CN112305509A - Radar track classification and identification method based on HOG _ SVM - Google Patents

Radar track classification and identification method based on HOG _ SVM Download PDF

Info

Publication number
CN112305509A
CN112305509A CN202011206346.9A CN202011206346A CN112305509A CN 112305509 A CN112305509 A CN 112305509A CN 202011206346 A CN202011206346 A CN 202011206346A CN 112305509 A CN112305509 A CN 112305509A
Authority
CN
China
Prior art keywords
track
svm
library
hog
radar
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
Application number
CN202011206346.9A
Other languages
Chinese (zh)
Inventor
陈龙坤
张涛
李洪涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Science and Technology
Original Assignee
Nanjing University of Science and Technology
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 Nanjing University of Science and Technology filed Critical Nanjing University of Science and Technology
Priority to CN202011206346.9A priority Critical patent/CN112305509A/en
Publication of CN112305509A publication Critical patent/CN112305509A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/36—Means for anti-jamming, e.g. ECCM, i.e. electronic counter-counter measures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/40—Means for monitoring or calibrating

Landscapes

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

Abstract

The invention discloses a radar track classification and identification method based on HOG _ SVM. The method is mainly suitable for classification and identification after the phased array radar track is formed. The invention comprises the following steps: (1) collecting motion tracks of a large number of different targets; (2) preprocessing the data acquired in the step 1 to form a track training library and a track testing library with labels; (3) extracting HOG characteristic values of data in a track training library with labels; (4) calculating HOG characteristic values and labels of data in a training library by using an SVM model, and constructing a sample comparison library; (5) and (3) verifying by using the track test library data with the tags, namely extracting the characteristic value of the track of the test library, predicting a track tag result by using an SVM (support vector machine) regression model, and outputting a track category tag. The invention belongs to the field of radar data processing, and aims at completing a classification recognition task aiming at a formed radar track.

Description

Radar track classification and identification method based on HOG _ SVM
Technical Field
The invention belongs to the technical field of radar data processing, and particularly relates to a radar track classification and identification method based on HOG _ SVM.
Background
The radar technology is developed rapidly, and can detect not only aerial targets such as unmanned aerial vehicles, civil aviation, birds and animals, but also ground targets such as pedestrians, vehicles and the like. The more powerful the radar, the more and more complex the types of targets that can be detected. Particularly, the ground and low-altitude targets with complex electromagnetic environments have a large amount of electromagnetic interference, and a stable false target track can be formed. The radar track formation mostly adopts a moving target display algorithm, and the influence caused by ground electromagnetic interference cannot be eliminated. The method for classifying and identifying the radar track based on the HOG _ SVM is characterized by extracting the characteristic value of the relevant parameter of the radar track, classifying the characteristic value by using an SVM technology to form a parameter base, and predicting unknown data by using an SVM regression model. The radar system using the method can give the category attribute of the target track in real time, and has high real-time performance and high efficiency.
Disclosure of Invention
The invention relates to a method for phased array radar track classification and identification based on HOG _ SVM, which comprises the following steps:
step 1: a large number of radar target tracks are collected, the tracks are in essence time-aligned point tracks, and each point track comprises the speed, energy, distance, azimuth, pitch, and signal-to-noise ratio of the target.
Step 2: preprocessing the data acquired in the step 1, including track smoothing: removing extreme data and taking fixed-length data track points; and (3) track labeling: and manually marking track labels according to the actual track forming reason, and arranging each track data point according to the time sequence.
And step 3: and (3) expanding and arranging the radar target parameters of each track in the flight path according to time to form a two-dimensional data table, and extracting the HOG characteristic values of the flight paths in the training library and the testing library. The characteristic value is calculated by the following mathematical model. Wherein G isx(x，y)、Gy(x, y) represents the horizontal gradient and the vertical gradient at the data point (x, y), respectively.
G (x, y) is the gradient magnitude, and α (x, y) represents the gradient direction.
Gx(x，y)＝(x+1，y)-(x-1，y)
Gy(x，y)＝(x，y+1)-(x，y-1)
Figure BSA0000223552560000021
Figure BSA0000223552560000022
And 4, step 4: and (4) calculating the HOG characteristic value and the label in the training library by using an SVM model, and constructing a sample comparison library. The SVM model idea is that a partition hyperplane is found based on a characteristic value sample space of a track training library, characteristic values of different labels are separated and classified into a characteristic value database. The SVM mathematical model is as follows. Where ω is the normal vector that determines the direction of the hyperplane and b is the displacement term that determines the distance between the hyperplane and the origin.
ωTxi+b＝yi
And 5: and (4) verifying by using the track of the test library, calculating a characteristic value of the test track by using an SVM regression algorithm model, predicting a track label and outputting a predicted label result. SVM regression model mathematical formula, wherein
Figure BSA0000223552560000023
Figure BSA0000223552560000024
Drawings
FIG. 1 is a flow chart of a HOG _ SVM radar based track classification recognition process.
Fig. 2 is a schematic diagram of the direction of HOG feature values.
Fig. 3 is an SVM support vector and interval.
Detailed Description
The invention relates to a method for phased array radar track classification and identification based on HOG _ SVM, the general flow chart is shown in figure 1, and the method comprises the following steps:
step 1: a large number of radar target tracks are collected, the tracks are in essence time-aligned point tracks, and each point track comprises the speed, energy, distance, azimuth, pitch, and signal-to-noise ratio of the target.
Step 2: preprocessing the data acquired in the step 1, including track smoothing: removing extreme data, taking fixed length data track points, as indicated by distance threshold (R)min，Rmax) The SNR threshold is expressed as (G)min，Gmax) The energy threshold is expressed as (E)min，Emax) The velocity threshold is expressed as (v)min，vmax) The wild value eliminating mathematical model is as follows:
Figure BSA0000223552560000025
1: represented as valid data points; 0 represents an invalid data point;
and (3) track labeling: and manually marking track labels according to the actual track forming reason, and arranging each track data point according to the time sequence.
And step 3: and (3) expanding and arranging the radar target parameters of each track in the flight path according to time to form a two-dimensional data table, and extracting the HOG characteristic values of the flight paths in the training library and the testing library. The characteristic value is calculated by the following mathematical model. Wherein G isx(x，y)、Gy(x, y) represents the horizontal gradient and the vertical gradient at (x, y) in the two-bit data table, respectively. G (x, y) is the gradient magnitude, and α (x, y) represents the gradient direction. Wherein the gradient direction is shown in FIG. 2, and the gradient direction range is 2 π
Figure BSA0000223552560000031
As a unit of block stepping, the gradient direction 6 may be divided into equal parts. Accumulating the gradient magnitude values of each gradient direction can convert the direction histogram into a single-dimensional vector feature value (x)1，x2，x3，x4，x5，x6)；
Gx(x，y)＝(x+1，y)-(x-1，y)
Gy(x，y)＝(x，y+1)-(x，y-1)
Figure BSA0000223552560000032
Figure BSA0000223552560000033
And 4, step 4: and (4) calculating the HOG characteristic value and the label in the training library by using an SVM model, and constructing a sample comparison library. The SVM model idea is that a partition hyperplane is found based on a characteristic value sample space of a track training library, characteristic values of different labels are separated and classified into a characteristic value database. The SVM mathematical model is as follows. Where ω is the normal vector determining the direction of the hyperplane, and b is the displacement term determining the distance between the hyperplane and the origin, and the absolute value of the displacement is taken to be 1, as shown in fig. 3.
ωTxi+b＝yi
And 5: and (4) verifying by using the track of the test library, calculating a characteristic value of the test track by using an SVM regression algorithm model, predicting a track label and outputting a predicted label result. SVM regression model algorithm formula, wherein
Figure BSA0000223552560000034
Figure BSA0000223552560000035
E∈The distance value of the characteristic value from the hyperplane after SVM regression, if and only if E∈And yiWhen exactly the same, the loss is zero. In practice, however, the losses cannot be zero and E can be tolerated∈And yiWith a deviation of e between.
Figure BSA0000223552560000036
When the predicted value is close to the true value;
Figure BSA0000223552560000037
the prediction is not close to the true value.

Claims (6)

1. A radar track classification and identification method based on HOG _ SVM is characterized in that: the method comprises the following steps:
step 1: and collecting a large number of radar tracks of different targets as sample data.
Step 2: and (3) preprocessing the data acquired in the step (1) and constructing a training library and a testing library with labels.
And step 3: and extracting the HOG characteristic values of the tracks in the training library with the labels and the test library.
And 4, step 4: and (4) calculating the HOG characteristic value and the label in the training library by using an SVM model, and constructing a sample comparison library.
And 5: and (4) verifying by using the track of the test library, calculating a characteristic value of the test track by using an SVM regression algorithm model, predicting a track label and outputting a predicted label result.
2. The HOG _ SVM-based radar track classification and identification method according to claim 1, wherein the radar track collected in step 1, the point track parameters in the track include speed, energy, distance, azimuth, pitch, and signal-to-noise ratio of the target.
3. The HOG _ SVM-based radar track classification and identification method according to claim 1, wherein step 2 is used for preprocessing the radar track, including track smoothing: removing extreme data and taking fixed-length data track points; and (3) track labeling: and manually marking track labels according to the actual track forming reason, and arranging each track data point according to the time sequence.
4. The HOG _ SVM based radar track classification and identification method according to claim 1, wherein the radar target parameters of each track in the step 3 are spread and arranged according to time to form a two-dimensional data table, and HOG characteristic values of the tracks in the training library and the testing library are extracted.
5. The HOG _ SVM based radar track classification and identification method according to claim 1, wherein the step 4 is to train HOG characteristic values and labels by using an SVM classification model to construct a label-based sample comparison library.
6. The HOG _ SVM based radar track classification and identification method according to claim 1, wherein the test track is input into SVM regression algorithm in step 5, and the track label is predicted and output.
CN202011206346.9A 2020-11-02 2020-11-02 Radar track classification and identification method based on HOG _ SVM Pending CN112305509A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011206346.9A CN112305509A (en) 2020-11-02 2020-11-02 Radar track classification and identification method based on HOG _ SVM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011206346.9A CN112305509A (en) 2020-11-02 2020-11-02 Radar track classification and identification method based on HOG _ SVM

Publications (1)

Publication Number Publication Date
CN112305509A true CN112305509A (en) 2021-02-02

Family

ID=74333876

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011206346.9A Pending CN112305509A (en) 2020-11-02 2020-11-02 Radar track classification and identification method based on HOG _ SVM

Country Status (1)

Country Link
CN (1) CN112305509A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112782666A (en) * 2021-03-22 2021-05-11 哈尔滨工程大学 Rapid detection method for marine radar target

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101837782A (en) * 2009-01-26 2010-09-22 通用汽车环球科技运作公司 Be used to collide the multiple goal Fusion Module of preparation system
CN103559791A (en) * 2013-10-31 2014-02-05 北京联合大学 Vehicle detection method fusing radar and CCD camera signals
JP2014126526A (en) * 2012-12-27 2014-07-07 Japan Radio Co Ltd Tracking support device
JP2015099571A (en) * 2013-11-20 2015-05-28 オリンパス株式会社 Object position identifying system and object position identifying method
JP2016151416A (en) * 2015-02-16 2016-08-22 三菱電機株式会社 Targets identification device
CN106469315A (en) * 2016-09-05 2017-03-01 南京理工大学 Based on the multi-mode complex probe target identification method improving One Class SVM algorithm
CN109117850A (en) * 2018-06-28 2019-01-01 上海交通大学 The recognition methods of corresponding infrared target image is carried out using visible light target image
KR20190049221A (en) * 2017-11-01 2019-05-09 주식회사 이노베스트 an Autonomous Vehicle of pedestrians facial features
CN109753874A (en) * 2018-11-28 2019-05-14 南京航空航天大学 A kind of low slow small classification of radar targets method based on machine learning
CN110018453A (en) * 2019-03-28 2019-07-16 西南电子技术研究所（中国电子科技集团公司第十研究所） Intelligent type recognition methods based on aircraft track feature
CN110221266A (en) * 2019-06-11 2019-09-10 哈尔滨工程大学 A kind of marine radar target rapid detection method based on support vector machines
CN110940959A (en) * 2019-12-13 2020-03-31 中国电子科技集团公司第五十四研究所 Man-vehicle classification and identification method for low-resolution radar ground target
CN110969174A (en) * 2018-09-29 2020-04-07 深圳市布谷鸟科技有限公司 Target identification method, device and system based on laser radar
CN111062322A (en) * 2019-12-17 2020-04-24 西安电子科技大学 Phased array radar behavior recognition method based on Support Vector Machine (SVM)
CN111428748A (en) * 2020-02-20 2020-07-17 重庆大学 Infrared image insulator recognition and detection method based on HOG characteristics and SVM

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101837782A (en) * 2009-01-26 2010-09-22 通用汽车环球科技运作公司 Be used to collide the multiple goal Fusion Module of preparation system
JP2014126526A (en) * 2012-12-27 2014-07-07 Japan Radio Co Ltd Tracking support device
CN103559791A (en) * 2013-10-31 2014-02-05 北京联合大学 Vehicle detection method fusing radar and CCD camera signals
JP2015099571A (en) * 2013-11-20 2015-05-28 オリンパス株式会社 Object position identifying system and object position identifying method
JP2016151416A (en) * 2015-02-16 2016-08-22 三菱電機株式会社 Targets identification device
CN106469315A (en) * 2016-09-05 2017-03-01 南京理工大学 Based on the multi-mode complex probe target identification method improving One Class SVM algorithm
KR20190049221A (en) * 2017-11-01 2019-05-09 주식회사 이노베스트 an Autonomous Vehicle of pedestrians facial features
CN109117850A (en) * 2018-06-28 2019-01-01 上海交通大学 The recognition methods of corresponding infrared target image is carried out using visible light target image
CN110969174A (en) * 2018-09-29 2020-04-07 深圳市布谷鸟科技有限公司 Target identification method, device and system based on laser radar
CN109753874A (en) * 2018-11-28 2019-05-14 南京航空航天大学 A kind of low slow small classification of radar targets method based on machine learning
CN110018453A (en) * 2019-03-28 2019-07-16 西南电子技术研究所（中国电子科技集团公司第十研究所） Intelligent type recognition methods based on aircraft track feature
CN110221266A (en) * 2019-06-11 2019-09-10 哈尔滨工程大学 A kind of marine radar target rapid detection method based on support vector machines
CN110940959A (en) * 2019-12-13 2020-03-31 中国电子科技集团公司第五十四研究所 Man-vehicle classification and identification method for low-resolution radar ground target
CN111062322A (en) * 2019-12-17 2020-04-24 西安电子科技大学 Phased array radar behavior recognition method based on Support Vector Machine (SVM)
CN111428748A (en) * 2020-02-20 2020-07-17 重庆大学 Infrared image insulator recognition and detection method based on HOG characteristics and SVM

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112782666A (en) * 2021-03-22 2021-05-11 哈尔滨工程大学 Rapid detection method for marine radar target

Similar Documents

Publication Publication Date Title
Miao et al. Insulator detection in aerial images for transmission line inspection using single shot multibox detector
Chen et al. Automatic defect detection of fasteners on the catenary support device using deep convolutional neural network
Yu et al. A new spatial-oriented object detection framework for remote sensing images
Dong et al. Deep metric learning-based for multi-target few-shot pavement distress classification
Kong et al. Hypernet: Towards accurate region proposal generation and joint object detection
CN113034548B (en) Multi-target tracking method and system suitable for embedded terminal
CN111161315B (en) Multi-target tracking method and system based on graph neural network
CN105528794A (en) Moving object detection method based on Gaussian mixture model and superpixel segmentation
Lan et al. Defect detection from UAV images based on region-based CNNs
CN106570490B (en) A kind of pedestrian's method for real time tracking based on quick clustering
CN110018453A (en) Intelligent type recognition methods based on aircraft track feature
CN107798691B (en) A kind of unmanned plane independent landing terrestrial reference real-time detection tracking of view-based access control model
Ye et al. CT-Net: An efficient network for low-altitude object detection based on convolution and transformer
CN103336971A (en) Target matching method among multiple cameras based on multi-feature fusion and incremental learning
CN112100435A (en) Automatic labeling method based on edge end traffic audio and video synchronization sample
CN112465854A (en) Unmanned aerial vehicle tracking method based on anchor-free detection algorithm
CN116109950A (en) Low-airspace anti-unmanned aerial vehicle visual detection, identification and tracking method
Fernández et al. Robust Real‐Time Traffic Surveillance with Deep Learning
Chen et al. A video-based method with strong-robustness for vehicle detection and classification based on static appearance features and motion features
CN115272882A (en) Discrete building detection method and system based on remote sensing image
Yoshihashi et al. Evaluation of bird detection using time-lapse images around a wind farm
CN114740441A (en) Low-slow small-target radar echo identification method based on small samples
CN112305509A (en) Radar track classification and identification method based on HOG _ SVM
Hsu et al. Learning from counting: Leveraging temporal classification for weakly supervised object localization and detection
CN112329830B (en) Passive positioning track data identification method and system based on convolutional neural network and transfer learning

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
DD01 Delivery of document by public notice
DD01 Delivery of document by public notice

Addressee: Chen Longkun

Document name: Notice of publication and substantive examination of invention patent application

DD01 Delivery of document by public notice
DD01 Delivery of document by public notice

Addressee: Chen Longkun

Document name: Notice of First Examination Opinion

DD01 Delivery of document by public notice

Addressee: Chen Longkun

Document name: Deemed withdrawal notice