CN108805028A - SAR image ground target detection based on electromagnetism strong scattering point and localization method - Google Patents
SAR image ground target detection based on electromagnetism strong scattering point and localization method Download PDFInfo
- Publication number
- CN108805028A CN108805028A CN201810423035.4A CN201810423035A CN108805028A CN 108805028 A CN108805028 A CN 108805028A CN 201810423035 A CN201810423035 A CN 201810423035A CN 108805028 A CN108805028 A CN 108805028A
- Authority
- CN
- China
- Prior art keywords
- target
- scattering point
- strong scattering
- feature
- point
- 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.)
- Pending
Links
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/2163—Partitioning the feature space
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The SAR image ground target based on electromagnetism strong scattering point that the invention discloses a kind of detects and localization method, including：The polarimetric SAR image of acquisition is pre-processed；The potential region of target is oriented with Threshold segmentation and interesting target region ROI is obtained by Morphological scale-space and cluster；Extract strong scattering point feature；It is established to obtain the threedimensional model of target according to target type library；Establish strong scattering point model of the target under different orientations；Echo data emulation is carried out according to the strong scattering point model of foundation, obtains one-dimensional range profile of the target under different orientations；Strong scattering point feature is extracted by one-dimensional range profile, feature is selected and is normalized；SVM classifier is trained to normalized strong scattering point feature and carries out characteristic matching with the strong scattering point feature extracted from ROI, finally obtains recognition result.The present invention can be used for ground target detection and Classification and Identification, effective position target location and can carry out subsequent target following.
Description
Technical field
The present invention relates to SAR image processing technology fields, and in particular to a kind of SAR image based on electromagnetism strong scattering point
Area Objects detect and localization method.
Background technology
The essence of SAR image target detection is that the feature difference showed according to target and clutter is realized to target
Detection, central step are the extraction of feature.It expands in terms of SAR image feature extraction and is largely ground both at home and abroad at present
Study carefully, is broadly divided into Extraction of Geometrical Features, gray distribution features extraction and scattering signatures extraction.
Geometric properties include deflection, main shaft, length and width, profile of target etc., can reflect target under different orientations
Size category information, have stability and invariance.This kind of feature extraction is simple, but is a kind of directviewing description to target,
The essential attribute of target cannot be accurately reflected, and secondary lobe present in SAR image can also seriously affect the extraction of target length and width
As a result.Gray distribution features include the texture of target, intensity etc., are the intuitive reflections of atural object character, but due to atural object characteristic
Complexity, simple gray distribution features are far from being enough for SAR image target identification, and such methods are to image
Quality requirement is higher and is affected by azimuth, the problems such as be easy to causeing a large amount of false-alarms and centre of location point position offset.It dissipates
It includes sharp peaks characteristic, scattering center feature etc. to penetrate feature, can reflect the fine physical arrangement of target, be obtained in field of target recognition
To being more and more widely used, but existing such methods are based primarily upon image segmentation, when the certain partial dispersion intensity of target
When weaker, it is difficult to correctly detect by image segmentation, this has resulted in the loss of target important feature, the meeting in Feature Points Matching
There are problems that model mismatch, causes the precision for finally detecting positioning relatively low, be unsatisfactory for pinpoint application demand.
Existing more commonly used target identification method is mainly based upon the characteristic matching of template.This method principle is simple,
As long as establishing a complete template library conveniently matches corresponding target, and can be obtained in target classification
To relatively good recognition effect.But complete template library is established, it needs to obtain radar incidence wave under different pitch angles to each
Class target sample different orientations correspondence image, in addition, different geomorphic features and background complexity can be to template library
Foundation bring great difficulty.Therefore, this method is not suitable for the application demand of Practical Project real-time high-efficiency.
Invention content
The purpose of the present invention is to provide a kind of, and the SAR image ground target based on electromagnetism strong scattering point detects and positioning side
Method proposes to use and establishes mesh on the scattering properties basis for fully considering the ground such as SAR image-forming principles and tank armored target
Mark strong scattering point model realizes ground target detection and positioning in the method for extracting strong scattering point feature, and it is big not only to eliminate acquisition
The practical operation for measuring sample SAR image is difficult, it is thus also avoided that the error or Character losing occurred when extraction sample scatter feature
The problems such as, to the target detection being more suitable in Practical Project under complex environment.
Realize that the technical solution of the object of the invention is：A kind of SAR image ground target detection based on electromagnetism strong scattering point
With localization method, comprise the following processes：
Step 1, the polarimetric SAR image of acquisition is pre-processed；
Step 2, the potential region of target is oriented with Threshold sementation, marginal point is filtered out by Morphological scale-space, and
The detection of interesting target region ROI is realized by cluster；
Step 3, the strong scattering point feature of extraction interesting target region ROI；
Step 4, it is established to obtain the threedimensional model of target by target type library, and different direction is established according to scattering point theory
Strong scattering point model under angle；
Step 5, echo data emulation is carried out according to the strong scattering point model of foundation, obtains target under different orientations
High resolution range profile；
Step 6, the strong scattering point feature for extracting object module, selects feature and is normalized, to normalized strong
Scattering point feature is trained, and obtains required SVM classifier；
Step 7, characteristic matching is carried out to the strong scattering point feature extracted from ROI, obtains final recognition result.
Compared with prior art, remarkable advantage of the invention is：(1) present invention is considering the three of ground armored target
On the basis of tieing up structure feature and scattering properties, the strong scattering point feature of target is obtained according to scattering point theoretical modeling, is not only exempted from
It has gone the practical operation of acquisition great amount of samples SAR image difficult, has also reduced the time overhead and target of great amount of samples test
Divide model mismatch problem caused by error, improves the recognition efficiency of target；(2) strong scattering point spy is obtained by modeling and simulating
Sign data have it is at low cost, be easily achieved, operate many advantages such as controllable, behind effective position target location and progress
Continuous target following；(3) the High Range Resolution feature of millimere-wave band is utilized, target Equivalent can be formed for multiple strong scattering points
Complex target, using target strong scattering point feature carry out object detection and recognition, existing target identification technology can be solved
The relatively low problem of middle model mismatch, identification probability improves positioning accuracy to reject false-alarm.
Description of the drawings
Fig. 1 is that the present invention is based on the detections of the SAR image ground target of strong scattering point and localization method flow chart.
Fig. 2 (a) is target original SAR image, and Fig. 2 (b) is to obtain after Target Segmentation, Morphological scale-space and clustering
Target binary map, Fig. 2 (c) are the result figure for finally detecting target centroid point.
Fig. 3 is to be distributed vertical view to the strong scattering point that the tank target threedimensional model of foundation obtains according to scattering point theory.
Fig. 4 is the RCS simulation result schematic diagrams of the tank target model to foundation in the present invention.
Specific implementation mode
In conjunction with Fig. 1, a kind of detection of SAR image ground target and localization method based on electromagnetism strong scattering point, this method packet
Include following steps：
(1) polarimetric SAR image of acquisition is pre-processed, removes the influence of coherent spot, clutter reduction background, enhancing and
Highlight target area；
(2) the potential region of target is oriented with Threshold sementation, filters out marginal point by Morphological scale-space, and pass through
Cluster realizes the detection of interesting target region ROI；
(3) the strong scattering point feature of extraction interesting target region ROI；
(4) it is established to obtain the threedimensional model of target by target type library, and different orientations is established according to scattering point theory
Under strong scattering point model；
(5) echo data emulation is carried out according to the strong scattering point model of foundation, obtains one of target under different orientations
Tie up high resolution range profile；
(6) the strong scattering point feature for extracting object module, selects feature and is normalized, to normalized strong scattering
Point feature is trained, and obtains required SVM classifier；
(7) characteristic matching is carried out to the strong scattering point feature extracted from ROI, obtains final recognition result.
Further, it is established to obtain the threedimensional model of target by target type library in the step (4), and according to scattering point
Theory establishes the strong scattering point model under different orientations, and detailed process is：
(4a) establishes threedimensional model in target type library according to the appearance and size of target；
(4b) sets scattering point according to the target 3 d structure model of foundation at the turning, edge and junction of object, from
And it models and obtains scattering center distribution map of the target under different orientations；
(4c) is scattered center modeling according to scattering center distribution map under obtained different orientations to target, forms mesh
Target strong scattering point model；
Further, echo data emulation is carried out according to the strong scattering point model of foundation in the step (5), obtains target
One-dimensional range profile under different orientations；Detailed process is：
(5a) in millimere-wave band, the total electromagnetic scattering of complex target is regarded as the strong scattering point tribute on certain local locations
The coherent superposition offered.It is assumed that object module has k strong scattering point, then back scattering of the object module on given frequency point
Response meets：
Wherein, y is back scattering response of the object module on given frequency point, AnFor the amplitude of strong scattering point, rnFor
The distance between strong scattering point and radar receiver, f are frequency, and c is the light velocity, and μ is noise.
(5b) projects the target strong scattering point echo-signal of acquisition in distance upwards, obtains target in different direction
High resolution range profile under angle；
Further, the strong scattering point feature that object module is extracted in the step (6) select simultaneously normalizing to feature
Change, normalized strong scattering point feature is trained, required SVM classifier is obtained；Detailed process is：
(6a) is special according to high resolution range profile extraction strong scattering point of the sample object model under different orientations
Obtain feature vector D=(d1,d2,d3), wherein three Feature Descriptors respectively represent target strong scattering point number d1, strong to dissipate
Hit heart Distribution Entropy d2And target radar scattering cross-section accumulates d3；
(6b) is using Principal Component Analysis PCA to feature vector D=(d1,d2,d3) dimensionality reduction is carried out, to remove redundancy,
Then it is normalized；
(6c) is trained normalized sampling feature vectors, obtains required SVM classifier.
With reference to embodiment and attached drawing, the present invention will be described in detail.
Embodiment
Realizing the technical thought of the present invention is：Millimere-wave band, the single scattering of target sufficiently high in resolution ratio
Response is gathered in certain isolated regions, therefore the high-frequency electromagnetic scattering response of high-resolution target can use the electricity of scattering center
The sum of magnetic scattering response indicates that the sum of response of that is, certain strong scattering points indicates.
As shown in Figure 1, a kind of detection of SAR image ground target and localization method based on electromagnetism strong scattering point, including with
Under several steps：
(1) Frost filtering is carried out to the polarimetric SAR image of acquisition, the output result for obtaining Frost filtering isWhereinFor the image after Frost filtering, the SAR image that I is, E is index impulse
The factor, InFor each pixel of filter window, EnFor exponential weighting factor.
(2) the potential region of target is oriented with maximum entropy threshold cutting techniques, shape is passed through to the bianry image after segmentation
State processing filters out marginal point and by holes filling with completion connected domain, and interesting target region is realized finally by cluster
The detection of ROI.Connected component labeling, definition are carried out to the bianry image after Target Segmentation using 8 neighborhood growth methodsH=
1,2 ... ..., L are the centre of form coordinate of each simply connected domain, construct the range search matrix of L × L
Wherein according to priori given threshold Y, work as dmnWhen < Y, n is classified as class m, target area is searched for pixel-by-pixel, will be searched
Belong to of a sort connected domain after rope to merge, and update connected domain, and then obtain object detection results, output is target
Barycenter point coordinates.Attached drawing 2 (a)~2 (c) is respectively the original SAR image of target, and Target Segmentation, Morphological scale-space and cluster are drawn
The schematic diagram of the target binary map and final detection result that are obtained after point.In an embodiment of the present invention, the picture point of final output
Coordinate is [68.8893,75.1268].
(3) the strong scattering point feature T (t of extraction interesting target region ROI1,t2,t3).Definition simultaneously orientation and away from
It is that the peak point of local maximum corresponds to the strong scattering central point in target on descriscent, then can be calculated by following formula
The strong scattering point feature of target：
Wherein Ii,jIt is a for the strong scattering central point of targeti,jFor current pixel value, P (ai,j) it is ai,jLocal neighborhood,
ap,qFor ai,jArbitrary pixel value in local neighborhood.
Three Feature Descriptors of above-mentioned strong scattering point feature respectively represent target strong scattering point number t1, strong scattering point is strong
Spend t2And strong scattering point Distribution Entropy t3。
(4) it establishes to obtain the threedimensional model of ground target according to the geometry feature of all kinds of targets in target type library,
The scattering point theoretical definition tank target of the turning of target object, edge and junction is generally all distributed according to scattering center
Strong scattering center appears in gun turret, gun barrel, edge and corner, thus establishes the strong scattering central distribution under different orientations
Figure.Center modeling is scattered to target further according to scattering center distribution map under obtained different orientations, forms the strong of target
Scatter times.Fig. 3 is to be distributed and overlook to the strong scattering point that the tank target threedimensional model of foundation obtains according to scattering point theory
Figure.
(5) according in millimere-wave band, the total electromagnetic scattering of complex target may be considered strong scattered on certain local locations
This theoretical premise of the coherent superposition of exit point contribution carries out echo data emulation using the target strong scattering point model of foundation, from
And obtain high resolution range profile of the target under different orientations.It is assumed that object module has k strong scattering point, then the mesh
Back scattering response of the model on given frequency point is marked to meet：
Wherein, y is back scattering response of the object module on given frequency point, AnIndicate the amplitude of strong scattering point, rn
Indicate that the distance between strong scattering point and radar receiver, f are frequency, c indicates that the light velocity, μ represent noise.
The target strong scattering point echo-signal of acquisition is projected into distance upwards again, target can be obtained in different orientations
Under high resolution range profile.
(6) high resolution range profile according to sample object model under different orientations extracts strong scattering point feature
Obtain feature vector D=(d1,d2,d3), wherein three Feature Descriptors respectively represent target strong scattering point number d1, strong scattering
Central distribution entropy d2And target radar scattering cross-section accumulates d3.Using PCA algorithms to obtained feature vector D=(d1,d2,d3) into
Row dimensionality reduction simultaneously normalizes, and to reduce redundancy, improves the training effectiveness of grader, while improving the recognition performance of target.Fig. 4 is this
To the RCS simulation result schematic diagrams of the tank target model of foundation in invention.
(7) to the strong scattering point feature extracted from ROI and the strong scattering point feature established by object module
Library carries out characteristic matching, obtains final recognition result.
In conclusion the present invention obtains the strong scattering point information of target by oneself modeling and simulating, for SAR image appearance
Mark, which is detected and is accurately positioned, provides new thinking.Certain local locations of ground armored target with extremely strong due to backward dissipating
Penetrate coefficient, it is some bright spots to be reflected in SAR image, i.e. strong scattering point, and these points will not because of image-forming condition variation and
Variation, it is more stable, accurate compared to traditional geometry and gray scale textural characteristics.In addition, the strong scattering point in the present invention is special
Sign is obtained by oneself simulation modeling, economical and effective, it is easy to accomplish, it is suitable for the application of SAR image target identification, and available
In subsequent positioning and tracking.
Claims (4)
1. a kind of detection of SAR image ground target and localization method based on electromagnetism strong scattering point, which is characterized in that including as follows
Process：
Step 1, the polarimetric SAR image of acquisition is pre-processed；
Step 2, the potential region of target is oriented with Threshold sementation, filters out marginal point by Morphological scale-space, and pass through
Cluster realizes the detection of interesting target region ROI；
Step 3, the strong scattering point feature of extraction interesting target region ROI；
Step 4, it is established to obtain the threedimensional model of target by target type library, and is established under different orientations according to scattering point theory
Strong scattering point model；
Step 5, echo data emulation is carried out according to the strong scattering point model of foundation, it is one-dimensional under different orientations obtains target
High resolution range profile；
Step 6, the strong scattering point feature for extracting object module, selects feature and is normalized, to normalized strong scattering
Point feature is trained, and obtains required SVM classifier；
Step 7, characteristic matching is carried out to the strong scattering point feature extracted from ROI, obtains final recognition result.
2. the detection of SAR image ground target and localization method according to claim 1 based on electromagnetism strong scattering point, special
Sign is that the detailed process of step 4 is：
(4a) establishes threedimensional model in target type library according to the appearance and size of target；
(4b) sets scattering point according to the target 3 d structure model of foundation at the turning, edge and junction of object, to build
Mould obtains scattering center distribution map of the target under different orientations；
(4c) is scattered center modeling according to scattering center distribution map under obtained different orientations to target, forms target
Strong scattering point model.
3. the detection of SAR image ground target and localization method according to claim 1 based on electromagnetism strong scattering point, special
Sign is that the detailed process of step 5 is：
(5a) assumes that object module has k strong scattering point, then back scattering response of the object module on given frequency point is full
Foot：
Wherein, y is back scattering response of the object module on given frequency point, AnFor the amplitude of strong scattering point, rnFor strong scattering
The distance between point and radar receiver, f is frequency, and c is the light velocity, and μ is noise.
(5b) projects the target strong scattering point echo-signal of acquisition in distance upwards, obtains target under different orientations
High resolution range profile.
4. the detection of SAR image ground target and localization method according to claim 1 based on electromagnetism strong scattering point, special
Sign is that the detailed process of step 6 is：
(6a) is obtained according to high resolution range profile extraction strong scattering point feature of the sample object model under different orientations
To feature vector D=(d1,d2,d3), wherein three Feature Descriptors respectively represent target strong scattering point number d1, in strong scattering
Heart Distribution Entropy d2And target radar scattering cross-section accumulates d3；
(6b) is using Principal Component Analysis to feature vector D=(d1,d2,d3) dimensionality reduction is carried out, then it is normalized；
(6c) is trained normalized sampling feature vectors, obtains required SVM classifier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810423035.4A CN108805028A (en) | 2018-05-05 | 2018-05-05 | SAR image ground target detection based on electromagnetism strong scattering point and localization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810423035.4A CN108805028A (en) | 2018-05-05 | 2018-05-05 | SAR image ground target detection based on electromagnetism strong scattering point and localization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108805028A true CN108805028A (en) | 2018-11-13 |
Family
ID=64093639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810423035.4A Pending CN108805028A (en) | 2018-05-05 | 2018-05-05 | SAR image ground target detection based on electromagnetism strong scattering point and localization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108805028A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109444840A (en) * | 2018-12-04 | 2019-03-08 | 南京航空航天大学 | A kind of radar clutter suppression method based on machine learning |
| CN109872350A (en) * | 2019-02-18 | 2019-06-11 | 重庆市勘测院 | A kind of new point cloud autoegistration method |
| CN109946698A (en) * | 2019-04-15 | 2019-06-28 | 北京市遥感信息研究所 | A kind of polarimetric synthetic aperture radar typical target feature base construction method and device |
| CN110223311A (en) * | 2019-05-24 | 2019-09-10 | 杭州世平信息科技有限公司 | Polarimetric radar edge Detection Method of Remote Sensing Images based on power drive adaptive windows |
| CN110441750A (en) * | 2019-07-19 | 2019-11-12 | 北京遥感设备研究所 | A kind of radar target strong scattering point extracting method |
| CN113421293A (en) * | 2021-06-30 | 2021-09-21 | 上海申瑞继保电气有限公司 | Transformer substation equipment image centroid calculation method |
| CN113820712A (en) * | 2021-09-07 | 2021-12-21 | 中山大学 | Ship target positioning method and system based on strong scattering points |
| CN115061136A (en) * | 2022-06-08 | 2022-09-16 | 江苏省水利科学研究院 | Method and system for detecting river and lake shoreline change point based on SAR image |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101281249A (en) * | 2008-05-20 | 2008-10-08 | 北京航空航天大学 | Method for modeling target dispersion characteristic of high resolution synthetic aperture radar |
| CN104730507A (en) * | 2015-04-01 | 2015-06-24 | 苏州闻捷传感技术有限公司 | Vehicle-mounted road barrier alarm method based on premodulation AIC radar range imaging |
| CN105930848A (en) * | 2016-04-08 | 2016-09-07 | 西安电子科技大学 | SAR-SIFT feature-based SAR image target recognition method |
| CN104459663B (en) * | 2014-11-27 | 2017-02-22 | 中国船舶重工集团公司第七二四研究所 | Naval vessel and cargo vessel classification method based on high-resolution one-dimensional range profile |
| CN106772307A (en) * | 2017-03-02 | 2017-05-31 | 电子科技大学 | A kind of true and false bullet target identification method based on many radar informations |
-
2018
- 2018-05-05 CN CN201810423035.4A patent/CN108805028A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101281249A (en) * | 2008-05-20 | 2008-10-08 | 北京航空航天大学 | Method for modeling target dispersion characteristic of high resolution synthetic aperture radar |
| CN104459663B (en) * | 2014-11-27 | 2017-02-22 | 中国船舶重工集团公司第七二四研究所 | Naval vessel and cargo vessel classification method based on high-resolution one-dimensional range profile |
| CN104730507A (en) * | 2015-04-01 | 2015-06-24 | 苏州闻捷传感技术有限公司 | Vehicle-mounted road barrier alarm method based on premodulation AIC radar range imaging |
| CN105930848A (en) * | 2016-04-08 | 2016-09-07 | 西安电子科技大学 | SAR-SIFT feature-based SAR image target recognition method |
| CN106772307A (en) * | 2017-03-02 | 2017-05-31 | 电子科技大学 | A kind of true and false bullet target identification method based on many radar informations |
Non-Patent Citations (1)
| Title |
|---|
| 田亚平: "雷达地面目标分类技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109444840A (en) * | 2018-12-04 | 2019-03-08 | 南京航空航天大学 | A kind of radar clutter suppression method based on machine learning |
| CN109444840B (en) * | 2018-12-04 | 2020-12-25 | 南京航空航天大学 | Radar clutter suppression method based on machine learning |
| CN109872350A (en) * | 2019-02-18 | 2019-06-11 | 重庆市勘测院 | A kind of new point cloud autoegistration method |
| CN109946698A (en) * | 2019-04-15 | 2019-06-28 | 北京市遥感信息研究所 | A kind of polarimetric synthetic aperture radar typical target feature base construction method and device |
| CN110223311A (en) * | 2019-05-24 | 2019-09-10 | 杭州世平信息科技有限公司 | Polarimetric radar edge Detection Method of Remote Sensing Images based on power drive adaptive windows |
| CN110441750A (en) * | 2019-07-19 | 2019-11-12 | 北京遥感设备研究所 | A kind of radar target strong scattering point extracting method |
| CN113421293A (en) * | 2021-06-30 | 2021-09-21 | 上海申瑞继保电气有限公司 | Transformer substation equipment image centroid calculation method |
| CN113421293B (en) * | 2021-06-30 | 2023-12-29 | 上海申瑞继保电气有限公司 | Substation equipment image centroid calculation method |
| CN113820712A (en) * | 2021-09-07 | 2021-12-21 | 中山大学 | Ship target positioning method and system based on strong scattering points |
| CN113820712B (en) * | 2021-09-07 | 2023-07-28 | 中山大学 | Ship target positioning method and system based on strong scattering points |
| CN115061136A (en) * | 2022-06-08 | 2022-09-16 | 江苏省水利科学研究院 | Method and system for detecting river and lake shoreline change point based on SAR image |
| CN115061136B (en) * | 2022-06-08 | 2024-01-09 | 江苏省水利科学研究院 | SAR image-based river and lake shoreline change point detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108805028A (en) | SAR image ground target detection based on electromagnetism strong scattering point and localization method | |
| CN108510467B (en) | SAR image target identification method based on depth deformable convolution neural network | |
| Zhang et al. | Airport detection and aircraft recognition based on two-layer saliency model in high spatial resolution remote-sensing images | |
| Liu et al. | A novel region-merging approach for coastline extraction from Sentinel-1A IW mode SAR imagery | |
| CN105427314B (en) | SAR image object detection method based on Bayes's conspicuousness | |
| Xiang et al. | Adaptive superpixel generation for polarimetric SAR images with local iterative clustering and SIRV model | |
| CN102214298B (en) | Method for detecting and identifying airport target by using remote sensing image based on selective visual attention mechanism | |
| Wang et al. | Frequency subband processing and feature analysis of forward-looking ground-penetrating radar signals for land-mine detection | |
| CN108256436A (en) | A kind of radar HRRP target identification methods based on joint classification | |
| CN104966081B (en) | Spine image-recognizing method | |
| CN111175718A (en) | Time-frequency domain combined ground radar automatic target identification method and system | |
| Xiang et al. | Adaptive superpixel generation for SAR images with linear feature clustering and edge constraint | |
| CN108333634A (en) | A kind of Ground Penetrating Radar buried target localization method based on frequency spectrum remnants conspicuousness detection methods | |
| Manno-Kovacs et al. | Image based robust target classification for passive ISAR | |
| CN109117739A (en) | One kind identifying projection properties extracting method based on neighborhood sample orientation | |
| CN111458688B (en) | Three-dimensional convolution network-based radar high-resolution range profile target recognition method | |
| Dong et al. | Keypoint-based local descriptors for target recognition in SAR images: A comparative analysis | |
| Tu et al. | Airport detection in SAR images via salient line segment detector and edge-oriented region growing | |
| Doo et al. | Reliable target feature extraction and classification using potential target information | |
| Zhang et al. | Vector phase symmetry for stable hyperbola detection in ground-penetrating radar images | |
| Manno-Kovacs et al. | Automatic target classification in passive ISAR range-crossrange images | |
| Kovacs et al. | Orientation based building outline extraction in aerial images | |
| CN116311067A (en) | Target comprehensive identification method, device and equipment based on high-dimensional characteristic map | |
| Wu et al. | Moving target shadow detection method based on improved ViBe in VideoSAR images | |
| CN115797773A (en) | Ultra-wideband radar target identification method based on edge detection and region segmentation |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |