CN113759362B - Method, device, equipment and storage medium for radar target data association - Google Patents
Method, device, equipment and storage medium for radar target data association Download PDFInfo
- Publication number
- CN113759362B CN113759362B CN202110859937.4A CN202110859937A CN113759362B CN 113759362 B CN113759362 B CN 113759362B CN 202110859937 A CN202110859937 A CN 202110859937A CN 113759362 B CN113759362 B CN 113759362B
- Authority
- CN
- China
- Prior art keywords
- reference point
- structure information
- signal structure
- information corresponding
- doppler
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 239000011159 matrix material Substances 0.000 claims description 91
- 238000012545 processing Methods 0.000 claims description 70
- 238000012549 training Methods 0.000 claims description 49
- 238000013528 artificial neural network Methods 0.000 claims description 45
- 238000001514 detection method Methods 0.000 claims description 44
- 238000005070 sampling Methods 0.000 claims description 36
- 238000009825 accumulation Methods 0.000 claims description 17
- 230000001427 coherent effect Effects 0.000 claims description 17
- 230000006835 compression Effects 0.000 claims description 17
- 238000007906 compression Methods 0.000 claims description 17
- 238000004891 communication Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 6
- 230000001133 acceleration Effects 0.000 description 11
- 238000004088 simulation Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/66—Radar-tracking systems; Analogous systems
- G01S13/70—Radar-tracking systems; Analogous systems for range tracking only
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for associating radar target data, which are used for acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data; inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point; if the similarity value is smaller than or equal to a preset similarity threshold value, determining that the first reference point and the second reference point are the same target. In the data association process, the structural information of the target echo signals is fully considered, the obtained similarity value is more accurate, and the correct association probability of the target is improved.
Description
Technical Field
The invention belongs to the technical field of radars, and particularly relates to a method, a device, equipment and a storage medium for radar target data association.
Background
Target tracking is used to correlate the same target from the environment where noise, clutter exists. In the initial stage of target tracking, target tracking can be performed by a data correlation method, namely, the target track is extracted from echo data which are scanned continuously for many times without introducing too many false tracks.
The traditional target tracking method is mainly designed based on a target uniform motion model, namely a sequential processing method and a batch processing method which assume that the target does uniform motion.
However, in practical application, if the target is maneuvering, that is, does not do uniform motion, the method is easy to cause model mismatch, and the accuracy of data association is not high.
The Western-type electronic technology university provides a 'radar track starting method based on position information and Doppler information' in a patent application with the application number of CN201610708469 and the application publication number of CN106405537B, wherein the method firstly calculates the maximum non-fuzzy speed and establishes a space position constraint condition according to the maximum speed; then selecting a track head from the measurement vector scanned for the kth time, calculating a radial speed set, and establishing a distance constraint condition according to the radial speed; in the (k+1) th scanning, the measurement vectors which are in the same Doppler channel with the track head and meet the two constraint conditions are associated; and finally, updating the distance constraint condition by using the effective measurement vector, and establishing a stable track according to the track starting criterion. The method assumes that the radial velocity of the target is in the same doppler channel between successive frames, and thus the associated accuracy for the maneuver target decreases.
The Chinese people's free army navy aviation engineering institute provides a Hough transformation rapid track starting method based on speed constraint in a patent application with the application publication number of CN201510061025 and the application publication number of CN 104569923B. The method comprises the steps of firstly carrying out combination pairing and speed constraint on sensor measurement data in each scanning period, deleting part of pairs formed by clutter, solving accurate intersection points of each pair in a parameter space by using a Hough transformation formula, extracting common intersection points by parameter space segmentation and threshold setting to obtain candidate tracks, and finally screening subsequent tracks by using the speed constraint to obtain final confirmed tracks. However, this method assumes that the target moves linearly, which is not compatible with the non-linear movement that would be caused when the target maneuvers, and thus performance degradation occurs when the target maneuvers.
Disclosure of Invention
In order to solve the technical problem of low accuracy of data association in the prior art, the invention provides a method, a device, equipment and a storage medium for radar target data association.
The technical problems to be solved by the invention are realized by the following technical scheme:
in a first aspect, the present invention provides a method of radar target data association, comprising:
Acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data;
inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point;
and if the similarity value is smaller than or equal to a preset similarity threshold value, determining that the first reference point and the second reference point are the same target.
Optionally, the acquiring the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point includes:
acquiring first frame echo data and second frame echo data;
performing target detection processing on the first frame echo data to obtain a plurality of first reference points;
acquiring signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points;
And obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data.
Optionally, the performing target detection processing on the first frame echo data to obtain a plurality of first reference points includes:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data, includes:
for each of the first parametersThe point of examination, confirm the distance unit and Doppler channel that the said first reference point corresponds to; acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Obtaining signal structure information corresponding to the first reference point by data of the Doppler channels, wherein N is r Is an integer greater than 0, N f Is an integer greater than 0;
the performing target detection processing on the second frame echo data to obtain a plurality of second reference points includes:
Sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data, including:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
Optionally, the distance unit and the doppler channel corresponding to the first reference point are centered, and the adjacent N is acquired from the first range-doppler matrix r N of sampling units f The data of the Doppler channels, obtaining the signal structure information corresponding to the first reference point, includes:
acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Data of the Doppler channels are obtained to obtain a first matrix;
taking a module value of the first matrix to obtain signal structure information corresponding to the first reference point;
the distance unit and the Doppler channel corresponding to the second reference point are taken as centers, and adjacent N is acquired from the second distance Doppler matrix r N of sampling units f The data of the Doppler channels, obtaining the signal structure information corresponding to the second reference point, includes:
acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f Obtaining data of the Doppler channels to obtain a second matrix;
and taking a module value of the second matrix to obtain signal structure information corresponding to the second reference point.
Optionally, the first reference point and the second reference point satisfy a maximum speed wave gate constraint condition.
Optionally, the similarity determination model is a trained twin neural network; the twin neural network comprises two full convolution neural networks with the same structure and shared weight.
Optionally, the method further comprises:
training the similarity determination model;
The training of the similarity determination model includes:
obtaining a training dataset comprising: a positive sample data set and a negative sample data set, the positive sample data set comprising a plurality of positive samples; the negative sample dataset comprises a plurality of negative samples; the number of the plurality of positive samples and the plurality of negative samples is the same;
and inputting the training data set into the similarity determination model for training until the change rate of the cost function is smaller than or equal to a preset threshold value, and stopping training to obtain a trained similarity determination model.
In a second aspect, an embodiment of the present invention provides an apparatus for radar target data association, including:
the acquisition module is used for acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data;
the processing module is used for inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point;
And if the similarity value is smaller than or equal to a preset similarity threshold value, determining that the first reference point and the second reference point are the same target.
Optionally, the acquiring module is specifically configured to:
acquiring first frame echo data and second frame echo data;
performing target detection processing on the first frame echo data to obtain a plurality of first reference points;
acquiring signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points;
and obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data.
Optionally, the acquiring module is specifically configured to:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data, includes:
Determining a distance unit and a Doppler channel corresponding to each first reference point; acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Obtaining signal structure information corresponding to the first reference point by data of the Doppler channels, wherein N is r Is an integer greater than 0, N f Is an integer greater than 0;
the performing target detection processing on the second frame echo data to obtain a plurality of second reference points includes:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data, including:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
Optionally, the acquiring module is specifically configured to:
acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Data of the Doppler channels are obtained to obtain a first matrix;
taking a module value of the first matrix to obtain signal structure information corresponding to the first reference point;
acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f Obtaining data of the Doppler channels to obtain a second matrix;
and taking a module value of the second matrix to obtain signal structure information corresponding to the second reference point.
Optionally, the first reference point and the second reference point satisfy a maximum speed wave gate constraint condition.
Optionally, the similarity determination model is a trained twin neural network; the twin neural network comprises two full convolution neural networks with the same structure and shared weight.
Optionally, the apparatus further includes:
The training module is used for training the similarity determination model;
the training module is specifically used for:
obtaining a training dataset comprising: a positive sample data set and a negative sample data set, the positive sample data set comprising a plurality of positive samples; the negative sample dataset comprises a plurality of negative samples; the number of the plurality of positive samples and the plurality of negative samples is the same;
and inputting the training data set into the similarity determination model for training until the change rate of the cost function is smaller than or equal to a preset threshold value, and stopping training to obtain a trained similarity determination model.
In a third aspect, an embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
a processor for performing the method steps as described in the first aspect above when executing a program stored on a memory.
In a fourth aspect, the present invention provides a computer-readable storage medium having stored therein a computer program which, when executed by a processor, carries out the method steps as described in the first aspect above.
The invention has the beneficial effects that:
acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, the second reference point is obtained according to second frame echo data, the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point are input into a similarity determination model, a similarity value between the first reference point and the second reference point is obtained, and if the similarity value is smaller than or equal to a preset similarity threshold value, the first reference point and the second reference point are determined to be the same target. According to the similarity between the signal structure information matrixes corresponding to the reference points in the two frames of echo data, whether the reference points in the two frames of echo data are the same target is determined, so that the structure information of the target echo signals is fully considered in the data association process, the obtained similarity value is more accurate, and the correct association probability of the target is improved.
Drawings
Fig. 1 is a schematic flow chart of a method for associating radar target data according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a twin neural network according to an embodiment of the present invention;
Fig. 3A is a schematic diagram of a radar with expected false alarm probability p=10 according to an embodiment of the present invention -6 When the target acceleration is 0,60m/s 2 ]Probability of correct association of targets within range;
fig. 3B shows that the expected false alarm probability of the radar according to the embodiment of the present invention is p=10 -5 When the target acceleration is 0,60m/s 2 ]Probability of correct association of targets within range;
fig. 3C shows that the expected false alarm probability of the radar according to the embodiment of the present invention is p=10 -4 When the target acceleration is 0,60m/s 2 ]Probability of correct association of targets within range;
fig. 4 is a schematic structural diagram of a device for radar target data association according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to specific examples, but embodiments of the present invention are not limited thereto.
In the field of target tracking, the echo data received by the radar can be subjected to data association, so that the same targets are associated, and the targets can be identified and analyzed later.
According to the radar target data association method provided by the embodiment of the invention, the similarity between the signal structure information matrixes corresponding to the reference points in the two frames of echo data is determined, and whether the reference points in the two frames of echo data are the same target is determined, so that the structure information of the target echo signal is fully considered in the data association process, the obtained similarity value is more accurate, and the correct association probability of the target is improved.
The technical scheme of the invention is described in detail below by specific examples.
Referring to fig. 1, fig. 1 is a flowchart of a method for associating radar target data according to an embodiment of the present invention, as shown in fig. 1, the method of the present embodiment is executed by a computer or a server, and the method provided by the present embodiment includes:
s101, acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point.
The first reference point is obtained according to the first frame echo data, and the second reference point is obtained according to the second frame echo data.
The first frame of echo data and the second frame of echo data may be continuous two frames of echo data or discontinuous two frames of echo data, which is not limited by the present invention.
Wherein the echo data in the present invention may also be referred to as radar echo data.
When a target is detected by the radar, multi-frame echo data can be obtained by echo signals received by the radar.
In the process of data association, two frames of echo data are acquired, wherein the two frames of echo data are respectively a first frame of echo data and a second frame of echo data. The corresponding reference point can be obtained according to the echo data, the first reference point is a reference point obtained according to the echo data of the first frame, and the second reference point is a reference point obtained according to the echo data of the second frame.
And respectively acquiring signal structure information corresponding to the first reference point and signal structure information corresponding to the second reference point.
The signal structure information corresponding to the first reference point is signal structure information in the echo signal corresponding to the first reference point.
The signal structure information corresponding to the second reference point is signal structure information in the echo signal corresponding to the second reference point.
S102, inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point.
And inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point. The similarity value is also the similarity value between the first reference point and the second reference point.
The similarity determination model is used for determining the similarity of the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point. The similarity determination model is a pre-trained network model.
S103, judging whether the similarity value is smaller than or equal to a preset similarity threshold value.
The preset similarity threshold is a preset value larger than zero.
If the similarity value is less than or equal to the preset similarity threshold, the process continues to S104. If the similarity value is greater than the preset similarity threshold, S105 is continued.
S104, determining the first reference point and the second reference point as the same target.
If the similarity value is less than or equal to the preset similarity threshold, the first reference point and the second reference point can be considered as the same target.
S105, determining the first reference point and the second reference point as different targets.
If the similarity value is greater than the preset similarity threshold, the first reference point and the second reference point can be considered as different targets.
In this embodiment, the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point are obtained according to the first frame echo data, the second reference point is obtained according to the second frame echo data, the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point are input into the similarity determining model, a similarity value between the first reference point and the second reference point is obtained, and if the similarity value is smaller than or equal to a preset similarity threshold, the first reference point and the second reference point are determined to be the same target. According to the similarity between the signal structure information matrixes corresponding to the reference points in the two frames of echo data, whether the reference points in the two frames of echo data are the same target is determined, so that the structure information of the target echo signals is fully considered in the data association process, the obtained similarity value is more accurate, and the correct association probability of the target is improved. In addition, the embodiment solves the problem of performance loss caused by insufficient utilization of echo information because only position information and Doppler information are utilized for target data association in the prior art, and is more effectively applied to data association of radar targets.
Further, S101 may be implemented by the following steps based on the embodiment shown in fig. 1:
s1011, acquiring first frame echo data and second frame echo data.
And S1012, performing target detection processing on the first frame echo data to obtain a plurality of first reference points.
And carrying out target detection processing on the echo data of the first frame to obtain a target detection result. The target detection result may include: the first frame of echo data does not contain a reference point, or the first frame of echo data contains a reference point.
If the first frame echo data does not contain the reference point, the subsequent processing is not carried out.
If the first frame echo data includes reference points, wherein the first frame echo data includes a plurality of first reference points, then S1013 is further performed.
S1013, obtaining signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data.
And S1014, performing target detection processing on the second frame echo data to obtain a second reference point.
The implementation of S1014 is similar to S1012 and will not be repeated here.
S1015, obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data.
The implementation of S1015 is similar to S1013 and will not be described here again.
According to the embodiment, the signal structure information corresponding to the first reference point is obtained according to the first reference point and the first frame echo data, the signal structure information corresponding to the second reference point is obtained according to the second reference point and the second frame echo data, the structure information of echo signals is fully considered, the obtained similarity value is more accurate, and the correct association probability of the target is improved.
Further, S1012 may be implemented as follows:
and sequentially carrying out pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix.
And performing pulse compression processing on the echo data of the first frame to obtain distance dimension information. And carrying out coherent accumulation processing on the echo data of the first frame to obtain Doppler channel information.
A first range-doppler matrix may be derived from the range-dimension information and the doppler channel information.
Accordingly, S1013 can be realized by:
determining a distance unit and a Doppler channel corresponding to each first reference point; acquiring adjacent N from a first range-Doppler matrix by taking a range unit corresponding to a first reference point and a Doppler channel as centers r N of sampling units f The data of the individual doppler channels,and obtaining signal structure information corresponding to the first reference point.
Wherein N is r Is an integer greater than 0, N f Is an integer greater than 0.
Further, taking a range unit corresponding to a first reference point and a Doppler channel as centers, and acquiring adjacent N from a first range-Doppler matrix r N of sampling units f The data of the Doppler channels, and the signal structure information corresponding to the first reference point are obtained, can be realized through the following steps:
acquiring adjacent N from a first range-Doppler matrix by taking a range unit corresponding to a first reference point and a Doppler channel as centers r N of sampling units f Data of the Doppler channels are obtained to obtain a first matrix;
and taking a modulus value of the first matrix to obtain signal structure information corresponding to the first reference point.
Further, S1014 may be implemented by, on the basis of the above embodiment, the following means:
and sequentially carrying out pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix.
And performing pulse compression processing on the echo data of the second frame to obtain distance dimension information. And carrying out coherent accumulation processing on the echo data of the second frame to obtain Doppler channel information.
A second range-doppler matrix may be derived from the range-dimension information and the doppler channel information.
Accordingly, S1015 may be implemented as follows:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit corresponding to the second reference point and the Doppler channel as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
Further, the distance unit corresponding to the second reference point is used forThe Doppler channel is taken as the center, and the adjacent N is acquired from the second range-Doppler matrix r N of sampling units f The data of the Doppler channels, and the signal structure information corresponding to the first reference point are obtained, can be realized through the following steps:
acquiring adjacent N from the second range-Doppler matrix by taking a range unit corresponding to the second reference point and the Doppler channel as centers r N of sampling units f Obtaining data of the Doppler channels to obtain a second matrix;
and taking a modulus value of the second matrix to obtain signal structure information corresponding to the second reference point.
On the basis of the above embodiment, further, the method further includes:
Determining N according to the sampling frequency of the radar and the bandwidth of a radar transmitting signal r 。
Alternatively, N r Can be obtained by the following formula (1):
wherein N is r F, the number of sampling units covered in the distance dimension for target echo structure information in the radar echo signal s For the sampling frequency of the radar, B is the bandwidth of the radar transmit signal,representing a rounding up operation.
Further, on the basis of the above embodiment, the first reference point and the second reference point satisfy a maximum speed wave gate constraint condition.
Further, based on the above embodiment, the similarity determination model may be a trained twin neural network.
Further, the twin neural network comprises two full convolution neural networks which have the same structure and share weights.
For example, a twin neural network is constructed which consists of two fully convolutional neural networks of identical structure, sharing weights, consisting of an input encoding layer, a hidden layer and an output decoding layer in turn.
Wherein: the input coding layer adopts a convolution layer with the convolution kernel size of 3*3 and the channel number of 32.
The hidden layer is formed by sequentially connecting 6 convolution layers, wherein:
the number of channels of the 1 st convolution layer is 32, and the convolution kernel size is 3*3;
the 2 nd convolution layer is 64 channels and the convolution kernel size is 3*3;
The 3 rd convolution layer is 64 channels and the convolution kernel size is 3*3;
the 4 th convolution layer is 128 channels and the convolution kernel size is 3*3;
the 5 th convolution layer is 128 channels and the convolution kernel size is 3*3;
and outputting the decoding layer, wherein a convolution layer with the channel number of 1 and the convolution kernel size of 1*1 is adopted.
According to the embodiment, the twin neural network is constructed, the defect that model mismatch is easy to occur when a target maneuver is faced by a model driving method in the prior art is overcome, the constraint of associated target observation can be learned from data in a data driving mode, and the radar target data association performance is improved.
The method provided by the embodiment of the invention is described below by taking the steps A to D as examples.
And step A, acquiring two frames of radar echo data, and respectively performing pulse compression, coherent accumulation and target detection processing on the two frames of radar echo data to obtain two frames of processed radar echo data.
Wherein the two frames of radar echo data comprise: start frame echo data (corresponding to the first frame echo data in the above embodiment) and associated frame echo data (corresponding to the second frame echo data in the above embodiment).
Step B, taking the detection result (the detection result corresponds to the reference point in the embodiment) which is not related to the known track of each processed two frames of radar echo data as the center, and sampling the number N of units according to the distance dimension r And Doppler channel number N f The extraction size isN r ×N f The echo signals covered by the matrix of the (a) form an echo signal matrix aiming at each detection result which is not related to the known flight path, and the echo signal matrix is subjected to modular value to obtain signal structure information.
And C, taking the detection result of each initial frame which is not associated with the known track as a reference point, and acquiring any detection result which is not associated with the known track in the associated frame as an associated point, wherein the detection results are combined into a pair of data pairs.
All data pairs are processed separately using the maximum speed wave gate constraint, excluding unlikely targets in the associated frame. The maximum speed waveguide constraint can be expressed by the following formula (2):
wherein ρ is 1 Represents the distance, θ, of the detected reference point 1 Azimuth information ρ representing the detected reference point 2 Represents the distance, theta, of the detected association point 2 Azimuth information, v, representing the detected correlation point max Representing the maximum possible speed of movement of the object, T representing the scanning period.
And D, inputting the signal structure information matrix of the reference point and the signal structure information matrix of the association point meeting the maximum speed wave gate constraint into a trained twin neural network to obtain the similarity D between the reference point and the association point, and comparing the similarity D with a threshold M obtained by network training, so that the association of the target is completed.
When d is less than or equal to M, the reference point and the association point belong to the same target; otherwise, they do not belong to the same target.
In some other embodiments, before the similarity determination model is used, further comprising: and training a similarity determination model. It will be appreciated that the process of training the similarity determination model may be performed separately or may be performed prior to the method steps of the above-described embodiments.
The step of training the similarity determination model may be performed by a computer or a server, which may be the same as or different from the computer or the server of the method of implementing the radar target data association in the above embodiment.
If the computer or server for training the similarity determination model is different from the computer or server for implementing the method for associating radar target data in the above embodiment, after the training is completed to obtain the similarity determination model, the trained similarity determination model may be sent to the computer or server for implementing the method for associating radar target data.
Wherein training the similarity determination model comprises:
a training dataset is acquired.
Wherein the training data set comprises: a positive sample data set and a negative sample data set, the positive sample data set comprising a plurality of positive samples; the negative sample dataset comprises a plurality of negative samples; the number of the plurality of positive samples and the plurality of negative samples is the same.
And inputting the training data set into the similarity determination model for training until the change rate of the cost function is smaller than or equal to a preset threshold value, and stopping training to obtain the trained similarity determination model.
How the similarity determination model is trained is described below with specific examples. According to the implementation example, aiming at the data association of the uniform acceleration moving target, the twin neural network is trained by constructing a training data set containing target signal structure information, so that the twin neural network learns from data to implement the target data association by using the signal structure information, and the correct association probability of the uniform acceleration moving target with different accelerations is improved.
And 1, constructing a training data set.
Specifically, step 1 may be implemented by the following steps 1.1 to 1.6:
step 1.1, according to the sampling frequency F of the radar s And bandwidth B of radar emission signals, calculating the number N of sampling units covered by target echo structure information in distance dimension in radar echo signals r N can be obtained, for example, by the method of the above formula (1) r 。
Step 1.2, when the obtained target exists, taking the target as a center point in the radar echo signal, and N adjacent distance dimensions r The data of all Doppler channels of each unit are taken as a target signal structure information matrix, and a target signal structure information matrix data set G is constructed t 。
Specifically, step 1.2 may be implemented by the following steps 1.2.1 to 1.2.5:
and 1.2.1, sequentially performing pulse compression and coherent accumulation processing on echo signals received by the radar when the target exists, and obtaining a range-Doppler matrix with the target.
And step 1.2.2, calculating a distance unit and a Doppler channel where the target is located.
Step 1.2.3, sampling the number N of units according to the distance dimension by taking the distance unit where the target is and the Doppler channel as the center from the distance Doppler matrix where the target exists r And Doppler channel number N f Extracting N r ×N f And taking a modulus of the echo of the region covered by the matrix of (2).
Step 1.2.4, taking the echo signal matrix subjected to the modulo processing as a target signal structure information matrix S t 。
Step 1.2.5, repeating the steps 1.2.1-1.2.4 to obtain the target signal structure information matrix S t Composed target signal structure information data set G t 。
Step 1.3, when the acquired target does not exist, taking the false alarm as a center in the radar echo signal, and N adjacent distance dimensions r The data of all Doppler channels of each unit are used as a false alarm signal structure information matrix to construct a false alarm signal structure information matrix data set G f 。
And 1.3.1, sequentially performing pulse compression and coherent accumulation processing on echo signals received by the radar when the target does not exist, and obtaining a range-Doppler matrix.
And 1.3.2, calculating a distance unit and a Doppler channel where the false alarm is located.
Step 1.3.3, from the range-Doppler matrix where no target exists, using the range cell where the false alarm is located andthe Doppler channel is taken as the center, and the number N of sampling units is sampled according to the distance dimension r And Doppler channel number N f Extracting N r ×N f And taking a modulus of the echo of the region covered by the matrix of (2).
Step 1.3.4, taking the echo signal matrix taking the modulus value as a false alarm signal structure information matrix S f 。
Step 1.3.5, returning to the execution of the steps 1.3.1-1.3.5 to obtain the false alarm signal structure information matrix S t Composed false alarm signal structure information matrix data set G f 。
Step 1.4, obtaining a target signal structure information matrix data set G t The signal structure information matrix between two groups of the same target continuous frames forms a positive sample data pair to construct a positive sample data set
Step 1.5, arbitrarily taking the target signal structure information matrix data set G t Is a group of target signal structure information matrix and false alarm signal structure information matrix data set G f A group of false alarm signal structure information matrixes in the data set G are formed into a first negative-sample data pair, and the false alarm signal structure information matrix data set G is arbitrarily taken f Two groups of false alarm signal structure information matrixes in the data-processing unit form a second negative sample data pair, and the first negative sample data pair and the second negative sample data pair which are equal in quantity are taken to form a negative sample data set
Step 1.6, collecting the same number of positive sample data setsAnd negative sample dataset +.>Composing training data set->
And 2, constructing a twin neural network.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a twin neural network according to an embodiment of the present invention. Step 2 is described in detail below in conjunction with fig. 2.
Step 2 may be implemented by steps 2.1-2.3 as follows:
step 2.1, constructing a twin neural network consisting of two full convolution neural networks with the same structure and shared weight, wherein the full convolution neural network consists of an input coding layer, a hidden layer and an output decoding layer in sequence, and the twin neural network comprises the following components:
the input coding layer adopts a convolution layer with the convolution kernel size of 3*3 and the channel number of 32;
the hidden layer is formed by sequentially connecting 6 convolution layers, wherein:
the number of channels of the 1 st convolution layer is 32, and the convolution kernel size is 3*3;
the 2 nd convolution layer is 64 channels and the convolution kernel size is 3*3;
the 3 rd convolution layer is 64 channels and the convolution kernel size is 3*3;
the 4 th convolution layer is 128 channels and the convolution kernel size is 3*3;
the 5 th convolution layer is 128 channels and the convolution kernel size is 3*3;
the output decoding layer adopts a convolution layer with the channel number of 1 and the convolution kernel size of 1*1;
and 2.2, respectively inputting two groups of data in the distance Doppler matrix data pair into two full convolution neural networks (twin neural networks) with the same structure and shared weight to obtain an output data pair, and calculating the distance between the two groups of data in the output data pair to obtain the output d of the twin neural network.
Alternatively, the similarity value d of the twin neural network output may be obtained by the following formula (3):
d＝||f(S1)-f(S 2 ) Formula (3)
Wherein, (S) 1 ,S 2 ) Representing pairs of twin neural network input data, (f (S) 1 ),f(S 2 ) (S) represents a data pair (S) 1 ,S 2 ) The true output data pair obtained when the full convolutional neural network is input, d represents the data pair (S 1 ,S 2 ) The true output obtained when the twin neural network is input.
Step 2.3, setting the cost function of the twin neural network as a cross entropy loss function
Alternatively, the cross entropy loss functionCan be obtained by the following formula (4):
wherein θ represents parameters of each layer of network connected in the twin neural network, which tend to be optimal along with the trend of the cost function in the process of training the twin neural network, L represents the number of data pairs in the training data set, L represents the sequence number of the data pairs, y l Indicating whether the two sets of data in the pair are from the same target, where if y l 1, indicating that the data in the data pair originate from the same target; if y l Equal to 0 indicates that the data in the data pair originates from a different target. d, d l Representing the real output obtained when the first data pair is input into the twin neural network, representing the similarity of two groups of data structure information, M is a preset threshold value of the association probability of the data pair, if the network output d l If the data is smaller than or equal to M, determining that the data in the data pair is from the same target, if the network outputs d l Greater than M, it is determined that the data in the data pair originates from a different target.
And step 3, training the twin neural network.
From training data setsAnd cost function->And training the twin neural network by using a minimum batch gradient descent method to obtain the trained twin neural network.
Specifically, step 3 may be implemented by the following steps 3.1 to 3.5:
and 3.1, setting the size P of the batch of the minimum batch gradient descent method and the network parameter updating step mu.
Step 3.2 from the training data setData with the size of P is randomly selected and is input into a twin neural network, and a corresponding cost function is calculated>/>
Step 3.3, calculating the cost function of the current twin neural networkGradient g with respect to network parameter θ.
And 3.4, updating the parameter theta of the twin neural network to be theta-mug.
Step 3.5, repeating the steps 3.1-3.4 until the cost function of the twin neural networkThe twin neural network is trained.
The beneficial effects of the technical scheme provided by the embodiment of the invention are further described below in combination with simulation experiments.
1. Simulation experiment conditions:
the hardware test platform of the simulation experiment of the invention is: the processor is CPU Xeon E5-2643, the main frequency is 3.4GHz, and the memory is 64GB; the software platform is as follows: ubuntu 18.04LTS,64 bit operating System, python 2.7.
The radar system set in the simulation experiment works in the L wave band, and the transmitted detection signal is a linear frequency modulation signal with the bandwidth of 5 MHz. Both training and testing scenarios assume that the scattering centers of the targets are randomly distributed over a 20m x 8m area.
Assuming that the echo of the target is subject to a Swerling type I distribution, the noise encountered in the detection process is assumed to be internal noise of the receiver, and is white noise subject to a complex gaussian distribution.
The signal-to-noise ratio of the simulation target is 13dB, and the target echo amplitude is subject to the Swerling I-type distribution.
2. Simulation content and simulation result analysis:
under the simulation conditions, the method and the method for starting the radar track based on the position information and the Doppler information are used for carrying out data association on targets with different false alarm probabilities for 1000 times in a complex Gaussian white noise environment, so that the target correct association probability of the target acceleration within the range of 0-60 is obtained.
Referring to fig. 3A to 3C, fig. 3A is a graph showing that the expected false alarm probability of the radar according to the embodiment of the present invention is p=10 -6 When the target acceleration is 0,60m/s 2 ]The probability of correct association of targets within range. Fig. 3B shows that the expected false alarm probability of the radar according to the embodiment of the present invention is p=10 -5 When the target acceleration is 0,60m/s 2 ]The probability of correct association of targets within range. Fig. 3C shows that the expected false alarm probability of the radar according to the embodiment of the present invention is p=10 -4 When the target acceleration is 0,60m/s 2 ]The probability of correct association of targets within range.
As can be seen from FIGS. 3A-3C, the correlation probability is lower than the Doppler information auxiliary correlation method for different false alarm probabilities only when the target moves at a constant speed, and the correlation probability is equal to the correlation probability of the Doppler information auxiliary correlation method for the acceleration (0, 60 m/s) 2 ]The probability of correct association of the target of the uniformly accelerated moving target in the range is higher than that of the association method assisted by Doppler information, which indicates that the method can be suitable for data association of maneuvering targets and can obtain higher association probability.
Referring to fig. 4, fig. 4 is a schematic structural diagram of a device for associating radar target data according to an embodiment of the present invention, and as shown in fig. 4, the device provided in this embodiment includes:
the obtaining module 401 is configured to obtain signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, where the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data;
The processing module 402 is configured to input signal structure information corresponding to the first reference point and signal structure information corresponding to the second reference point into a similarity determination model, so as to obtain a similarity value between the first reference point and the second reference point;
if the similarity value is smaller than or equal to a preset similarity threshold value, determining that the first reference point and the second reference point are the same target.
Optionally, the obtaining module 401 is specifically configured to:
acquiring first frame echo data and second frame echo data;
performing target detection processing on the first frame echo data to obtain a plurality of first reference points;
acquiring signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points;
and obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data.
Optionally, the obtaining module 401 is specifically configured to:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix;
According to each first reference point and the first frame echo data, respectively obtaining signal structure information corresponding to each first reference point, including:
determining a distance unit and a Doppler channel corresponding to each first reference point; centering on a range cell and a Doppler channel corresponding to a first reference point, and selecting a range-Doppler matrix from the first range-Doppler matrixAcquiring adjacent N r N of sampling units f Obtaining signal structure information corresponding to a first reference point from data of the Doppler channels, wherein N is r Is an integer greater than 0, N f Is an integer greater than 0;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points, including:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix;
according to each second reference point and the second frame echo data, respectively obtaining signal structure information corresponding to each second reference point, including:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit corresponding to the second reference point and the Doppler channel as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
Optionally, the obtaining module 401 is specifically configured to:
acquiring adjacent N from a first range-Doppler matrix by taking a range unit corresponding to a first reference point and a Doppler channel as centers r N of sampling units f Data of the Doppler channels are obtained to obtain a first matrix;
taking a modulus value of the first matrix to obtain signal structure information corresponding to a first reference point;
acquiring adjacent N from the second range-Doppler matrix by taking a range unit corresponding to the second reference point and the Doppler channel as centers r N of sampling units f Obtaining data of the Doppler channels to obtain a second matrix;
and taking a modulus value of the second matrix to obtain signal structure information corresponding to the second reference point.
Optionally, the first reference point and the second reference point satisfy a maximum speed wave gate constraint condition.
Optionally, the similarity determination model is a trained twin neural network; the twin neural network comprises two full convolution neural networks which have the same structure and share weights.
Optionally, the apparatus provided in this embodiment further includes:
the training module is used for training the similarity determination model;
the training module is specifically used for:
Acquiring a training data set, the training data set comprising: a positive sample data set and a negative sample data set, the positive sample data set comprising a plurality of positive samples; the negative sample dataset comprises a plurality of negative samples; the number of the plurality of positive samples and the plurality of negative samples are the same;
and inputting the training data set into the similarity determination model for training until the change rate of the cost function is smaller than or equal to a preset threshold value, and stopping training to obtain the trained similarity determination model.
The device of the above embodiment may be used to implement the technical solution of the above method embodiment, and its implementation principle and technical effects are similar, and are not repeated here.
The embodiment of the invention provides electronic equipment, which comprises a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are communicated with each other through the communication bus;
a memory for storing a computer program;
and a processor configured to implement the method steps of any of the embodiments described above when executing a program stored on a memory.
An embodiment of the present invention provides a computer-readable storage medium, in which a computer program is stored, which when executed by a processor implements the method steps of any of the embodiments described above.
In the description of the present invention, it should be understood that the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
In the present invention, unless expressly stated or limited otherwise, a first feature "above" or "below" a second feature may include both the first and second features being in direct contact, as well as the first and second features not being in direct contact but being in contact with each other through additional features therebetween. Moreover, a first feature being "above," "over" and "on" a second feature includes the first feature being directly above and obliquely above the second feature, or simply indicating that the first feature is higher in level than the second feature. The first feature being "under", "below" and "beneath" the second feature includes the first feature being directly under and obliquely below the second feature, or simply means that the first feature is less level than the second feature.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Further, one skilled in the art can engage and combine the different embodiments or examples described in this specification.
The foregoing is a further detailed description of the invention in connection with the preferred embodiments, and it is not intended that the invention be limited to the specific embodiments described. It will be apparent to those skilled in the art that several simple deductions or substitutions may be made without departing from the spirit of the invention, and these should be considered to be within the scope of the invention.
Claims (8)
1. A method of radar target data correlation, comprising:
Acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data;
inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point;
if the similarity value is smaller than or equal to a preset similarity threshold value, determining that a first reference point and the second reference point are the same target;
the obtaining the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point includes:
acquiring first frame echo data and second frame echo data;
performing target detection processing on the first frame echo data to obtain a plurality of first reference points;
acquiring signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points;
Obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data;
the target detection processing is performed on the first frame echo data to obtain a plurality of first reference points, including:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data, includes:
determining a distance unit and a Doppler channel corresponding to each first reference point; acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Obtaining signal structure information corresponding to the first reference point by data of the Doppler channels, wherein N is r Is an integer greater than 0, N f Is an integer greater than 0;
the performing target detection processing on the second frame echo data to obtain a plurality of second reference points includes:
Sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data, including:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
2. The method of claim 1, wherein the acquiring adjacent N from the first range-doppler matrix is centered on a range bin and a doppler channel corresponding to the first reference point r N of sampling units f The data of the Doppler channels, obtaining the signal structure information corresponding to the first reference point, includes:
acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Data of the Doppler channels are obtained to obtain a first matrix;
taking a module value of the first matrix to obtain signal structure information corresponding to the first reference point;
the distance unit and the Doppler channel corresponding to the second reference point are taken as centers, and adjacent N is acquired from the second distance Doppler matrix r N of sampling units f The data of the Doppler channels, obtaining the signal structure information corresponding to the second reference point, includes:
acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f Obtaining data of the Doppler channels to obtain a second matrix;
and taking a module value of the second matrix to obtain signal structure information corresponding to the second reference point.
3. The method according to any of claims 1-2, wherein the first reference point and the second reference point satisfy a maximum speed wave gate constraint.
4. The method of any one of claims 1-2, wherein the similarity determination model is a trained twin neural network; the twin neural network comprises two full convolution neural networks with the same structure and shared weight.
5. The method according to any one of claims 1-2, wherein the method further comprises:
training the similarity determination model;
the training of the similarity determination model includes:
obtaining a training dataset comprising: a positive sample data set and a negative sample data set, the positive sample data set comprising a plurality of positive samples; the negative sample dataset comprises a plurality of negative samples; the number of the plurality of positive samples and the plurality of negative samples is the same;
and inputting the training data set into the similarity determination model for training until the change rate of the cost function is smaller than or equal to a preset threshold value, and stopping training to obtain a trained similarity determination model.
6. An apparatus for radar target data association, comprising:
the acquisition module is used for acquiring signal structure information corresponding to a first reference point and signal structure information corresponding to a second reference point, wherein the first reference point is obtained according to first frame echo data, and the second reference point is obtained according to second frame echo data;
the processing module is used for inputting the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point into a similarity determination model to obtain a similarity value between the first reference point and the second reference point; if the similarity value is smaller than or equal to a preset similarity threshold value, determining that a first reference point and the second reference point are the same target;
The obtaining the signal structure information corresponding to the first reference point and the signal structure information corresponding to the second reference point includes:
acquiring first frame echo data and second frame echo data;
performing target detection processing on the first frame echo data to obtain a plurality of first reference points;
acquiring signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data;
performing target detection processing on the second frame echo data to obtain a plurality of second reference points;
obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data;
the target detection processing is performed on the first frame echo data to obtain a plurality of first reference points, including:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the first frame echo data to obtain a plurality of first reference points and a first range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each first reference point according to each first reference point and the first frame echo data, includes:
Determining a distance unit and a Doppler channel corresponding to each first reference point; acquiring adjacent N from the first range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the first reference point as centers r N of sampling units f Obtaining signal structure information corresponding to the first reference point by data of the Doppler channels, wherein N is r Is an integer greater than 0, N f Is an integer greater than 0;
the performing target detection processing on the second frame echo data to obtain a plurality of second reference points includes:
sequentially performing pulse compression processing, coherent accumulation processing and target detection processing on the second frame echo data to obtain a plurality of second reference points and a second range-Doppler matrix;
the step of respectively obtaining signal structure information corresponding to each second reference point according to each second reference point and the second frame echo data, including:
determining a distance unit and a Doppler channel corresponding to each second reference point; acquiring adjacent N from the second range-Doppler matrix by taking a range unit and a Doppler channel corresponding to the second reference point as centers r N of sampling units f And obtaining signal structure information corresponding to the first reference point by the data of the Doppler channels.
7. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
a processor for carrying out the method steps of any one of claims 1-5 when executing a program stored on a memory.
8. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored therein a computer program which, when executed by a processor, implements the method steps of any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110859937.4A CN113759362B (en) | 2021-07-28 | 2021-07-28 | Method, device, equipment and storage medium for radar target data association |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110859937.4A CN113759362B (en) | 2021-07-28 | 2021-07-28 | Method, device, equipment and storage medium for radar target data association |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113759362A CN113759362A (en) | 2021-12-07 |
| CN113759362B true CN113759362B (en) | 2024-02-23 |
Family
ID=78788086
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110859937.4A Active CN113759362B (en) | 2021-07-28 | 2021-07-28 | Method, device, equipment and storage medium for radar target data association |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113759362B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114966595A (en) * | 2022-05-20 | 2022-08-30 | 加特兰微电子科技(上海)有限公司 | MIMO sensor, arrival angle approximation degree judging method and target information matching method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008084465A2 (en) * | 2007-01-08 | 2008-07-17 | Rafael - Armament Development Authority Ltd. | Method and system for synthetic aperture radar mapping and navigation using digital pulse compression |
| CN102073043A (en) * | 2010-11-04 | 2011-05-25 | 电子科技大学 | Multi-frame phase coherence accumulation target tracking-before-detecting method |
| CN104535977A (en) * | 2014-09-04 | 2015-04-22 | 武汉滨湖电子有限责任公司 | GSM signal based radar target detection method |
| CN107765228A (en) * | 2017-09-29 | 2018-03-06 | 西安电子科技大学 | A kind of online radar target detection method based on region similitude |
| CN111209462A (en) * | 2020-01-02 | 2020-05-29 | 北京字节跳动网络技术有限公司 | Data processing method, device and equipment |
| CN111540027A (en) * | 2020-04-23 | 2020-08-14 | 上海高仙自动化科技发展有限公司 | Detection method, detection device, electronic equipment and storage medium |
| WO2021077287A1 (en) * | 2019-10-22 | 2021-04-29 | 华为技术有限公司 | Detection method, detection device, and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2026280B1 (en) * | 2007-07-23 | 2013-10-23 | Esaote S.p.A. | Method and corresponding apparatus for quantitative measurements on sequences of images, particularly ultrasonic images |
-
2021
- 2021-07-28 CN CN202110859937.4A patent/CN113759362B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008084465A2 (en) * | 2007-01-08 | 2008-07-17 | Rafael - Armament Development Authority Ltd. | Method and system for synthetic aperture radar mapping and navigation using digital pulse compression |
| CN102073043A (en) * | 2010-11-04 | 2011-05-25 | 电子科技大学 | Multi-frame phase coherence accumulation target tracking-before-detecting method |
| CN104535977A (en) * | 2014-09-04 | 2015-04-22 | 武汉滨湖电子有限责任公司 | GSM signal based radar target detection method |
| CN107765228A (en) * | 2017-09-29 | 2018-03-06 | 西安电子科技大学 | A kind of online radar target detection method based on region similitude |
| WO2021077287A1 (en) * | 2019-10-22 | 2021-04-29 | 华为技术有限公司 | Detection method, detection device, and storage medium |
| CN111209462A (en) * | 2020-01-02 | 2020-05-29 | 北京字节跳动网络技术有限公司 | Data processing method, device and equipment |
| CN111540027A (en) * | 2020-04-23 | 2020-08-14 | 上海高仙自动化科技发展有限公司 | Detection method, detection device, electronic equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 一种逆合成孔径激光雷达成像算法;杨小优;池龙;何劲;雷强;张群;;激光与红外(第08期);全文 * |
| 利用回波幅度信息辅助的模糊聚类关联算法;欧阳超;刘强;舒建国;;计算机工程与设计(第20期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113759362A (en) | 2021-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108460382B (en) | Optical remote sensing image ship detection method based on deep learning single-step detector | |
| CN107220606B (en) | Radar radiation source signal identification method based on one-dimensional convolutional neural network | |
| CN112184849B (en) | Intelligent processing method and system for complex dynamic multi-target micro-motion signals | |
| CN112001270B (en) | Ground radar automatic target classification and identification method based on one-dimensional convolutional neural network | |
| CN112990082B (en) | Detection and identification method of underwater sound pulse signal | |
| CN111880157A (en) | Method and system for detecting target in radar image | |
| CN114895263A (en) | Radar active interference signal identification method based on deep migration learning | |
| CN112684427A (en) | Radar target identification method based on serial quadratic reinforcement training | |
| CN113759362B (en) | Method, device, equipment and storage medium for radar target data association | |
| CN113516037A (en) | Marine vessel track segment association method, system, storage medium and equipment | |
| CN114117912A (en) | Sea clutter modeling and inhibiting method under data model dual drive | |
| Dai et al. | An efficient water segmentation method for SAR images | |
| CN113608193A (en) | Radar multi-target distance and speed estimation method based on UNet | |
| CN115601604B (en) | Multi-task microbubble track tracking method based on long-time and short-time memory network | |
| Chen et al. | Multi-Function Radar Work Mode Recognition Based on Encoder-Decoder Model | |
| CN114898222B (en) | Ship target track identification method and device | |
| Pan et al. | Ship detection using online update of clutter map based on fuzzy statistics and spatial property | |
| CN115909086A (en) | SAR target detection and identification method based on multistage enhanced network | |
| CN115600101A (en) | Unmanned aerial vehicle signal intelligent detection method and device based on priori knowledge | |
| CN113126086B (en) | Life detection radar weak target detection method based on state prediction accumulation | |
| CN115345216A (en) | FMCW radar interference elimination method fusing prior information | |
| CN113376610B (en) | Narrow-band radar target detection method based on signal structure information | |
| Song et al. | High-Accuracy Gesture Recognition using Mm-Wave Radar Based on Convolutional Block Attention Module | |
| CN112698295A (en) | Knowledge-assisted radar detection and tracking integrated method and system | |
| CN117310641A (en) | Sea clutter suppression method combining phase processing mechanism and LSTM network model |
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 |