Radar ground-sea clutter environment multi-feature identification method

Abstract

The invention relates to a radar ground-sea clutter environment multi-feature identification method, and belongs to the technical field of radar signal processing. According to the method, firstly, feature extraction is carried out on each data sequence, and normalization is completed; forming a feature vector by four feature values of each resolution unit, sending the feature vector to a random forest classifier to complete judgment, and setting the resolution unit which is judged to be sea clutter as 0; and the resolution unit for determining the ground clutter is set to be 1. And converting a 0 / 1 matrix obtained according to the judgment result into a grey-scale map, and taking the grey-scale map as a primary identification result. In order to improve the identification result, closed operation and hole filling are sequentially used for processing the primary identification result, the identification effect is improved, and a secondary processing result is obtained. The method has higher robustness and accuracy, and the extraction of the potential target area is beneficial to improving the target detection efficiency.

Description

Technical Field

[0001] The invention relates to a radar land and sea clutter environment multi-feature identification method, belonging to the technical field of radar signal processing. Background Art

[0002] There are many types of interference problems such as ground and sea clutter in the detection of nearshore sea targets. The ground clutter generated by islands and land will cause patches or discrete strong clutter points, which seriously affect the subsequent detection effect. Therefore, it is of great significance to separate ground and sea clutter in the early stage of target detection. After completing the land and sea identification, the ground clutter interference is reduced, and the detection work is focused on the identified sea area, which helps to complete the target detection and clutter analysis work more efficiently in the future, and obtain more accurate detection results.

[0003] At present, many land and sea identification schemes have been proposed based on radar echo data. Xia Xiaoyun et al. proposed a phase-based land and sea clutter identification method for shore-based radars. This method uses the phase change between adjacent pulses of the same unit to characterize the differences between land and sea clutter. Deng Jiaqi et al. used a heuristic identification algorithm for the clutter amplitude mean statistic of radar echo data, thereby globally separating the stable echo areas in each region in the distance dimension, and distinguishing the land and sea clutter areas according to the proportion of the DC component. Although the two methods have good identification effects, as single-measure identification algorithms, they both have potential feature limitations and are easily affected by multiple factors such as sea conditions, platform conditions, and radar systems. Bandt et al. proposed permutation entropy in 2002, which is used to describe the complexity of time series or chaotic dynamic systems. Permutation entropy shows high sensitivity to time, has strong anti-noise performance, and can better detect the mutation information of complex nonlinear systems. Therefore, permutation entropy has a certain characterization ability for the differences in the non-stationary characteristics of land and sea clutter.

[0004] In summary, with the increase of data volume and resolution, higher requirements are placed on the speed of land-sea separation algorithm. At the same time, it is also necessary to combine different features to provide information at different levels and angles to cope with more complex measurement environments. Therefore, it is necessary to propose a more accurate and faster land-sea clutter identification method. Summary of the invention

[0005] The purpose of the present invention is to solve the deficiencies of the above-mentioned prior art and to propose a radar land-sea clutter environment multi-feature identification method, which has higher robustness and accuracy, and the extraction of potential target areas helps to improve target detection efficiency.

[0006] The radar land-sea clutter environment multi-feature identification method is special in that it includes the following steps:

[0007] Step 1: Extract the echo data of 13 adjacent pulses of the resolution unit to be detected in the same distance unit to form a data sequence. The sequence is analyzed to extract the DC component ratio feature, and then the spectrum kurtosis feature is extracted by spectrum analysis. The sequence is modulo-calculated to calculate the permutation entropy feature, and the phase quadratic fitting feature is calculated after extracting the sequence phase information to complete the feature extraction of the data to be detected;

[0008] Step 2: construct a data set based on the features extracted in step 1 and normalize them respectively. Then, the four types of eigenvalues ​​of each resolution unit form a feature vector and send it to the random forest classifier to complete the judgment. The 0 / 1 matrix obtained by the judgment result is transformed into a grayscale image and used as the primary recognition result. The primary recognition result is processed by removing isolated points, closing operations and hole filling in turn to improve the recognition effect and obtain the secondary processing result.

[0009] Step 3: Use the two empirical methods of the number of condensed points and the area width greater than N meters to complete the classification of large and small areas in the secondary processing results obtained in step 2. Extract the amplitude mean of the radar echo modulo along the azimuth dimension in each small area, compress the two-dimensional area into a one-dimensional vector, and the vector length is the same as the number of distance units occupied by the area. By performing a spectral peak search on the one-dimensional vector, the number of peaks and the peak position are obtained. According to the number of peaks, the type of small area is identified and it is judged to be a target or a small island.

[0010] Preferably, the specific steps of extracting the spectrum kurtosis feature in step 1 are:

[0011] The echo data of 13 adjacent pulses of the resolution unit to be detected in the same distance unit are extracted to form a data sequence. A 256-point FFT is performed on each data vector. The spectrum kurtosis of each resolution unit is calculated based on the frequency domain data after FFT. The spectrum kurtosis of all resolution units is obtained in one scanning data of the radar:

[0012]

[0013] Where K is the kurtosis of the data sequence after FFT, x i is the resolution unit to be detected, n is the number of scans to x i The total number of beams, is the average value of the unit to be tested and the adjacent units. The kurtosis of the standard normal distribution is 3. When the peak value of the time series curve is higher than that of the normal distribution, the kurtosis is greater than 3, that is, it presents a peak distribution. When it is lower than that of the normal distribution, the kurtosis is less than 3, that is, the peak value is relatively flat. The peak value has no dimension, and its value range is between 1 and the number of sample data.

[0014] Preferably, the specific steps of extracting the DC component ratio feature in step 1 are:

[0015] The DC power and average power of the data sequence are expressed as formula (2) and (3), respectively denoted as s and p:

[0016]

[0017]

[0018] Where N is the number of pulses, x i is the i-th radar echo data. By calculating the ratio of the two, the DC component ratio can be obtained, which is recorded as:

[0019]

[0020] The DC component ratio ranges from [0,1], the DC component ratio of ground clutter is close to 1, and the DC component ratio of sea clutter is close to 0.

[0021] Preferably, the specific steps of extracting the phase quadratic fitting features in step 1 are:

[0022] Select 13 adjacent echo data under the same distance unit of the detection unit, extract the phase value of each data to form a new sequence, unwrap the phase, and after the phase is unwrapped, use the least squares method to perform quadratic fitting on the new data sequence, and complete feature extraction by calculating the fitting error. Use the least squares method to perform quadratic fitting on the new data sequence, and compare each original phase data with the corresponding fitting value to calculate the residual, that is, the residual is equal to the original phase value minus the fitting phase value. Then, sum the squares of all residuals to obtain the total fitting error, which is the phase quadratic fitting feature;

[0023] Preferably, the specific steps of permutation entropy feature extraction in step 1 are:

[0024] Select N adjacent pulse data under the same distance unit of the detection unit and take the modulus to form a time series {X(1),X(2),...,X(N)}. Use the phase space reconstruction method to obtain the reconstruction matrix:

[0025]

[0026] The time series N is 13, m and t represent the embedding dimension and delay time respectively, the embedding dimension m is 4, the delay time t is 1, each row in the reconstruction matrix is ​​a reconstruction component, a total of K, the reconstruction components are rearranged in ascending order to obtain a new sequence, and the column index of each element position in the new sequence constitutes a set of symbol sequences:

[0027] S(l)={j 1 ,j 2 ,…,j m}, l = 1, 2, ..., k, and k ≤ m! (6)

[0028] There are a total of m! kinds of symbol sequences mapped in the m-dimensional phase space. The probability of occurrence of each symbol sequence is obtained by dividing the number of occurrences of each symbol sequence by the total number of occurrences of m! different symbol sequences, that is, P 1 ,P 2 ,...,P k , according to the probability of occurrence of various symbol sequences, the permutation entropy is obtained:

[0029]

[0030] When P j =1 / m!, H pe (m) obtains the maximum value ln(m!), and normalizes the permutation entropy, that is: 0≤H pe =H pe (m) / ln(m!)≤1.

[0031] Preferably, the specific steps of step 2 are:

[0032] Step 2-1: construct a data set for the four features extracted in step 1 and normalize them respectively to obtain an N*4 data matrix, where N is the number of samples to be detected. The four eigenvalues ​​of each resolution unit form a feature vector and are sent to the random forest classifier to complete the judgment. The resolution unit judged as sea clutter is set to 0; the resolution unit judged as ground clutter is set to 1. The result obtained in this step is the random forest judgment result;

[0033] Step 2-2: A 0 / 1 matrix is ​​formed according to the random forest judgment result obtained in step 2-1, and the matrix is ​​transformed into a grayscale image, and used as the primary identification result. The primary identification result is processed in sequence using isolated point removal, closing operation and hole filling to obtain the optimized post-processing result; wherein the isolated point removal is to delete all connected components with less than P pixels in the primary identification result, wherein the P value can be determined according to the specific situation of the data to be detected, the closing operation is to design a certain structural element X, and the result after removing the isolated points is first expanded by X and then eroded by X, and X also needs to be determined according to the situation of the data to be detected. The result of the above operation is to disconnect the small connections in each area as much as possible to facilitate the subsequent step 3. There is no parameter setting for the hole filling operation, and it is directly used for the result after the closing operation.

[0034] Preferably, the specific steps of step 3 are:

[0035] First, use the two empirical methods of condensed point number and area width greater than N meters to complete the classification of large and small areas in the secondary processing results obtained in step 2. The empirical method of condensed point number refers to extracting the connected areas determined as ground clutter in the secondary processing results one by one, checking the number of sampling points occupied by the connected areas, and treating the areas with a number of sampling points greater than P as land areas. The P value here is an empirical value, which needs to be determined in combination with the actual area size represented by a single sampling point in the echo data to be measured. It is required that P*single sampling point area>conventional small islands. The empirical method of area width greater than N meters is to calculate the center of mass of each connected area for the remaining connected areas after the condensed point number, and use the horizontal coordinate K of the center of mass as the reference to check how many azimuth codes the connected area occupies in the vertical coordinate. At this time, let the number of occupied azimuth codes be J, and the angle interval occupied by each azimuth code be x, then the area width L is:

[0036] L=2Ktan(Jx / 2) (8)

[0037] Thus, the areas with a width greater than N meters are selected as land areas, where the width value N is an empirical value;

[0038] For the remaining small areas after screening in the above steps, the amplitude mean of the radar echo after modulus is extracted along the azimuth dimension, and the two-dimensional area is compressed into a one-dimensional vector. The vector length is the same as the number of distance units occupied by the small area. By performing spectral peak search on the one-dimensional vector, the number and position of peaks are obtained.

[0039] The radar land and sea clutter environment multi-feature identification method of the present invention selects spectrum peaks, DC component proportion, phase quadratic fit and permutation entropy to form feature vectors, uses random forest learning algorithm to classify sea clutter data and land clutter data, forms a primary identification image, and then uses morphological filtering to eliminate isolated points and fill holes in the primary identification results, thereby obtaining a secondary identification image. Since the anchored ships and small islands in the radar echo data (after pulse compression) are similar in feature expression. Therefore, the patent algorithm determines the potential target area and the island area in the identification result, and extracts the location information of the potential target in the potential target area. Experimental results show that the algorithm can obtain accurate land and sea clutter identification results and provide more accurate regional guidance for sea surface target detection. The present invention analyzes the land and sea clutter characteristics, selects four features to input random forest classification, and the classification results are morphologically filtered to ensure regional connectivity. By compressing the two-dimensional area to a one-dimensional vector, the potential target area and the island are distinguished according to the difference in the number of spectrum peaks. The verification of measured data shows that the present invention can effectively identify the land and sea areas, complete the identification of potential target areas and island areas, and provide potential target location information. Compared with other identification methods, the present invention has higher robustness and accuracy, and the extraction of potential target areas helps to improve target detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 :It is the algorithm flow chart of the present invention;

[0041] Figure 2 : It is the process diagram of fitting error extraction;

[0042] Figure 3 : The distribution result diagram of permutation entropy value; (a) in the figure shows the distribution result diagram of permutation entropy value when the embedding dimension is 3 in the same region; (b) shows the distribution result diagram of permutation entropy value when the embedding dimension is 4 in the same region; (c) shows the distribution result diagram of permutation entropy value when the embedding dimension is 5 in the same region;

[0043] Figure 4 : The result diagrams of each optimization step of the random forest recognition result; (a) in the figure shows the recognition result diagram obtained after sending the four eigenvalues ​​into the random forest; (b) shows the image after removing isolated points by performing a closing operation on the random forest recognition result; (c) shows the closing operation diagram of the image after removing noise points; (d) shows the hole filling diagram of the image after the closing operation;

[0044] Figure 5 : The classification result diagram of potential target areas; (a) is the large area screening diagram; (b) is the extraction area diagram with a width of >350 meters; (c) is the extraction area diagram of potential target areas. DETAILED DESCRIPTION

[0045] For better understanding and implementation, a specific embodiment is given below to describe in detail a radar land and sea clutter environment multi-feature identification method of the present invention. The processing flow of the present invention is described in detail below:

[0046] The algorithm flow chart of this embodiment is as follows Figure 1As shown. First, the radar echo data after pulse compression is analyzed, and the features of each data sequence are extracted and normalized. Then, the four eigenvalues ​​of each resolution unit form a feature vector and are sent to the random forest classifier to complete the judgment. The resolution unit judged as sea clutter is set to 0; the resolution unit judged as ground clutter is set to 1. The 0 / 1 matrix obtained by the judgment result is transformed into a grayscale image and used as the primary identification result. In order to improve the identification result, the primary identification result is processed by closing operation and hole filling in turn to improve the identification effect and obtain the secondary processing result. In order to solve the problem of judging the target area and small islands that are greatly affected by the pulse pressure range sidelobe, the patent of the present invention first uses the two empirical standards of the number of condensation points and the area width greater than N meters to screen the large area, and then extracts multiple small areas left after screening. The amplitude mean of the radar echo after modulus is extracted along the azimuth dimension in each small area, and the two-dimensional area is compressed into a one-dimensional vector, and the vector length is the same as the number of distance units occupied by the area. By searching the spectrum peak of the one-dimensional vector, the number of peaks and the peak position are obtained. According to the number of peaks, a series of small areas are identified to determine whether they are targets or small islands. The following is a step-by-step explanation:

[0047] Step 1 (ground and sea clutter feature extraction):

[0048] (1) Spectral kurtosis feature extraction

[0049] The radar beam is petal-shaped. In actual work, the data of a certain resolution unit is often collected at the maximum gain point. Due to the influence of the main lobe beam width, the non-maximum gain point of the main lobe also carries the information of adjacent resolution units, that is, the same resolution unit can be scanned by multiple adjacent beams. Therefore, combined with the measured data used in this patent, this patent chooses to extract the echo data of 13 adjacent pulses of the resolution unit to be detected at the same distance unit to form a data sequence. Perform a 256-point FFT on each data vector, and calculate the spectral kurtosis of each resolution unit based on the frequency domain data after FFT. Through this extraction method, the spectral kurtosis of all resolution units can be obtained in one scan data of the radar.

[0050]

[0051] Where K is the kurtosis of the data sequence after FFT, x i is the resolution unit to be detected, n is the number of scans to x i The total number of beams, is the average value of the unit to be tested and the adjacent units. The kurtosis of the standard normal distribution is 3. When the peak value of the time series curve is higher than that of the normal distribution, the kurtosis is greater than 3, that is, it presents a peak distribution; when it is lower than that of the normal distribution, the kurtosis is less than 3, that is, the peak value is relatively flat. The peak value has no dimension and its value range is between 1 and the number of sample data.

[0052] (2) Feature extraction of DC component ratio

[0053] The proportion of the DC component of the signal can be obtained by the ratio of the DC power to the average power. The DC power and average power expressions of the data sequence are as shown in equations (2) and (3), denoted as s and p respectively:

[0054]

[0055]

[0056] Where N is the number of pulses, x i is the i-th radar echo data. By calculating the ratio of the two, the proportion of the DC component can be obtained, which is recorded as:

[0057]

[0058] The DC component ratio ranges from [0,1], the DC component ratio of ground clutter is close to 1, and the DC component ratio of sea clutter is close to 0.

[0059] (3) Phase quadratic fit feature extraction

[0060] Select 13 adjacent echo data under the same distance unit of the detection unit, extract the phase value of each data to form a new sequence, and pay attention to the unwrapping of phase wrapping. Phase wrapping means that the calculated phase value is not the real phase, but the real phase falls on [-π,π] after circulating with a period of 2π, so the calculated value needs to be unwrapped.

[0061] After phase unwrapping, the least squares method can be used to perform quadratic fitting on the new data sequence, and feature extraction can be completed by calculating the fitting error. Figure 2 shown.

[0062] (4) Permutation Entropy Feature Extraction

[0063] Select N adjacent pulse data under the same distance unit of the detection unit and take the modulus to form a time series {X(1), X(2), ..., X(N)}, and use the phase space reconstruction method to obtain the reconstruction matrix.

[0064]

[0065] The time series N is 13, m and t represent the embedding dimension and delay time respectively. According to the experimental conclusions in the previous article, the embedding dimension m is 4. Since the number of pulses is small, there is no need to downsample the signal, so the delay time t is 1. Each row in the reconstruction matrix is ​​a reconstruction component, a total of K, and the reconstruction components are rearranged in ascending order to obtain a new sequence. The column index of each element position in the new sequence constitutes a set of symbol sequences:

[0066] S(l)={j 1 ,j 2 ,…,j m}, l = 1, 2, ..., k, and k ≤ m! (6)

[0067] There are a total of m! kinds of symbol sequences mapped in the m-dimensional phase space. The probability of occurrence of each symbol sequence is obtained by dividing the number of occurrences of each symbol sequence by the total number of occurrences of m! different symbol sequences, that is, P 1 ,P 2 ,...,P k . The permutation entropy is obtained based on the probability of occurrence of various symbol sequences:

[0068]

[0069] When P j =1 / m!, H pe (m) obtains the maximum value ln(m!), so the permutation entropy can be normalized, that is: 0≤H pe =H pe (m) / ln(m!)≤1.

[0070] The extracted features are used to construct a data set and normalized respectively. Then, the four types of eigenvalues ​​of each resolution unit are combined into a eigenvector and sent to the random forest classifier to complete the judgment. The 0 / 1 matrix obtained by the judgment result is transformed into a grayscale image and used as the primary recognition result. The primary recognition result is processed by removing isolated points, closing operations and hole filling in turn to improve the recognition effect and obtain the secondary processing results, as shown below;

[0071] The four features extracted in step 1 are used to construct a data set and normalized respectively to obtain an N*4 data matrix (N is the number of samples to be detected). The four eigenvalues ​​of each resolution unit form a feature vector and are sent to the random forest classifier to complete the judgment. The resolution unit judged as sea clutter is set to 0; the resolution unit judged as ground clutter is set to 1. The result obtained in this step is the random forest judgment result.

[0072] The above is the identification result of the trained model directly. If you need to analyze based on a self-constructed data set, you need to retrain the random forest model and adjust the model parameters based on the relationship between generalization error and model complexity. This patent continuously adjusts the model complexity by drawing a learning curve and grid search method to get as close to the lowest point of generalization error as possible (Note: There is no strict restriction on the parameter adjustment method in the actual method, and third-party libraries can also be used to achieve automatic parameter adjustment). The training parameters that this patent mainly focuses on are the number of subtrees, the maximum growth depth of the tree, and the minimum number of leaf samples.

[0073] The random forest judgment results obtained according to the above steps form a 0 / 1 matrix, and the matrix is ​​transformed into a grayscale image, and used as the primary identification result. The primary identification results are processed in sequence using isolated point removal, closing operation and hole filling to obtain the optimized post-processing results. Among them, removing isolated points means deleting all connected components with less than P pixels in the primary identification results, where the P value can be determined according to the specific situation of the data to be detected. The closing operation is to design a certain structural element X, and the result after removing isolated points is first expanded by X and then eroded by X. X also needs to be determined according to the situation of the data to be detected. The result of the above operation is to disconnect the small connections in each area as much as possible to facilitate the subsequent step 3. There is no parameter setting for the hole filling operation, and it can be directly used on the result after the closing operation.

[0074] The classification of large and small areas in the secondary processing results obtained in the above steps is completed by using the two empirical methods of the number of condensed points and the area width greater than N meters. The amplitude mean after the radar echo is modulated along the azimuth dimension in each small area is extracted, and the two-dimensional area is compressed into a one-dimensional vector. The vector length is the same as the number of distance units occupied by the area. By searching the spectrum peak of the one-dimensional vector, the number of peaks and the peak position are obtained. According to the number of peaks, the type of small area is identified and it is judged as a target or a small island. The details are as follows:

[0075] First, the classification of large and small areas in the secondary processing results obtained in the above steps is completed by using the two empirical methods of condensed point count and area width greater than N meters. The empirical method of condensed point count refers to extracting the connected areas determined as ground clutter in the secondary processing results one by one, checking the number of sampling points occupied by the connected areas, and treating the areas with more than P occupied sampling points as land areas. The P value here is an empirical value, which needs to be determined in combination with the actual area size represented by a single sampling point in the echo data to be measured, requiring P*single sampling point area>conventional small islands. The empirical method of area width greater than N meters is to calculate the center of mass of each connected area for the remaining connected areas after the condensed point count, and use the horizontal coordinate K of the center of mass (i.e., the distance unit in the B display image) as the reference to check how many azimuth codes the connected area occupies in the vertical coordinate. At this time, let the number of occupied azimuth codes be J, and the angle interval occupied by each azimuth code be x, then the area width L is:

[0076] L=2Ktan(Jx / 2) (8)

[0077] Thus, the area with a width greater than N meters is selected as the land area. The width value N is an empirical value, which is usually not greater than the length of a large ship, but it also needs to be selected based on the actual situation.

[0078] For the remaining small areas after screening in the above steps, the amplitude mean after the radar echo is modulated along the azimuth dimension, and the two-dimensional area is compressed into a one-dimensional vector, and the vector length is the same as the number of distance units occupied by the small area. By performing a spectrum peak search on the one-dimensional vector, the number of peaks and the peak position are obtained. Since the amplitude value of the potential target to be detected in this paper is more prominent in the time series, there is no strict requirement for the spectrum peak search algorithm in this step, but the minimum peak height needs to be considered to prevent the algorithm performance from being reduced due to too many peaks. Under ideal conditions, there will be a more obvious double peak in the peak number histogram of the island area and the potential target area, and the cluster with a smaller number of peaks is the potential target area. At this point, the small area type identification based on the number of peaks has been completed, and the extraction of potential target areas has been basically achieved.

[0079] Since the embedding dimension is an important parameter for calculating the permutation entropy and affects the final calculation results, in order to achieve better recognition results, further experiments are needed to analyze the impact of the embedding dimension on the permutation entropy value and determine the value of the embedding dimension when the difference in land and sea clutter is most significant. Figure 3 (a), (b), and (c) represent the distribution results of the permutation entropy values ​​when the embedding dimensions are 3, 4, and 5 in the same region, respectively.

[0080] As shown by the experimental results, Figure 3 Middle: (a) When the embedding dimension is 3, the entropy curve of the ground and sea clutter is relatively chaotic, with many overlaps, and it is difficult to find a threshold to distinguish the ground and sea clutter. (b) When the embedding dimension is 4, it can be seen that the sea clutter curve is significantly larger than the ground clutter curve as a whole. Compared with the embedding dimension of 3, the sea clutter entropy curve is more stable, and the overlap points of the two clutter curves are significantly reduced. (c) When the embedding dimension is 5, the entropy curves of the two types of clutter tend to be stable. However, compared with the embedding dimension of 4, the distinction between the two clutters has not increased significantly, and the increase in the embedding dimension makes the calculation more complicated, and the amount of calculation also increases accordingly. According to the above research, considering the recognition effect and the amount of calculation, the embedding dimension selected in the subsequent feature extraction is 4.

[0081] Step 2: Random forest generates classification results and optimizes classification results:

[0082] The four eigenvalues ​​of each resolution unit form a eigenvector and are sent to the random forest classifier to complete the judgment. The resolution unit judged as sea clutter is set to 0; the resolution unit judged as ground clutter is set to 1. The 0 / 1 matrix obtained from the judgment result is transformed into a grayscale image and used as the primary identification result. In order to improve the identification result, the primary identification result is processed by closing operation and hole filling in turn to improve the identification effect and obtain the optimized post-processing result.

[0083] Figure 4(a) is the recognition result obtained after sending the four eigenvalues ​​into the random forest. (b) is the image after removing isolated points by performing a closing operation on the random forest recognition result. (c) is the image after removing noise points by performing a closing operation. (d) is the image after closing operation to fill holes.

[0084] It can be seen from the experimental results that compared with the single-feature identification method, the multi-feature identification results have improved the sidelobe effect of the target area and the land area fragmentation problem mentioned above to a certain extent. It can be seen from the land and sea clutter boundary obtained after hole filling that the land and sea clutter boundary obtained by the method of this patent is closer to the coastline in the radar P display image. Morphological filtering is mainly realized by set operations of structural elements of different forms. The structural elements are composed of 0 / 1 matrices, and the area with 1 defines the scope of the morphological operation. Under the premise of ensuring the basic outline of the land and sea clutter boundary, the structural elements used in this patent when processing the random forest identification results are smaller than the structural elements used by the above-mentioned single-feature method, which means that more boundary details are retained.

[0085] Therefore, the method of this embodiment comprehensively utilizes the complementarity and correlation between multiple land and sea clutter classification features to more comprehensively describe the difference characteristics of different types of clutter, thereby improving the accuracy and stability of classification.

[0086] Step 3: Extract potential target areas

[0087] Figure 5 It is to screen the large area of ​​the multi-feature recognition results, extract the area with width>N meters (the value of N can be comprehensively considered the size of common ships in the monitored sea area), and then classify and judge the remaining small areas, and provide the location information of suspected targets in the area. In (a) and (b), the dark blue is the land area determined by the algorithm, and the light blue and yellow in (c) represent the suspected island area and the suspected target area, respectively.

[0088] Depend on Figure 5 (b) It can be seen that after filtering the large areas and extracting the areas with width > 350 meters, there are still many small areas in the figure. Figure 5(c) It can be seen that after the potential target area detection, most of the islands and land parts of concern in this data can be identified. In addition, the area that is misjudged as an island (such as the red box in Figure 15 (c)) is a sidelobe area that is separated from the actual target after image processing. Such areas will not affect subsequent target detection. For island areas that are misjudged as targets, the detection algorithm can be used to further detect the suspected target small area. Comprehensively extracting the potential target area and analyzing its classification results, the land and sea clutter identification algorithm basically meets the requirements of practical applications, and possible false alarms can also be further eliminated through subsequent processing. In addition, compared with the full-area detection in previous algorithms, the method of performing target detection in the potential target area greatly saves the amount of calculation and detection time. In practical applications, the use of this patented method can complete land and sea identification and potential target area identification at one time, which helps to optimize the marine target detection process and effectively save computing resources.

[0089] The above calculation examples of the present invention are only used to explain the calculation model and calculation process of the present invention in detail, and are not intended to limit the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A multi-feature identification method for radar land and sea clutter environment, characterized in that The following steps are involved: Step 1: Extract the echo data of 13 adjacent pulses of the resolution unit to be detected in the same distance unit to form a data sequence. The sequence is analyzed to extract the DC component ratio feature, and then the spectrum kurtosis feature is extracted by spectrum analysis. The sequence is modulo-calculated to calculate the permutation entropy feature, and the phase quadratic fitting feature is calculated after extracting the sequence phase information to complete the feature extraction of the data to be detected; Step 2: construct a data set based on the features extracted in step 1 and normalize them respectively. Then, the four types of eigenvalues ​​of each resolution unit form a feature vector and send it to the random forest classifier to complete the judgment. The 0 / 1 matrix obtained by the judgment result is transformed into a grayscale image and used as the primary recognition result. The primary recognition result is processed by removing isolated points, closing operations and hole filling in turn to improve the recognition effect and obtain the secondary processing result. Step 3: Use the two empirical methods of the number of condensed points and the area width greater than N meters to complete the classification of large and small areas in the secondary processing results obtained in step 2. Extract the amplitude mean of the radar echo modulo along the azimuth dimension in each small area, compress the two-dimensional area into a one-dimensional vector, and the vector length is the same as the number of distance units occupied by the area. By performing a spectral peak search on the one-dimensional vector, the number of peaks and the peak position are obtained. According to the number of peaks, the type of small area is identified and it is judged to be a target or a small island.

2. The radar ground-sea clutter environment multi-feature identification method according to claim 1, characterized in that The specific steps of extracting the spectrum kurtosis feature in step 1 are: The echo data of 13 adjacent pulses of the resolution unit to be detected in the same distance unit are extracted to form a data sequence. A 256-point FFT is performed on each data vector. The spectrum kurtosis of each resolution unit is calculated based on the frequency domain data after FFT. The spectrum kurtosis of all resolution units is obtained in one scanning data of the radar: Where K is the kurtosis of the data sequence after FFT, x i is the resolution unit to be detected, n is the number of scans to x i The total number of beams, is the average value of the unit to be tested and the adjacent units. The kurtosis of the standard normal distribution is 3. When the peak value of the time series curve is higher than that of the normal distribution, the kurtosis is greater than 3, that is, it presents a peak distribution. When it is lower than that of the normal distribution, the kurtosis is less than 3, that is, the peak value is relatively flat. The peak value has no dimension, and its value range is between 1 and the number of sample data.

3. The radar ground-sea clutter environment multi-feature identification method according to claim 1, characterized in that The specific steps of DC component ratio feature extraction in step 1 are: The DC power and average power of the data sequence are expressed as formula (2) and (3), respectively denoted as s and p: Where N is the number of pulses, x i is the i-th radar echo data. By calculating the ratio of the two, the DC component ratio can be obtained, which is recorded as: The DC component ratio ranges from [0,1], the DC component ratio of ground clutter is close to 1, and the DC component ratio of sea clutter is close to 0.

4. The radar ground-sea clutter environment multi-feature identification method according to claim 1, characterized in that The specific steps of extracting the phase quadratic fitting feature in step 1 are: Select 13 adjacent echo data under the same distance unit of the detection unit, extract the phase value of each data to form a new sequence, and unwrap the phase. After the phase is unwrapped, use the least squares method to perform quadratic fitting on the new data sequence, and complete the feature extraction by calculating the fitting error. Use the least squares method to perform quadratic fitting on the new data sequence, and compare each original phase data with the corresponding fitting value to calculate the residual, that is, the residual is equal to the original phase value minus the fitting phase value. Then, sum the squares of all residuals to get the total fitting error, which is the phase quadratic fitting characteristic.

5. The radar land-sea clutter environment multi-feature identification method according to claim 1, characterized in that The specific steps of permutation entropy feature extraction in step 1 are: Select N adjacent pulse data under the same distance unit of the detection unit and take the modulus to form a time series {X(1), X(2), …, X(N)}. Use the phase space reconstruction method to obtain the reconstruction matrix: The time series N is 13, m and t represent the embedding dimension and delay time respectively, the embedding dimension m is 4, the delay time t is 1, each row in the reconstruction matrix is ​​a reconstruction component, a total of K, the reconstruction components are rearranged in ascending order to obtain a new sequence, and the column index of each element position in the new sequence constitutes a set of symbol sequences: S(l)={j1,j2,…,j m }, l = 1, 2, ..., k, and k ≤ m! (6) There are a total of m! kinds of symbol sequences mapped in the m-dimensional phase space. The probability of occurrence of each symbol sequence is obtained by dividing the number of occurrences of each symbol sequence by the total number of occurrences of m! different symbol sequences, that is, P1, P2, ..., P k , according to the probability of occurrence of various symbol sequences, the permutation entropy is obtained: When P j =1 / m!, H pe (m) obtains the maximum value ln(m!), and normalizes the permutation entropy, that is: 0≤H pe =H pe (m) / ln(m!)≤1.

6. The radar land-sea clutter environment multi-feature identification method according to claim 1, characterized in that The specific steps of step 2 are: Step 2-1: construct a data set for the four features extracted in step 1 and normalize them respectively to obtain an N*4 data matrix, where N is the number of samples to be detected. The four eigenvalues ​​of each resolution unit form a feature vector and are sent to the random forest classifier to complete the judgment. The resolution unit judged as sea clutter is set to 0; the resolution unit judged as ground clutter is set to 1. The result obtained in this step is the random forest judgment result; Step 2-2: A 0 / 1 matrix is ​​formed according to the random forest judgment result obtained in step 2-1, and the matrix is ​​transformed into a grayscale image, and used as the primary identification result. The primary identification result is processed in sequence using isolated point removal, closing operation and hole filling to obtain the optimized post-processing result; wherein the isolated point removal is to delete all connected components with less than P pixels in the primary identification result, wherein the P value can be determined according to the specific situation of the data to be detected, the closing operation is to design a certain structural element X, and the result after removing the isolated points is first expanded by X and then eroded by X, and X also needs to be determined according to the situation of the data to be detected. The result of the above operation is to disconnect the small connections in each area as much as possible to facilitate the subsequent step 3. There is no parameter setting for the hole filling operation, and it is directly used for the result after the closing operation.

7. The radar land-sea clutter environment multi-feature identification method according to claim 6, characterized in that The specific steps of step 3 are: First, use the two empirical methods of condensed point number and area width greater than N meters to complete the classification of large and small areas in the secondary processing results obtained in step 2. The empirical method of condensed point number refers to extracting the connected areas determined as ground clutter in the secondary processing results one by one, checking the number of sampling points occupied by the connected areas, and treating the areas with a number of sampling points greater than P as land areas. The P value here is an empirical value, which needs to be determined in combination with the actual area size represented by a single sampling point in the echo data to be measured. It is required that P*single sampling point area>conventional small islands. The empirical method of area width greater than N meters is to calculate the center of mass of each connected area for the remaining connected areas after the condensed point number, and use the horizontal coordinate K of the center of mass as the reference to check how many azimuth codes the connected area occupies in the vertical coordinate. At this time, let the number of occupied azimuth codes be J, and the angle interval occupied by each azimuth code be x, then the area width L is: L=2Ktan(Jx / 2) (8) Thus, the areas with a width greater than N meters are selected as land areas, where the width value N is an empirical value; For the remaining small areas after screening in the above steps, the amplitude mean of the radar echo after modulus is extracted along the azimuth dimension, and the two-dimensional area is compressed into a one-dimensional vector. The vector length is the same as the number of distance units occupied by the small area. By performing spectral peak search on the one-dimensional vector, the number and position of peaks are obtained.

Images