Aerial target classification and identification method based on RCS feature distribution

Abstract

The invention provides an aerial target classification and recognition method based on RCS feature distribution, and belongs to the technical field of target recognition. The method comprises the following steps: locking an air target by using a radar system, setting radar parameters, and tracking the air target in real time to collect real-time RCS data; preprocessing and sorting the RCS data, extracting a plurality of feature parameters from the RCS data, and forming a feature vector by the plurality of feature parameters to form RCS feature description; confidence interval calculation is carried out on the extracted feature parameters, and features with non-overlapped confidence intervals are selected as a basis for target identification; generating RCS data of different types of air targets at various angles, postures and frequencies so as to establish an RCS simulation library; and matching the RCS feature vector acquired in real time with data in the RCS simulation library, and outputting an identified target type and a related confidence analysis result.

Description

Technical Field

[0001] The invention belongs to the technical field of target recognition, and in particular relates to an aerial target classification and recognition method based on RCS feature distribution. Background Art

[0002] With the development and application of aviation technology, the types of aerial targets are increasing day by day. Identifying and classifying these targets has become an important task in aviation safety and defense applications. Traditional target recognition methods usually rely on radar cross-sectional area (RCS) characteristics. RCS is an important physical quantity that describes the scattering characteristics of the target to the radar signal and is usually used for target detection and recognition in radar systems. However, the existing technology has some defects and deficiencies in processing the RCS characteristics of aerial targets:

[0003] 1. Low recognition accuracy: Existing recognition methods rely on static RCS features, which change significantly under different observation angles, frequencies and target postures, resulting in low recognition accuracy, especially in complex environments.

[0004] 2. Feature confusion: Some RCS features cannot be effectively distinguished between different targets, and may even be confused, resulting in reduced recognition accuracy. This phenomenon is particularly evident when feature confidence intervals overlap.

[0005] 3. Limited recognition capabilities: Traditional RCS libraries may not be comprehensive enough and lack simulation data support for different targets under various conditions, which limits the recognition system's ability to recognize new and complex targets. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention proposes an aerial target classification and identification scheme based on RCS feature distribution.

[0007] The first aspect of the present invention discloses a method for classifying and identifying aerial targets based on RCS feature distribution, the method comprising:

[0008] Step S1, using the radar system to lock the aerial target, setting radar parameters, and tracking the aerial target in real time to collect real-time RCS data;

[0009] Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description;

[0010] Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition;

[0011] Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library;

[0012] Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

[0013] According to the method of the first aspect, in step S1:

[0014] Using a radar system to lock onto the aerial target and determine the initial position and flight direction of the aerial target;

[0015] Radar parameter settings include: setting the radar's scanning frequency, beam width, transmit power, etc., so that the radar can capture the target's RCS data at multiple angles and frequencies;

[0016] Starting the radar system to track the aerial target in real time and capture its RCS data at different angles and frequencies;

[0017] The RCS data is stored in the form of a time series, and each data point corresponds to an RCS value of an observation angle and frequency.

[0018] According to the method of the first aspect, in step S2:

[0019] Preprocess the collected RCS data, including denoising, data filtering, and interpolation; arrange the preprocessed RCS data in time series;

[0020] Extract multiple characteristic parameters from RCS data; including:

[0021] Position characteristics: mean, median, maximum, minimum;

[0022] Dispersion characteristics: range, variance, standard deviation, coefficient of variation;

[0023] Distribution characteristics: skewness coefficient, kurtosis coefficient;

[0024] Transformation features: frequency domain features after Fourier transformation, features after wavelet transformation;

[0025] The extracted multiple feature parameters are combined into a feature vector to form the RCS feature description of the aerial target for subsequent target recognition.

[0026] According to the method of the first aspect, in step S3:

[0027] Perform statistical analysis on each extracted characteristic parameter and calculate its distribution under different observation conditions; calculate the confidence interval for each characteristic value, using a 95% confidence level;

[0028] The corresponding confidence interval is calculated for each eigenvalue according to its distribution characteristics. The eigenvalues ​​at different angles and frequencies reflect the characteristics of the aerial target.

[0029] Among multiple feature parameters, features with non-overlapping confidence intervals are selected as the basis for target recognition to distinguish different types of aerial targets.

[0030] According to the method of the first aspect, in step S4:

[0031] Generate RCS simulation data of different types of aerial targets at various angles, attitudes and frequencies using electromagnetic simulation software or experimental measurement; the RCS simulation data covers various observation conditions, thereby establishing an RCS simulation library;

[0032] The generated RCS simulation data are classified and stored in the database according to target type, observation conditions, and characteristic parameters; the RCS simulation library is updated regularly.

[0033] According to the method of the first aspect, in step S5:

[0034] Match the RCS feature vector obtained in real time with the data in the RCS simulation library, and use the Mahalanobis distance of the feature vector or the matching algorithm to calculate the matching degree; based on the matching degree, filter out the simulation data entry that is closest to the real-time data;

[0035] According to the matching results, the target is classified and identified to determine the specific type of the target; the classification and identification process is based on the feature that the confidence intervals do not overlap;

[0036] Compare the recognition result with the actual situation to verify the accuracy of the recognition; if the recognition result meets the expectation, output the final classification result; otherwise, adjust the parameters and re-recognize;

[0037] Output the final identified target type and related confidence analysis results.

[0038] A second aspect of the present invention discloses an aerial target classification and identification system based on RCS feature distribution, the system comprising a processing unit, the processing unit being configured to execute:

[0039] Step S1, obtaining the collected real-time RCS data; wherein: calling the radar system to lock the air target, setting the radar parameters, the radar system is used to track the air target in real time, thereby obtaining the real-time RCS data

[0040] Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description;

[0041] Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition;

[0042] Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library;

[0043] Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

[0044] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in the first aspect of the present disclosure are implemented.

[0045] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in the first aspect of the present disclosure are implemented.

[0046] In summary, the beneficial technical effects brought by the present invention include:

[0047] 1. High recognition accuracy: By analyzing the dynamic distribution of the target RCS characteristics, the accuracy of target recognition is improved, especially in complex environments.

[0048] 2. Strong recognition capability: The introduction of multi-angle and multi-frequency RCS matching methods enhances the recognition capability of various aerial targets.

[0049] 3. Good flexibility: This method can adapt to new target types by continuously updating the RCS simulation library and has strong scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0051] Figure 1 The figure is a flow chart of a method for classifying and identifying aerial targets based on RCS feature distribution according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0053] The present invention proposes a classification and identification scheme for aerial targets based on RCS feature distribution. The scheme mainly includes:

[0054] 1. Dynamic RCS feature analysis: This invention generates the target's RCS feature distribution map by dynamically analyzing the target's multi-angle and multi-frequency RCS data, thereby improving recognition accuracy. This dynamic analysis overcomes the limitations of static features in traditional methods, making the recognition results more accurate.

[0055] 2. Establishment and matching of RCS simulation library: The present invention creates a comprehensive RCS simulation library, which contains RCS values ​​of different targets at various angles, attitudes and frequencies. By matching with the real-time observed RCS data, the present invention can quickly and accurately identify the target type.

[0056] 3. Confidence analysis: The confidence analysis method is introduced to calculate the confidence interval of the RCS feature, and by screening features with non-overlapping confidence intervals, feature confusion is reduced, thereby enhancing the accuracy of target recognition.

[0057] The present invention effectively solves the problems of low recognition accuracy, limited recognition capability and incomplete RCS library in the prior art, and significantly improves the accuracy and reliability of aerial target recognition. The present invention aims to improve the classification and recognition accuracy and capability of aerial targets, reduce feature confusion by dynamically analyzing the RCS features of the targets and combining confidence interval analysis, and establish a simulation library containing multiple target RCS features to improve the accuracy and reliability of recognition.

[0058] The first aspect of the present invention discloses a method for classifying and identifying aerial targets based on RCS feature distribution, the method comprising (eg Figure 1 shown):

[0059] Step S1, using the radar system to lock the aerial target, setting radar parameters, and tracking the aerial target in real time to collect real-time RCS data;

[0060] Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description;

[0061] Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition;

[0062] Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library;

[0063] Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

[0064] According to the method of the first aspect, in step S1:

[0065] Using a radar system to lock onto the aerial target and determine the initial position and flight direction of the aerial target;

[0066] Radar parameter settings include: setting the radar's scanning frequency, beam width, transmit power, etc., so that the radar can capture the target's RCS data at multiple angles and frequencies;

[0067] Starting the radar system to track the aerial target in real time and capture its RCS data at different angles and frequencies;

[0068] The RCS data is stored in the form of a time series, and each data point corresponds to an RCS value of an observation angle and frequency.

[0069] According to the method of the first aspect, in step S2:

[0070] Preprocess the collected RCS data, including denoising, data filtering, and interpolation; arrange the preprocessed RCS data in time series;

[0071] Extract multiple characteristic parameters from RCS data; including:

[0072] Position characteristics: mean, median, maximum, minimum;

[0073] Dispersion characteristics: range, variance, standard deviation, coefficient of variation;

[0074] Distribution characteristics: skewness coefficient, kurtosis coefficient;

[0075] Transformation features: frequency domain features after Fourier transformation, features after wavelet transformation;

[0076] The extracted multiple feature parameters are combined into a feature vector to form the RCS feature description of the aerial target for subsequent target recognition.

[0077] According to the method of the first aspect, in step S3:

[0078] Perform statistical analysis on each extracted characteristic parameter and calculate its distribution under different observation conditions; calculate the confidence interval for each characteristic value, using a 95% confidence level;

[0079] The corresponding confidence interval is calculated for each eigenvalue according to its distribution characteristics. The eigenvalues ​​at different angles and frequencies reflect the characteristics of the aerial target.

[0080] Among multiple feature parameters, features with non-overlapping confidence intervals are selected as the basis for target recognition to distinguish different types of aerial targets.

[0081] According to the method of the first aspect, in step S4:

[0082] Generate RCS simulation data of different types of aerial targets at various angles, attitudes and frequencies using electromagnetic simulation software or experimental measurement; the RCS simulation data covers various observation conditions, thereby establishing an RCS simulation library;

[0083] The generated RCS simulation data are classified and stored in the database according to target type, observation conditions, and characteristic parameters; the RCS simulation library is updated regularly.

[0084] According to the method of the first aspect, in step S5:

[0085] Match the RCS feature vector obtained in real time with the data in the RCS simulation library, and use the Mahalanobis distance of the feature vector or the matching algorithm to calculate the matching degree; based on the matching degree, filter out the simulation data entry that is closest to the real-time data;

[0086] According to the matching results, the target is classified and identified to determine the specific type of the target; the classification and identification process is based on the feature that the confidence intervals do not overlap;

[0087] Compare the recognition result with the actual situation to verify the accuracy of the recognition; if the recognition result meets the expectation, output the final classification result; otherwise, adjust the parameters and re-recognize;

[0088] Output the final identified target type and related confidence analysis results.

[0089] First embodiment

[0090] This embodiment provides an aerial target classification and recognition method based on RCS feature distribution, which mainly includes:

[0091] Step 1: Target RCS feature distribution analysis

[0092] In the radar system, the multi-angle and multi-frequency RCS characteristic data of the target are captured and analyzed to generate the target's RCS characteristic distribution map. These data can be obtained through experimental measurement or simulation.

[0093] Step 2: Build the target RCS simulation library

[0094] Based on the actual data of different types of aerial targets, a comprehensive RCS simulation library is established, which records the RCS values ​​of different targets at various angles, attitudes and frequencies.

[0095] Step 3: Confidence Analysis

[0096] The confidence intervals of the RCS features are calculated and analyzed, and those features whose confidence intervals do not overlap are identified as the basis for target recognition.

[0097] Step 4: Target Identification and Classification

[0098] Based on the RCS data acquired in real time, the best matching features are determined by matching with the simulation library and confidence analysis to complete the classification and identification of the target.

[0099] Second embodiment

[0100] In practical applications, the radar system is first used to obtain the real-time RCS data of the target. Then, feature analysis is performed based on the acquired data and matched with the data in the pre-established RCS simulation library. The confidence analysis method is used to identify features with non-overlapping confidence intervals, and these features are used as the basis for target recognition.

[0101] This embodiment provides an aerial target classification and recognition method based on RCS feature distribution. The method mainly includes the following steps: data collection, feature extraction, confidence interval analysis, establishment of RCS simulation library, and target matching and classification. The following is a detailed implementation process of each step.

[0102] 1. Data Collection

[0103] Step 1.1: Target lock and radar parameter setting

[0104] First, use the radar system to lock onto the aerial target and determine the target's initial position and flight direction.

[0105] Set the radar's scanning frequency, beam width, transmit power and other parameters to ensure that the radar can capture the target's RCS data at multiple angles and frequencies.

[0106] Step 1.2: Real-time RCS data acquisition

[0107] Start the radar system to track the target in real time and capture its RCS data at different angles and frequencies.

[0108] RCS data are usually stored in the form of time series, and each data point corresponds to an RCS value of an observation angle and frequency.

[0109] 2. Feature extraction

[0110] Step 2.1: Data preprocessing

[0111] The collected RCS data are preprocessed, including denoising, data filtering, interpolation, etc., to reduce the impact of noise on subsequent analysis.

[0112] Arrange the processed RCS data in time series to prepare for feature extraction.

[0113] Step 2.2: Characteristic parameter calculation

[0114] Extract multiple feature parameters from the preprocessed RCS data. These feature parameters include but are not limited to:

[0115] Position characteristics: such as mean, median, maximum and minimum.

[0116] Dispersion characteristics: such as range, variance, standard deviation, and coefficient of variation.

[0117] Distribution characteristics: such as skewness coefficient and kurtosis coefficient.

[0118] Transformation features: such as frequency domain features after Fourier transform, features after wavelet transform, etc.

[0119] Step 2.3: Feature vector generation

[0120] The extracted multiple feature parameters are combined into feature vectors to form the RCS feature description of the target. These feature vectors will be used for subsequent target recognition.

[0121] 3. Confidence Interval Analysis

[0122] Step 3.1: Eigenvalue statistics

[0123] Statistical analysis is performed on each extracted feature parameter to calculate its distribution under different observation conditions.

[0124] A confidence interval is calculated for each eigenvalue, typically with a 95% confidence level.

[0125] Step 3.2: Confidence interval calculation

[0126] For each eigenvalue, the corresponding confidence interval is calculated based on its distribution characteristics to ensure that the eigenvalues ​​at different angles and frequencies can correctly reflect the characteristics of the target.

[0127] Step 3.3: Feature Selection

[0128] Among multiple feature parameters, features with non-overlapping confidence intervals are selected as the basis for target recognition. These features can effectively distinguish different types of aerial targets.

[0129] 4. Establish RCS simulation library

[0130] Step 4.1: Simulation data generation

[0131] Use electromagnetic simulation software or experimental measurement methods to generate RCS data of different types of aerial targets at various angles, attitudes and frequencies.

[0132] Ensure that the simulation data covers a wide range of observation conditions to build a comprehensive RCS simulation library.

[0133] Step 4.2: Data storage and management

[0134] The generated RCS simulation data are classified and stored in the database according to target type, observation conditions, characteristic parameters, etc.

[0135] The simulation library is updated regularly to ensure its adaptability to new targets and different environments.

[0136] 5. Target matching and classification

[0137] Step 5.1: Matching real-time data with simulation data

[0138] The RCS feature vector obtained in real time is matched with the data in the RCS simulation library, and the matching degree is calculated using the Mahalanobis distance of the feature vector or other matching algorithms.

[0139] Based on the matching degree, the simulation data entries that are closest to the real-time data are selected.

[0140] Step 5.2: Target Classification

[0141] According to the matching results, the target is classified and identified to determine the specific type of the target. The classification process is based on the feature that the confidence intervals do not overlap, thereby improving the accuracy of identification.

[0142] Step 5.3: Result verification and output

[0143] Compare the recognition result with the actual situation to verify the accuracy of the recognition. If the recognition result meets the expectation, the final classification result is output; otherwise, adjust the parameters and re-recognize.

[0144] Output the final identified target type and related confidence analysis results.

[0145] A second aspect of the present invention discloses an aerial target classification and identification system based on RCS feature distribution, the system comprising a processing unit, the processing unit being configured to execute:

[0146] Step S1, obtaining the collected real-time RCS data; wherein: calling the radar system to lock the air target, setting the radar parameters, the radar system is used to track the air target in real time, thereby obtaining the real-time RCS data

[0147] Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description;

[0148] Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition;

[0149] Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library;

[0150] Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

[0151] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in the first aspect of the present disclosure are implemented.

[0152] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in the first aspect of the present disclosure are implemented.

[0153] In summary, the beneficial technical effects brought by the present invention include:

[0154] 1. High recognition accuracy: By analyzing the dynamic distribution of the target RCS characteristics, the accuracy of target recognition is improved, especially in complex environments.

[0155] 2. Strong recognition capability: The introduction of multi-angle and multi-frequency RCS matching methods enhances the recognition capability of various aerial targets.

[0156] 3. Good flexibility: This method can adapt to new target types by continuously updating the RCS simulation library and has strong scalability.

[0157] Please note that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above-mentioned embodiments only express several implementation methods of the present application, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present application, several variations and improvements can be made, which all belong to the scope of protection of the present application. Therefore, the scope of protection of the patent in this application shall be based on the attached claims.

Claims

1. A method for classifying and identifying aerial targets based on RCS feature distribution, characterized in that: The method comprises: Step S1, using the radar system to lock the aerial target, setting radar parameters, and tracking the aerial target in real time to collect real-time RCS data; Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description; Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition; Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library; Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

2. The method for classifying and identifying aerial targets based on RCS feature distribution according to claim 1 is characterized in that: In step S1: Using a radar system to lock onto the aerial target and determine the initial position and flight direction of the aerial target; Radar parameter settings include: setting the radar's scanning frequency, beam width, transmit power, etc., so that the radar can capture the target's RCS data at multiple angles and frequencies; Starting the radar system to track the aerial target in real time and capture its RCS data at different angles and frequencies; The RCS data is stored in the form of a time series, and each data point corresponds to an RCS value of an observation angle and frequency.

3. The method for classifying and identifying aerial targets based on RCS feature distribution according to claim 2 is characterized in that: In step S2: Preprocess the collected RCS data, including denoising, data filtering, and interpolation; Arrange the preprocessed RCS data in time series; Extract multiple characteristic parameters from RCS data; including: Position characteristics: mean, median, maximum, minimum; Dispersion characteristics: range, variance, standard deviation, coefficient of variation; Distribution characteristics: skewness coefficient, kurtosis coefficient; Transformation features: frequency domain features after Fourier transformation, features after wavelet transformation; The extracted multiple feature parameters are combined into a feature vector to form the RCS feature description of the aerial target for subsequent target recognition.

4. The method for classifying and identifying aerial targets based on RCS feature distribution according to claim 3 is characterized in that: In step S3: Perform statistical analysis on each extracted characteristic parameter and calculate its distribution under different observation conditions; calculate the confidence interval for each characteristic value, using a 95% confidence level; The corresponding confidence interval is calculated for each eigenvalue according to its distribution characteristics. The eigenvalues ​​at different angles and frequencies reflect the characteristics of the aerial target. Among multiple feature parameters, features with non-overlapping confidence intervals are selected as the basis for target recognition to distinguish different types of aerial targets.

5. The method for classifying and identifying aerial targets based on RCS feature distribution according to claim 4 is characterized in that: In step S4: Generate RCS simulation data of different types of aerial targets at various angles, attitudes and frequencies using electromagnetic simulation software or experimental measurement; the RCS simulation data covers various observation conditions, thereby establishing an RCS simulation library; The generated RCS simulation data are classified and stored in the database according to target type, observation conditions, and characteristic parameters; the RCS simulation library is updated regularly.

6. The method for classifying and identifying aerial targets based on RCS feature distribution according to claim 5 is characterized in that: In step S5: Match the RCS feature vector obtained in real time with the data in the RCS simulation library, and use the Mahalanobis distance of the feature vector or the matching algorithm to calculate the matching degree; based on the matching degree, filter out the simulation data entry that is closest to the real-time data; According to the matching results, the target is classified and identified to determine the specific type of the target; The classification recognition process is based on features whose confidence intervals do not overlap; Compare the recognition result with the actual situation to verify the accuracy of the recognition; if the recognition result meets the expectation, output the final classification result; otherwise, adjust the parameters and re-recognize; Output the final identified target type and related confidence analysis results.

7. An aerial target classification and recognition system based on RCS feature distribution, characterized in that: The system comprises a processing unit configured to perform: Step S1, obtaining the collected real-time RCS data; wherein: calling the radar system to lock the air target, setting the radar parameters, the radar system is used to track the air target in real time, thereby obtaining the real-time RCS data Step S2, preprocessing and sorting the RCS data, and extracting multiple feature parameters therefrom, and forming a feature vector with the multiple feature parameters to form an RCS feature description; Step S3, calculating the confidence interval of the extracted feature parameters, and selecting features whose confidence intervals do not overlap as the basis for target recognition; Step S4, generating RCS data of different types of aerial targets at various angles, attitudes and frequencies to establish an RCS simulation library; Step S5: Match the RCS feature vector acquired in real time with the data in the RCS simulation library, and output the identified target type and related confidence analysis results.

8. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps in the method for classifying and identifying aerial targets based on RCS feature distribution described in any one of claims 1 to 6 are implemented.

Images