CN114076924A - Method for realizing classification and identification of large complex target based on multi-frequency echo data - Google Patents
Method for realizing classification and identification of large complex target based on multi-frequency echo data Download PDFInfo
- Publication number
- CN114076924A CN114076924A CN202010845657.3A CN202010845657A CN114076924A CN 114076924 A CN114076924 A CN 114076924A CN 202010845657 A CN202010845657 A CN 202010845657A CN 114076924 A CN114076924 A CN 114076924A
- Authority
- CN
- China
- Prior art keywords
- target
- model
- echo data
- classification
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 89
- 238000013145 classification model Methods 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims description 25
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 18
- 239000000463 material Substances 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 12
- 230000001413 cellular effect Effects 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 10
- 238000011176 pooling Methods 0.000 claims description 10
- 230000005526 G1 to G0 transition Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 5
- 230000003287 optical effect Effects 0.000 claims description 5
- 239000010410 layer Substances 0.000 description 40
- 238000004891 communication Methods 0.000 description 14
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 8
- 235000006506 Brasenia schreberi Nutrition 0.000 description 6
- 238000005259 measurement Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000004088 simulation Methods 0.000 description 4
- 238000005452 bending Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 229910001092 metal group alloy Inorganic materials 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000003973 paint Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000010206 sensitivity analysis Methods 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
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
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a method for realizing large-scale complex target classification and identification based on multi-frequency echo data, which comprises the following steps: receiving radar echo data; inputting radar echo data into a target classification model which is trained in advance to obtain a target classification recognition result; the target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in multiple frequencies. The invention can improve the identification efficiency of radar target identification and reduce the false alarm rate.
Description
Technical Field
The invention belongs to the technical field of computational electromagnetism, and particularly relates to a method for realizing large-scale complex target classification and identification based on multi-frequency echo data.
Background
The radar target identification is an important development direction of modern radars, has high technical difficulty, has wide application prospect in aerospace technology and strategic early warning, and is a key and hot research direction in the field of radars.
In the prior art, a radar target identification method is mainly realized based on radar image analysis, and in the process of forming a radar image, a series of processing such as pulse compression, matched filtering and the like needs to be carried out by utilizing actually measured echo data, so that the computing resources are consumed, and the identification efficiency is low; in addition, a part of echo information in the original echo data can be lost in the process of forming the radar image, so that the false alarm rate of target identification is higher.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a method for realizing large-scale complex target classification and identification based on multi-frequency echo data.
The technical problem to be solved by the invention is realized by the following technical scheme:
in a first aspect, the present invention provides a method for implementing classification and identification of large complex targets based on multi-frequency echo data, including:
receiving radar echo data;
inputting the radar echo data into a target classification model which is trained in advance to obtain a target classification recognition result;
the target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in the multiple frequencies.
Optionally, the multi-frequency echo database is constructed in a manner that:
constructing geometric models of various targets;
aiming at each target, respectively constructing corresponding dielectric parameter models for all surface elements included in the geometric model of the target based on the material and the structure of the target shell to obtain an initial modeling model of the target;
respectively constructing a first scattered field model of each target based on the initial modeling model of each target;
aiming at each target, analyzing and determining a plurality of sensitive parameter combinations of the first scattered field model of the target by using an orthogonal analysis method, and substituting each sensitive parameter combination into the first scattered field model of the target respectively to obtain a plurality of second scattered field models of the target; the sensitive parameter combination is a combination of target parameters and/or radar parameters, wherein the influence on the echo amplitude of the target exceeds a threshold value, and the target parameters are parameters of the target;
and respectively substituting various different radar working frequencies into each second scattered field model to obtain the multi-frequency echo database.
Optionally, the constructing of the geometric models of the various objects includes:
acquiring initial interpolation surface node information and initial interpolation grid units of each target;
aiming at each target belonging to a flying target, constructing a geometric model of the target by using a bi-quadric surface interpolation method based on the obtained interpolation surface node information of the target and the interpolation grid unit;
aiming at each target belonging to the water body target, based on the obtained interpolation surface node information of the target and the interpolation grid unit, a plane structure and an edge structure on a target shell are constructed by using a non-uniform triangular grid division method, and a preset fine structure on the target shell is constructed by using a cellular grid division method, so that a geometric model of the target is obtained.
Optionally, the separately constructing a first scattered field model of each target based on the initial modeling model of each target includes:
aiming at each target belonging to a flying target, based on an initial modeling model of the target, a scattering field of a plane structure and a scattering field of a curved surface structure on a target shell are constructed by utilizing an interpolation surface integral method of a stationary phase method, and a scattering field of a marginal structure on the target shell is constructed by utilizing an interpolation surface diffraction integral method, so that a first scattering field model of the target is obtained;
aiming at each target belonging to a water body target, based on an initial modeling model of the target, a triangular patch optical tracking method is utilized to construct a scattered field of a surface structure on a target shell, a cellular grid differential method is utilized to construct a scattered field of a preset fine structure on the target shell, and a triangular surface diffraction integral method is utilized to construct a scattered field of a seamed edge structure on the target shell, so that a first scattered field model of the target is obtained; wherein, the face class structure includes at least: multi-angle surface structure, plane structure, curved surface structure and cavity structure.
Optionally, the network structure of the object classification model includes: the device comprises a feature extraction layer, a full connection layer and a Softmax output layer;
the characteristic extraction layer comprises a plurality of cascaded characteristic extraction units, and each characteristic extraction unit comprises a convolution layer and a pooling layer; in the feature extraction layer, adjacent convolution layers and pooling layers are connected through an activation function;
the fully-connected layer includes one or more.
Optionally, the echo data in the multi-frequency echo database are normalized echo data;
the step of inputting the radar echo data into a pre-trained target classification model to obtain a target classification recognition result comprises the following steps:
carrying out normalization processing on the radar echo data;
and inputting the radar echo data after the normalization processing into a target classification model which is trained in advance to obtain a target classification recognition result.
In a second aspect, the present invention provides an apparatus for implementing classification and identification of large complex targets based on multi-frequency echo data, including:
the receiving module is used for receiving radar echo data;
the model application module is used for inputting the radar echo data into a pre-trained target classification model to obtain a target classification recognition result;
the target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in the multiple frequencies.
Optionally, the multi-frequency echo database is constructed in a manner that:
constructing geometric models of various targets;
aiming at each target, respectively constructing corresponding dielectric parameter models for all surface elements included in the geometric model of the target based on the material and the structure of the target shell to obtain an initial modeling model of the target;
respectively constructing a first scattered field model of each target based on the initial modeling model of each target;
aiming at each target, analyzing and determining a plurality of sensitive parameter combinations of the first scattered field model of the target by using an orthogonal analysis method, and substituting each sensitive parameter combination into the first scattered field model of the target respectively to obtain a plurality of second scattered field models of the target; the sensitive parameter combination is a combination of target parameters and/or radar parameters, wherein the influence on the echo amplitude of the target exceeds a threshold value, and the target parameters are parameters of the target;
and respectively substituting various different radar working frequencies into each second scattered field model to obtain the multi-frequency echo database.
Optionally, the constructing of the geometric models of the various objects includes:
acquiring initial interpolation surface node information and initial interpolation grid units of each target;
aiming at each target belonging to a flying target, constructing a geometric model of the target by using a bi-quadric surface interpolation method based on the obtained interpolation surface node information of the target and the interpolation grid unit;
aiming at each target belonging to the water body target, based on the obtained interpolation surface node information of the target and the interpolation grid unit, a plane structure and an edge structure on a target shell are constructed by using a non-uniform triangular grid division method, and a preset fine structure on the target shell is constructed by using a cellular grid division method, so that a geometric model of the target is obtained.
Optionally, the separately constructing a first scattered field model of each target based on the initial modeling model of each target includes:
aiming at each target belonging to a flying target, based on an initial modeling model of the target, a scattering field of a plane structure and a scattering field of a curved surface structure on a target shell are constructed by utilizing an interpolation surface integral method of a stationary phase method, and a scattering field of a marginal structure on the target shell is constructed by utilizing an interpolation surface diffraction integral method, so that a first scattering field model of the target is obtained;
aiming at each target belonging to a water body target, based on an initial modeling model of the target, a triangular patch optical tracking method is utilized to construct a scattered field of a surface structure on a target shell, a cellular grid differential method is utilized to construct a scattered field of a preset fine structure on the target shell, and a triangular surface diffraction integral method is utilized to construct a scattered field of a seamed edge structure on the target shell, so that a first scattered field model of the target is obtained; wherein, the face class structure includes at least: multi-angle surface structure, plane structure, curved surface structure and cavity structure.
Optionally, the network structure of the object classification model includes: the device comprises a feature extraction layer, a full connection layer and a Softmax output layer;
the characteristic extraction layer comprises a plurality of cascaded characteristic extraction units, and each characteristic extraction unit comprises a convolution layer and a pooling layer; in the feature extraction layer, adjacent convolution layers and pooling layers are connected through an activation function;
the fully-connected layer includes one or more.
Optionally, the echo data in the multi-frequency echo database are normalized echo data;
the model application module is specifically configured to:
carrying out normalization processing on the radar echo data;
and inputting the radar echo data after the normalization processing into a target classification model which is trained in advance to obtain a target classification recognition result.
In a third aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the method steps of the method for realizing the classification and identification of the large-scale complex target based on the multi-frequency echo data when executing the program stored in the memory.
In a fourth aspect, the present invention provides a computer-readable storage medium, in which a computer program is stored, and the computer program, when being executed by a processor, implements the method steps of the above method for implementing classification and identification of large and complex targets based on multi-frequency echo data.
In yet another aspect of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the method steps of the above-described method for performing classification and identification of large complex targets based on multi-frequency echo data.
In the method for realizing large-scale complex target classification and identification based on multi-frequency echo data, a radar image is not required to be formed, and a target classification and identification result can be obtained directly based on original radar echo data, so that computing resources are saved, and the identification efficiency is higher; in addition, because the original radar echo data does not lose echo information, the target classification and identification result obtained by the method is accurate, and compared with the prior art, the false alarm rate can be reduced.
In addition, after the second scattered field models of various targets under different frequencies and different parameter combinations are constructed, the multi-frequency echo database with huge data volume can be directly obtained without actual measurement, and manpower and material resources required by actual measurement are saved.
The present invention will be described in further detail with reference to the accompanying drawings.
Drawings
Fig. 1 is a schematic flowchart of a method for implementing classification and identification of a large complex target based on multi-frequency echo data according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a construction method of a multi-frequency echo database according to an embodiment of the present invention;
FIG. 3(a) is a schematic diagram of an exemplary geometric model of a large complex water target;
FIG. 3(b) is a schematic diagram of an exemplary geometric model of a large complex flying target;
fig. 4(a) is a schematic simulation diagram of a set of multi-frequency echo data obtained after a second scattering field model is constructed for the large complex water target in fig. 3(a) by using the flow shown in fig. 2;
fig. 4(b) is a schematic simulation diagram of a set of multi-frequency echo data obtained after a second scattered field model is constructed for the large complex flying target in fig. 3(b) by using the flow shown in fig. 2;
FIG. 5 is a schematic diagram of a network structure of a target classification model provided by an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an apparatus for implementing classification and identification of a large complex target based on multi-frequency echo data according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of an electronic device 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 the embodiments of the present invention are not limited thereto.
In order to improve the identification efficiency of radar target identification and reduce the false alarm rate, the embodiment of the invention provides a method, a device and electronic equipment for realizing large-scale complex target classification identification based on multi-frequency echo data. The device provided by the embodiment of the invention is an execution main body of the method provided by the embodiment of the invention, and the device can be applied to the electronic equipment provided by the embodiment of the invention. In practical applications, the electronic device may be a desktop computer, a portable computer, a radar, or the like, which is not limited herein, and any electronic device that can implement the present invention is within the protection scope of the present invention.
First, a method for implementing classification and identification of large complex targets based on multi-frequency echo data according to an embodiment of the present invention is described in detail. As shown in fig. 1, the method may include the steps of:
s101: and receiving radar echo data.
It will be appreciated that the radar return data is received by a receiver of the radar.
S102: and inputting the radar echo data into a pre-trained target classification model to obtain a target classification recognition result.
The target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in multiple frequencies.
It can be understood that, in the embodiment of the invention, a novel radar target identification method is provided, and the method does not need to form a radar image, and can obtain a target classification identification result directly based on original radar echo data, so that the calculation resource is saved, and the identification efficiency is higher; in addition, because the original radar echo data does not lose echo information, the target classification and identification result obtained by the embodiment of the invention is accurate, and compared with the prior art, the false alarm rate can be reduced.
In order to obtain a large-data-volume multi-frequency echo database and improve the accuracy of a target classification recognition result, an embodiment of the present invention provides a preferred method for constructing a multi-frequency echo database, which includes the following steps, as shown in fig. 2:
s201: geometric models of various objects are constructed.
In practical application, various targets involved in an actual radar target recognition scene can be referred to, and a geometric model of the various targets can be respectively constructed. The various targets referred to herein may include various large complex targets such as various flight targets and various water targets, among others. Of course, the method for constructing the multi-frequency echo database provided by the embodiment of the invention is also applicable to small targets.
Fig. 3(a) schematically shows the geometry of a large complex water target, and fig. 3(b) schematically shows the geometry of a large complex flying target. It will be appreciated that for large complex objects, different parts of their shell may be constructed using different geometric modelling approaches. For clarity of the layout of the solution, a specific implementation of building a geometric model of various objects is illustrated in the following.
S202: and aiming at each target, respectively constructing corresponding dielectric parameter models for all surface elements included in the geometric model of the target based on the material and the structure of the target shell to obtain an initial modeling model of the target.
It can be understood that not only is the geometric structure of the large complex target complex, but also the material of the large complex target changes correspondingly due to different functions realized by different parts. For example, the outer shell of the water body target is mostly made of metal alloy, and in order to prolong the service life of the water body target to the maximum extent, the surface of the outer shell is usually covered with paint coating materials and the like; meanwhile, in order to improve the navigational speed and the flexibility, composite materials are mostly adopted at the positions of the deck, the cabin body and the like of the water body target. Therefore, in the embodiment of the invention, a corresponding dielectric parameter model is established according to the actual material and structure of the target shell. The structure of the target shell may be a single-layer structure or a multi-layer structure, so that when constructing the dielectric parameter model, in addition to the material of the target shell, the actual structure of each part of the target shell needs to be considered, and a corresponding dielectric parameter model is constructed, so as to obtain an initial modeling model of the target. It will be appreciated that the initial modeling model may express the geometry of the target and the electromagnetic properties with respect to the electrical conductivity.
S203: and respectively constructing a first scattered field model of each target based on the initial modeling model of each target.
It can be understood that the structure of the large-scale flying target is relatively simple and mainly comprises a cylinder-like surface, a plane, a partial edge and the like, while the large-scale water body target comprises a plane-like structure, a multi-angle structure, an edge, a sharp top, a bending discontinuity and the like, the two types of targets have different structural characteristics, and the construction modes of the corresponding first scattered field models can be distinguished. The manner in which the first scatterfield models of the flight target and the water target are constructed will be described in the following.
S204: and aiming at each target, analyzing and determining a plurality of sensitive parameter combinations of the first scattered field model of the target by using an orthogonal analysis method, and substituting each sensitive parameter combination into the first scattered field model of the target respectively to obtain a plurality of second scattered field models of the target.
The sensitive parameter is a combination of target parameters and/or radar parameters, which affect the echo amplitude of the target beyond a threshold, where the target parameters refer to parameters of the target, and may include various parameters such as the type, size, posture, and position of the target, and are not limited to those shown here; the radar parameters may include various parameters such as the operating frequency, bandwidth, angle, and polarization of the radar, and are not limited to those shown herein.
It can be understood that the influence degrees of different parameters on the echo amplitude of the target are inconsistent, and in order to train a more accurate target classification model by using the multi-frequency echo database, the embodiment of the invention further performs parameter sensitivity analysis on the basis of the first scattered field model of the target, thereby determining a plurality of groups of typical parameters which have important influence on the echo amplitude characteristic and taking the typical parameters as a plurality of sensitive parameter combinations. Then, substituting each sensitive parameter combination into the first scattered field model of the target respectively, thereby obtaining a plurality of second scattered field models; these second scatter field models thus encompass the scatter field of the object in various situations in actual operation.
S205: and respectively substituting various different radar working frequencies into each second scattered field model to obtain a multi-frequency echo database.
Specifically, parameters representing the radar working frequency in each second scattered field model are set to different frequencies respectively, so that the real scattered fields of the target under different frequencies can be simulated, massive multi-frequency echo data can be obtained, and a multi-frequency echo database is formed. Therefore, after the second scattered field models of various targets under different frequencies and different parameter combinations are constructed, the embodiment of the invention can directly obtain the multi-frequency echo database with huge data volume without actual measurement, thereby saving manpower and material resources required by actual measurement. Therefore, the established neural network model is trained by using the multi-frequency echo database, and a trained target classification model can be obtained.
Next, a specific implementation of constructing the geometric model of each object in step S201 will be exemplarily described.
Illustratively, constructing geometric models of various objects may include:
acquiring initial interpolation surface node information and initial interpolation grid units of each target;
aiming at each target belonging to a flying target, constructing a geometric model of the target by using a bi-quadric surface interpolation method based on the obtained interpolation surface node information of the target and the interpolation grid unit;
aiming at each target belonging to a water body target, constructing a plane structure and an edge structure on a target shell by using a non-uniform triangular mesh generation method and constructing a fine structure on the target shell by using a cellular mesh generation method based on the obtained interpolation surface node information of the target and an interpolation mesh unit to obtain a geometric model of the target; the predetermined fine structure may be selected by human experience, including but not limited to the edge, peak, and discontinuous bending structure mentioned in step S203.
In practical application, the initial interpolation surface node information and the initial interpolation grid unit of each target can be generated by using relevant software of a geometric modeling class. Then, for different targets, the interpolation grid unit is further developed according to the corresponding method shown in the embodiment of the invention, and a point cloud model of the target, namely a geometric model of the target, is obtained.
It is understood that, for a water body target or a flight target with a shell structure different from a conventional target of the same type, the specific construction mode of the geometric model is not limited to that shown here, and the geometric model can be constructed by referring to the modeling methods suitable for the various structures shown above according to the structural characteristics of the various parts of the shell.
Accordingly, in step S203, constructing a first scattered field model of each target based on the initial modeling model of each target, respectively, may include:
aiming at each target belonging to a flying target, based on an initial modeling model of the target, a scattering field of a plane structure and a scattering field of a curved surface structure on a target shell are constructed by utilizing an interpolation surface integral method of a stationary phase method, and a scattering field of a marginal structure on the target shell is constructed by utilizing an interpolation surface diffraction integral method, so that a first scattering field model of the target is obtained;
aiming at each target belonging to a water body target, based on an initial modeling model of the target, a triangular patch optical tracking method is utilized to construct a scattered field of a surface structure on a target shell, a cellular grid differential method is utilized to construct a scattered field of a preset fine structure on the target shell, and a triangular surface diffraction integral method is utilized to construct a scattered field of a seamed edge structure on the target shell, so that a first scattered field model of the target is obtained; wherein, the surface structure of the water body target at least comprises: multi-angle surface structure, plane structure, curved surface structure and cavity structure.
Similarly, for a water body target or a flight target with a shell structure different from a similar conventional target, the specific construction method of the first scattered field model is not limited to that shown here, and the first scattered field model can be constructed according to the structural characteristics of each part of the shell of the first scattered field model and by referring to the scattered field construction methods suitable for the various structures shown above.
For example, according to the above process, after the second scattering field model is constructed for the large complex water target in fig. 3(a), the obtained simulation diagram of a set of multi-frequency echo data is shown in fig. 4(a), and after the second scattering field model is constructed for the large complex flying target in fig. 3(b), the obtained simulation diagram of a set of multi-frequency echo data is shown in fig. 4 (b). In fig. 4(a) and 4(b), Φ and θ represent an incident angle and a scattering angle of the electromagnetic wave, respectively; RCS (radar Cross section) represents a radar scattering Cross section.
Next, the object classification model used in the embodiment of the present invention is exemplified. As shown in fig. 5, the network structure of the object classification model may include: a feature extraction layer 501, a full connection layer 502 and a Softmax output layer 503; the feature extraction layer comprises a plurality of cascaded feature extraction units 5011, and each feature extraction unit 5011 comprises a convolution layer and a pooling layer; in the feature extraction layer 501, adjacent convolutional layers and pooling layers are connected by an activation function, which may be a Linear correction Unit (ReLu).
In practical application, when a target classification model is trained, training samples can be constructed based on each echo data in the multi-frequency echo database and a target classification label corresponding to each echo data, the training samples are used for training the target classification model, and a loss value of the model is calculated; and when the loss value is smaller than the preset loss value threshold, finishing the training to obtain the trained target classification model.
In addition, in order to facilitate training and using the target classification model, the echo data in the multi-frequency echo database can be processed into normalized echo data; correspondingly, in step S101 of fig. 1, the step of inputting the radar echo data into the pre-trained target classification model to obtain the target classification recognition result may include:
carrying out normalization processing on radar echo data;
and inputting the radar echo data after the normalization processing into a target classification model which is trained in advance to obtain a target classification recognition result.
Corresponding to the above method for implementing classification and identification of large and complex targets based on multi-frequency echo data, an embodiment of the present invention further provides an apparatus for implementing classification and identification of large and complex targets based on multi-frequency echo data, as shown in fig. 6, the apparatus may include:
the receiving module 601 is configured to receive radar echo data.
The model application module 602 is configured to input radar echo data to a target classification model which is trained in advance, so as to obtain a target classification recognition result;
the target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in multiple frequencies.
Optionally, the multi-frequency echo database is constructed in a manner that:
constructing geometric models of various targets;
aiming at each target, respectively constructing corresponding dielectric parameter models for all surface elements included in the geometric model of the target based on the material and the structure of the target shell to obtain an initial modeling model of the target;
respectively constructing a first scattered field model of each target based on the initial modeling model of each target;
aiming at each target, analyzing and determining a plurality of sensitive parameter combinations of the first scattered field model of the target by using an orthogonal analysis method, and substituting each sensitive parameter combination into the first scattered field model of the target respectively to obtain a plurality of second scattered field models of the target; the sensitive parameter combination is a combination of target parameters and/or radar parameters, the influence of the sensitive parameter combination on the target echo amplitude exceeds a threshold value, and the target parameters are parameters of a target;
and respectively substituting various different radar working frequencies into each second scattered field model to obtain the multi-frequency echo database.
Optionally, geometric models of various objects are constructed, including:
acquiring initial interpolation surface node information and initial interpolation grid units of each target;
aiming at each target belonging to a flying target, constructing a geometric model of the target by using a bi-quadric surface interpolation method based on the obtained interpolation surface node information of the target and the interpolation grid unit;
aiming at each target belonging to the water body target, based on the obtained interpolation surface node information of the target and the interpolation grid unit, a plane structure and an edge structure on a target shell are constructed by using a non-uniform triangular grid division method, and a preset fine structure on the target shell is constructed by using a cellular grid division method, so that a geometric model of the target is obtained.
Optionally, the respectively constructing a first scattered field model of each target based on the initial modeling model of each target and attribution information of the target belonging to a flying target or a water body target, includes:
aiming at each target belonging to a flying target, based on an initial modeling model of the target, a scattering field of a plane structure and a curved surface structure on a target shell is constructed by utilizing an interpolation surface integral method of a stationary phase method, and a scattering field of a marginal structure on the target shell is constructed by utilizing an interpolation surface diffraction integral method, so that a first scattering field model of the target is obtained;
aiming at each target belonging to a water body target, based on an initial modeling model of the target, a triangular patch optical tracking method is utilized to construct a scattered field of a surface structure on a target shell, a cellular grid differential method is utilized to construct a scattered field of a preset fine structure on the target shell, and a triangular surface diffraction integral method is utilized to construct a scattered field of a seamed edge structure on the target shell, so that a first scattered field model of the target is obtained; wherein, face class structure includes at least: multi-angle surface structure, plane structure, curved surface structure and cavity structure.
Optionally, the network structure of the object classification model includes: the device comprises a feature extraction layer, a full connection layer and a Softmax output layer;
the characteristic extraction layer comprises a plurality of cascaded characteristic extraction units, and each characteristic extraction unit comprises a convolution layer and a pooling layer; in the feature extraction layer, adjacent convolution layers and pooling layers are connected through an activation function;
the fully-connected layer includes one or more.
Optionally, the echo data in the multi-frequency echo database are normalized echo data;
the model application module 602 is specifically configured to include:
carrying out normalization processing on radar echo data;
and inputting the radar echo data after the normalization processing into a target classification model which is trained in advance to obtain a target classification recognition result.
In the device for realizing the classification and identification of the large-scale complex target based on the multi-frequency echo data, which is provided by the embodiment of the invention, the target classification and identification result can be obtained directly based on the original radar echo data without forming a radar image, so that the computing resource is saved, and the identification efficiency is higher; in addition, because the original radar echo data does not lose echo information, the target classification and identification result obtained by the embodiment of the invention is accurate, and compared with the prior art, the false alarm rate can be reduced.
In addition, in the embodiment of the invention, after the second scattered field models of various targets under different frequencies and different parameter combinations are established, the multi-frequency echo database with huge data volume can be directly obtained without actual measurement, and the manpower and material resources required by the actual measurement are saved.
An embodiment of the present invention further provides an electronic device, as shown in fig. 7, including a processor 701, a communication interface 702, a memory 703 and a communication bus 704, where the processor 701, the communication interface 702, and the memory 703 complete mutual communication through the communication bus 704,
a memory 703 for storing a computer program;
the processor 701 is configured to implement the method steps of the method for implementing classification and identification of a large complex target based on multi-frequency echo data when executing the program stored in the memory 703.
The communication bus mentioned in the electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is used to illustrate, but not only one bus or one type of bus.
The communication interface is used for communication between the electronic equipment and other equipment.
The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components.
The invention also provides a computer readable storage medium. In the computer-readable storage medium, a computer program is stored, which, when being executed by a processor, implements the method steps of any of the above-mentioned methods for implementing classification and identification of large complex objects based on multi-frequency echo data.
Alternatively, the computer-readable storage medium may be a Non-Volatile Memory (NVM), such as at least one disk Memory.
Optionally, the computer readable memory may also be at least one memory device located remotely from the processor.
In yet another embodiment of the present invention, a computer program product containing instructions is also provided, which when run on a computer causes the computer to perform the method steps of the above-described method for performing classification and identification of large complex targets based on multi-frequency echo data.
It should be noted that, for the device/electronic apparatus/storage medium/computer program product embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant points, refer to the partial description of the method embodiment.
It should be noted that the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
In the description of the specification, reference to the description of the term "one embodiment", "some embodiments", "an example", "a specific example", or "some examples", etc., means that a particular feature or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples described in this specification can be combined and combined by those skilled in the art.
While the present application has been described in connection with various embodiments, other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed application, from a review of the drawings, the disclosure, and the appended claims.
The foregoing is a more detailed description of the invention in connection with specific preferred embodiments and it is not intended that the invention be limited to these specific details. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (6)
1. A method for realizing classification and identification of large complex targets based on multi-frequency echo data is characterized by comprising the following steps:
receiving radar echo data;
inputting the radar echo data into a target classification model which is trained in advance to obtain a target classification recognition result;
the target classification model is obtained by training based on a multi-frequency echo database and a target classification label corresponding to each echo data in the multi-frequency echo database, wherein each echo data corresponds to one frequency in the multiple frequencies.
2. The method of claim 1, wherein the multi-frequency echo database is constructed by:
constructing geometric models of various targets;
aiming at each target, respectively constructing corresponding dielectric parameter models for all surface elements included in the geometric model of the target based on the material and the structure of the target shell to obtain an initial modeling model of the target;
respectively constructing a first scattered field model of each target based on the initial modeling model of each target;
aiming at each target, analyzing and determining a plurality of sensitive parameter combinations of the first scattered field model of the target by using an orthogonal analysis method, and substituting each sensitive parameter combination into the first scattered field model of the target respectively to obtain a plurality of second scattered field models of the target; the sensitive parameter combination is a combination of target parameters and/or radar parameters, wherein the influence on the target echo amplitude exceeds a threshold value, and the target parameters are parameters of a target;
and respectively substituting various different radar working frequencies into each second scattered field model to obtain the multi-frequency echo database.
3. The method of claim 2, wherein said constructing a geometric model of the various objects comprises:
acquiring initial interpolation surface node information and initial interpolation grid units of each target;
aiming at each target belonging to a flying target, constructing a geometric model of the target by using a bi-quadric surface interpolation method based on the obtained interpolation surface node information of the target and the interpolation grid unit;
aiming at each target belonging to the water body target, based on the obtained interpolation surface node information of the target and the interpolation grid unit, a plane structure and an edge structure on a target shell are constructed by using a non-uniform triangular grid division method, and a preset fine structure on the target shell is constructed by using a cellular grid division method, so that a geometric model of the target is obtained.
4. The method of claim 3, wherein separately constructing a first model of the scattered field for each target based on the initial modeling model for each target comprises:
aiming at each target belonging to a flying target, based on an initial modeling model of the target, a scattering field of a plane structure and a scattering field of a curved surface structure on a target shell are constructed by utilizing an interpolation surface integral method of a stationary phase method, and a scattering field of a marginal structure on the target shell is constructed by utilizing an interpolation surface diffraction integral method, so that a first scattering field model of the target is obtained;
aiming at each target belonging to a water body target, based on an initial modeling model of the target, a triangular patch optical tracking method is utilized to construct a scattered field of a surface structure on a target shell, a cellular grid differential method is utilized to construct a scattered field of a preset fine structure on the target shell, and a triangular surface diffraction integral method is utilized to construct a scattered field of a seamed edge structure on the target shell, so that a first scattered field model of the target is obtained; wherein, the face class structure includes at least: multi-angle surface structure, plane structure, curved surface structure and cavity structure.
5. The method of claim 1, wherein the network structure of the object classification model comprises: the device comprises a feature extraction layer, a full connection layer and a Softmax output layer;
the characteristic extraction layer comprises a plurality of cascaded characteristic extraction units, and each characteristic extraction unit comprises a convolution layer and a pooling layer; in the feature extraction layer, adjacent convolution layers and pooling layers are connected through an activation function;
the fully-connected layer includes one or more.
6. The method of claim 5, wherein the echo data in the multi-frequency echo database are normalized echo data;
the step of inputting the radar echo data into a pre-trained target classification model to obtain a target classification recognition result comprises the following steps:
carrying out normalization processing on the radar echo data;
and inputting the radar echo data after the normalization processing into a target classification model which is trained in advance to obtain a target classification recognition result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010845657.3A CN114076924A (en) | 2020-08-20 | 2020-08-20 | Method for realizing classification and identification of large complex target based on multi-frequency echo data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010845657.3A CN114076924A (en) | 2020-08-20 | 2020-08-20 | Method for realizing classification and identification of large complex target based on multi-frequency echo data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114076924A true CN114076924A (en) | 2022-02-22 |
Family
ID=80282345
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010845657.3A Pending CN114076924A (en) | 2020-08-20 | 2020-08-20 | Method for realizing classification and identification of large complex target based on multi-frequency echo data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114076924A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101738603A (en) * | 2009-07-02 | 2010-06-16 | 北京理工大学 | Method for simulating verisimilar target echo |
| CN103018729A (en) * | 2012-11-23 | 2013-04-03 | 北京航空航天大学 | Method for calculating radar scattering cross section of metal cylindrical calibration body |
| CN107632291A (en) * | 2017-08-18 | 2018-01-26 | 上海无线电设备研究所 | A kind of method based on the anti-corner reflector interference of polarimetric radar |
| CN107886121A (en) * | 2017-11-03 | 2018-04-06 | 北京清瑞维航技术发展有限公司 | Target identification method, apparatus and system based on multiband radar |
| CN109239684A (en) * | 2018-10-15 | 2019-01-18 | 中国航天科工集团八五研究所 | Radar target simulation system and its analogy method based on EM scatter model |
| CN109901159A (en) * | 2019-04-04 | 2019-06-18 | 哈尔滨理工大学 | A kind of metre wave radar target identification method based on deep learning |
| CN110967678A (en) * | 2019-12-20 | 2020-04-07 | 安徽博微长安电子有限公司 | Data fusion algorithm and system for multiband radar target identification |
-
2020
- 2020-08-20 CN CN202010845657.3A patent/CN114076924A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101738603A (en) * | 2009-07-02 | 2010-06-16 | 北京理工大学 | Method for simulating verisimilar target echo |
| CN103018729A (en) * | 2012-11-23 | 2013-04-03 | 北京航空航天大学 | Method for calculating radar scattering cross section of metal cylindrical calibration body |
| CN107632291A (en) * | 2017-08-18 | 2018-01-26 | 上海无线电设备研究所 | A kind of method based on the anti-corner reflector interference of polarimetric radar |
| CN107886121A (en) * | 2017-11-03 | 2018-04-06 | 北京清瑞维航技术发展有限公司 | Target identification method, apparatus and system based on multiband radar |
| CN109239684A (en) * | 2018-10-15 | 2019-01-18 | 中国航天科工集团八五研究所 | Radar target simulation system and its analogy method based on EM scatter model |
| CN109901159A (en) * | 2019-04-04 | 2019-06-18 | 哈尔滨理工大学 | A kind of metre wave radar target identification method based on deep learning |
| CN110967678A (en) * | 2019-12-20 | 2020-04-07 | 安徽博微长安电子有限公司 | Data fusion algorithm and system for multiband radar target identification |
Non-Patent Citations (5)
| Title |
|---|
| DU W W, ET AL.: "Modeling and simulation of radar echo signal of aircraft targets with GRECO", 《2008 8TH INTERNATIONAL SYMPOSIUM ON ANTENNAS, PROPAGATION AND EM THEORY》, 31 December 2008 (2008-12-31), pages 859 - 862 * |
| XI LUO, ET AL: "A Novel 2-D Geometry Reconstruction Approach for Space Debris via Interpolation-Free Operation under Low SNR Conditions", 《REMOTE SENSING》, 26 June 2020 (2020-06-26), pages 1 - 17 * |
| 赵慧波等: "支持向量机的 OTHR 多频特征目标分类识别法", 《火力与指挥控制》, 29 February 2012 (2012-02-29), pages 16 - 19 * |
| 郭立新等: "目标与复杂地海面复合电磁散射研究现状综述", 《电波科学学报》, 29 February 2020 (2020-02-29), pages 69 - 84 * |
| 陆洪: "基于多频和极点特征的高频超视距雷达目标识别", 《中国优秀硕士学位论文全文数据库 信息科技辑》, 15 August 2005 (2005-08-15) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pérez et al. | Analysis of antennas on board arbitrary structures modeled by NURBS surfaces | |
| Wen et al. | Deep learning based smart radar vision system for object recognition | |
| CN110362877B (en) | Environmental electromagnetic scattering characteristic analysis method for uncertain factors | |
| CN117332532B (en) | Method and device for constructing axial flow impeller mechanical blade model | |
| CN114076924A (en) | Method for realizing classification and identification of large complex target based on multi-frequency echo data | |
| CN116401498A (en) | Electromagnetic scattering analysis method, equipment and medium based on MB-CRWG (China-microwave broadcasting) basis function | |
| CN115700760A (en) | Multi-mode data-based total-space laser radar scattering cross section calculation method | |
| CN111832183B (en) | Electromagnetic simulation method based on octree structure MLFMA region decomposition | |
| Bilal et al. | Multi‐frequency analysis of Gaussian process modelling for aperiodic RCS responses of a parameterised aircraft model | |
| Mansukhani et al. | Rcs based target classification using deep learning methods | |
| CN115659607A (en) | Method for determining multilayer wave-transparent structure | |
| CN114910892A (en) | Laser radar calibration method and device, electronic equipment and storage medium | |
| US11280899B2 (en) | Target recognition from SAR data using range profiles and a long short-term memory (LSTM) network | |
| Zhuang et al. | Image processing and deep normalized CNN for the location measurement and reading distance prediction of RFID multi-tags | |
| CN111368398A (en) | Method and device for analyzing electromagnetic scattering characteristics of large electric target with uncertain structure | |
| CN116092072B (en) | Spacecraft target detection method, spacecraft target detection system, storage medium and electronic equipment | |
| CN114076923B (en) | Target recognition method based on time domain echo model of multilayer background and target | |
| He et al. | Comparison of satellite target recognition methods based on ISAR images | |
| CN107884754A (en) | Radar scattering face effective area detection means and method | |
| Delgado et al. | Fast monostatic RCS computation using the near-field sparse approximate inverse and the multilevel fast multipole algorithm | |
| CN115629552B (en) | Method and device for checking main target identification all-link model of radio frequency detection system | |
| CN111414801B (en) | Classification and identification method for electrically large non-cooperative target with uncertain shape | |
| CN116522781B (en) | Sample data generation method, model training method and device | |
| Santosa et al. | Safe distance reminder system on ship against port for the standing process using image processing | |
| Li et al. | Parallelized Multilevel Characteristic Basis function Method (MLCBFM) combined with Adaptive Cross Approximation (ACA) for the analysis of the scattering from electrically large rough surfaces |
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 |