海大学 A kind of improved image significance detection method based on super-pixel
CN107113442A (en) * 2014-10-31 2017-08-29 三星电子株式会社 Method and apparatus for being encoded/decoded to motion vector

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107113442A (en) * 2014-10-31 2017-08-29 三星电子株式会社 Method and apparatus for being encoded/decoded to motion vector
CN104376330A (en) * 2014-11-19 2015-02-25 西安电子科技大学 Polarization SAR image ship target detection method based on superpixel scattering mechanism
CN105354599A (en) * 2015-11-25 2016-02-24 安徽创世科技有限公司 Color recognition method based on improved SLIC super-pixel segmentation algorithm
CN105844292A (en) * 2016-03-18 2016-08-10 南京邮电大学 Image scene labeling method based on conditional random field and secondary dictionary study
CN106056159A (en) * 2016-06-03 2016-10-26 西安电子科技大学 Image fine classification method based on Fisher Vector
CN106778821A (en) * 2016-11-25 2017-05-31 西安电子科技大学 Classification of Polarimetric SAR Image method based on SLIC and improved CNN
CN106651884A (en) * 2016-12-30 2017-05-10 西安电子科技大学 Sketch structure-based mean field variational Bayes synthetic aperture radar (SAR) image segmentation method
CN106815842A (en) * 2017-01-23 2017-06-09 河海大学 A kind of improved image significance detection method based on super-pixel

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
ALEXANDER SCHICK ET AL.: ""Improving foreground segmentations with probabilistic superpixel Markov random fields"", 《2012 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS》 *
於敏 等: ""基于相似性和统计性的超像素的图像分割"", 《计算机工程与应用》 *
柚子皮: ""马尔可夫随机场 MRF"", 《CSDN》 *

Also Published As

Publication number Publication date
CN110008785B (en) 2022-09-02

Similar Documents

Publication Publication Date Title
CN103413151B (en) Hyperspectral image classification method based on figure canonical low-rank representation Dimensionality Reduction
CN108647591A (en) Activity recognition method and system in a kind of video of view-based access control model-semantic feature
CN108492258A (en) A kind of radar image denoising method based on generation confrontation network
CN105608421B (en) A kind of recognition methods of human action and device
Chen et al. Adversarial example in remote sensing image recognition
CN108460341A (en) Remote sensing image object detection method based on integrated depth convolutional network
CN103345630B (en) A kind of traffic signs localization method based on spherical panoramic video
CN109740604A (en) A kind of method and apparatus of running region detection
CN103984966A (en) SAR image target recognition method based on sparse representation
CN109886218A (en) SAR image Ship Target Detection method based on super-pixel statistics diversity
CN103679216B (en) Method for detecting infrared small/weak targets based on space-time sparse representation
CN103425995B (en) Hyperspectral image classification method based on region similarity low rank expression dimension reduction
CN105184830A (en) Symmetry image symmetric axis detection positioning method
Li et al. Gaofen-3 sea ice detection based on deep learning
CN107742133A (en) A kind of sorting technique for Polarimetric SAR Image
CN117237740B (en) SAR image classification method based on CNN and Transformer
CN108021886A (en) A kind of unmanned plane repeats texture image part remarkable characteristic matching process
Xu et al. On-board ship detection in SAR images based on L-YOLO
Jia et al. Depth measurement based on a convolutional neural network and structured light
CN106649607A (en) Remote sensing image online target recognition and monitoring method and system based on vector mask
CN110008785A (en) A kind of target identification method and device
Yang et al. Real-time human activity classification from radar with CNN-LSTM network
CN108875798A (en) A kind of super-pixel grade feature extracting method based on spatial pyramid pond
Qayyum et al. Fusion of CNN and sparse representation for threat estimation near power lines and poles infrastructure using aerial stereo imagery
CN103337075B (en) A kind of image significance computing method based on isophote

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