Precession target scattering center separation method based on three-dimensional radar data cube

Abstract

The invention discloses a precession target scattering center separation method based on a three-dimensional radar data cube, and the method comprises the steps: carrying out the processing of a radar echo pulse signal of a conical precession target, and building a precession model of the target, so as to distinguish the different signal characteristics of a local scattering center and a sliding scattering center; carrying out two-dimensional distance-Doppler imaging on the radar echo signal, arranging multiple frames of obtained distance-Doppler images according to a time sequence within a preset imaging time, and constructing a distance-Doppler image sequence; based on a feature enhancement method of a binary mask, automatically separating data of a local scattering center and a sliding scattering center from the distance-Doppler image sequence, extracting coordinate information of each scattering center, and generating a three-dimensional radar data cube containing distance, frequency and time dimensions; and performing trajectory tracking and separation on the sliding scattering center signal in the three-dimensional radar data cube by using a three-dimensional optimal path selection method.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and more specifically, to a method for separating the scattering center of a precessing target based on a three-dimensional radar data cube. Background Technology

[0002] Micro-motion phenomena are widespread in various moving targets, such as the rotation of helicopter rotors, the vibration of aircraft wings, and the precession of ballistic targets. The micro-Doppler effect in radar echo signals refers to the Doppler frequency modulation phenomenon caused by minute movements of a target or its components. This effect can reflect the target's structure and micro-motion characteristics, thus providing important information for target identification. Because micro-motion is typically small-amplitude, uncontrolled motion, it is difficult to imitate artificially. Therefore, target identification technology based on micro-motion characteristics is considered one of the most promising directions in radar target identification and has been widely applied in fields such as human motion recognition, UAV detection, and animal behavior analysis.

[0003] In radar signal processing, the separation of scattering centers is typically achieved based on different types of radar images, such as two-dimensional range-Doppler (RD) images, time-frequency distribution maps, and high-resolution range profiles (HRRP). The acquisition of these images depends on the corresponding signal processing methods. Therefore, to effectively characterize the multi-scale decomposition phenomenon in a signal, an appropriate radar image representation format must be selected.

[0004] Currently, scattering center separation and estimation based on two-dimensional radar images has become a widely researched area. For example, the CSRDI-MGPTF method is used to extract micro-Doppler curves, and parameter estimation methods under occlusion effects are discussed. Although this algorithm has the ability to quickly separate scattering centers and has high parameter estimation accuracy, its method still relies on the processing of two-dimensional images or time-frequency distributions, and its ability to separate complex motion trajectories is limited.

[0005] Relatedly, a micro-Doppler parametric model (mDPM) for a conical ballistic target was established, and the scattering center was analyzed using the genetic algorithm-generalized parametric time-frequency transform (GA-GPTF) method. The effectiveness of this method is overly dependent on the performance of the time-frequency analysis tool; when there is ambiguity or overlap in the time-frequency distribution, the accuracy of model matching and parameter estimation is significantly affected.

[0006] Furthermore, scattering center separation estimation based on two-dimensional radar images can also be achieved by establishing a framework for signal decomposition and data correlation using independent component analysis (ICA) in a distributed radar network. While this method saves computational resources and enables scattering center analysis, its system complexity is high, requiring multiple radars to work together, making it unsuitable for the common and important application scenario of monostation radar.

[0007] In summary, existing scattering center separation methods suffer from the following drawbacks: most are limited to analysis in the two-dimensional signal domain (such as RD images and time-frequency planes), making it difficult to handle complex scattering centers with overlapping trajectories in two-dimensional space; their separation effect is heavily dependent on the performance of specific signal processing tools (such as time-frequency analysis), resulting in insufficient robustness; and some effective solutions (such as radar network-based methods) have stringent implementation conditions and lack universality.

[0008] Therefore, there is an urgent need for a new technical means to effectively separate the scattering center of a precessing target in a higher dimension, thereby overcoming the limitations of two-dimensional analysis. Summary of the Invention

[0009] To address at least one deficiency or improvement requirement of existing technologies, this invention provides a method for separating the scattering centers of precessing targets based on a three-dimensional radar data cube. This method solves the problems of existing methods being unable to effectively distinguish scattering centers whose trajectories overlap on a two-dimensional plane, and being overly reliant on complex signal processing tools with insufficient robustness. It achieves efficient, automatic, and accurate separation of echo signals from each scattering center on a precessing target, providing a reliable data foundation for subsequent micro-motion parameter estimation.

[0010] To achieve the above objectives, according to a first aspect of the present invention, a method for separating the scattering centers of a precessing target based on a three-dimensional radar data cube is provided. The method includes: processing the radar echo pulse signal of a conical precessing target to establish a precession model of the target, thereby distinguishing the different signal characteristics of local scattering centers and sliding scattering centers; performing two-dimensional range-Doppler imaging on the radar echo signal, and arranging the obtained multi-frame range-Doppler images in chronological order within a preset imaging time to construct a range-Doppler image sequence; automatically separating the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence using a feature enhancement method based on a binary mask, and extracting the coordinate information of each scattering center to generate a three-dimensional radar data cube containing range, frequency, and time dimensions; and using a three-dimensional optimal path selection method to track and separate the sliding scattering center signal in the three-dimensional radar data cube.

[0011] In an exemplary embodiment, the micro-motion distance and micro-motion frequency of the local scattering center are sinusoidal in the time domain, while the micro-motion distance and micro-motion frequency of the sliding scattering center are non-sinusoidal, modulated by both target precession and spin.

[0012] In an exemplary embodiment, the feature enhancement method based on a binary mask, which automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence and extracts the coordinate information of each scattering center to generate a three-dimensional radar data cube containing range, frequency, and time dimensions, includes: performing pixel-level adaptive Wiener filtering on each frame of range-Doppler image to suppress noise and retain image features; applying the Otsu thresholding algorithm to the filtered image matrix to calculate multiple hierarchical thresholds and generating a first-level binary mask based on the highest threshold; performing a Hadamard product operation on the first-level binary mask and the original range-Doppler image to extract an image matrix containing only local scattering center information, and associating the coordinates of local scattering centers in all frames to achieve three-dimensional trajectory association; performing a logical NOT operation on the first-level binary mask and then performing a Hadamard product operation on the original range-Doppler image to obtain an intermediate image matrix that does not contain local scattering centers; and performing filtering and thresholding on the intermediate matrix again to generate a second-level binary mask to extract an image matrix containing sliding scattering center information.

[0013] In an exemplary embodiment, the first-level binary mask The generation method is as follows:

[0014] in, For matrix The pixels in This is the image matrix after Wiener filtering. This represents the highest threshold obtained using the Otsu algorithm.

[0015] In an exemplary embodiment, the three-dimensional optimal path selection method specifically comprises: defining a path as a sequence of position coordinates of the scattering center in each frame of the range-Doppler image sequence; finding the optimal path through an optimization algorithm, wherein the optimal path is the path with the minimum total motion cost over the entire time series; the total motion cost is composed of the sum of the cost of change in the distance dimension and the cost of change in the frequency dimension.

[0016] In an exemplary embodiment, the cost of the change in the distance dimension is calculated as follows: when the absolute value of the change in the distance unit coordinates of the scattering center between two adjacent frames is not greater than a preset distance change threshold, the cost of the change in the distance dimension is considered to be zero; when the absolute value of the change in the distance unit coordinates is greater than the distance change threshold, the cost of the change in the distance dimension is proportional to the square of the portion of the change that exceeds the threshold, and is adjusted by a distance penalty factor.

[0017] In an exemplary embodiment, the cost of frequency dimension change is calculated as follows: when the absolute value of the change in frequency unit coordinates between two adjacent frames is not greater than a preset frequency change threshold, the cost of frequency dimension change is considered to be zero; when the absolute value of the change in frequency unit coordinates is greater than the frequency change threshold, the cost of frequency dimension change is proportional to the square of the portion of the change exceeding the threshold, and is adjusted by a frequency penalty factor.

[0018] According to a second aspect of the embodiments of this application, a precession target scattering center separation system is also provided for implementing the above-described method for separating the scattering centers of a precession target based on a three-dimensional radar data cube. This system may include: a modeling unit for processing the radar echo pulse signal of a conical precession target and establishing a precession model of the target to distinguish the different signal characteristics of local scattering centers and sliding scattering centers; a construction unit for performing two-dimensional range-Doppler imaging on the radar echo signal and arranging the obtained multi-frame range-Doppler images in chronological order within a preset imaging time to construct a range-Doppler image sequence; a preliminary separation unit for automatically separating the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence based on a binary mask feature enhancement method, and extracting the coordinate information of each scattering center to generate a three-dimensional radar data cube containing range, frequency, and time dimensions; and a precise separation unit for using a three-dimensional optimal path selection method to track and separate the sliding scattering center signal in the three-dimensional radar data cube.

[0019] According to a third aspect of the invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to execute the above-described method for separating the precession target scattering center based on a three-dimensional radar data cube at runtime.

[0020] According to a fourth aspect of the present invention, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described method for separating the precession scattering center based on a three-dimensional radar data cube through the computer program.

[0021] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: (1) This invention provides a method for separating the scattering centers of precessing targets based on a three-dimensional radar data cube. By constructing a range-frequency-time three-dimensional data cube, an independent trajectory representation space is provided for each scattering center. This allows trajectories that may intersect or overlap on a two-dimensional plane to be clearly unfolded and distinguished in three-dimensional space. This fundamentally solves the technical bottleneck of the difficulty in separating scattering centers due to trajectory overlap, and significantly improves the success rate and reliability of the separation task.

[0022] (2) Using the binary mask feature enhancement method, based on the physical characteristic that local scattering centers usually have a stronger radar cross-section, they are automatically and quickly separated from complex echoes. This method does not rely on complex parameter adjustments, has a high degree of automation, and high execution efficiency. For sliding scattering centers, a three-dimensional optimal path selection algorithm is adopted. This algorithm simulates the basic physical law of the smoothness and continuity of the scattering center's motion, and can intelligently find the optimal path that best matches the actual motion state among many possible paths. This avoids the problem of traditional methods that may require manual intervention for trajectory association, and realizes fully automatic and high-precision trajectory tracking.

[0023] (3) Because the scattering center is separated more thoroughly and the trajectory is extracted more completely, it provides high-quality input data for subsequent estimation of micro-motion parameters (such as precession angle, precession period, structural size, etc.). This improves the accuracy of final target recognition and parameter inversion. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating an optional method for separating the scattering center of a precessing target based on a three-dimensional radar data cube, provided for an embodiment of this application; Figure 2 A flowchart illustrating another optional method for separating the scattering center of a precessing target based on a three-dimensional radar data cube, provided for an embodiment of this application; Figure 3 A schematic diagram of a CAD model of an optional electromagnetic calculation target provided for an embodiment of this application; Figure 4 A schematic diagram of the geometric structure of an optional electromagnetic computing target provided for an embodiment of this application; Figure 5 A schematic diagram illustrating an optional RCS effect provided for an embodiment of this application; Figure 6 A schematic diagram of an optional RD sequence provided for an embodiment of this application; Figure 7 A schematic diagram of an optional radar data cube associated with local scattering centers provided for an embodiment of this application; Figure 8 A schematic diagram of an optional radar data cube with all scattering centers associated, provided as an embodiment of this application; Figure 9 A schematic diagram of an optional time-range projection of a radar data cube provided for an embodiment of this application; Figure 10 A schematic diagram of an optional time-frequency projection of a radar data cube provided for an embodiment of this application; Figure 11 A schematic diagram of the range-frequency projection of an optional radar data cube provided for an embodiment of this application; Figure 12 This is a schematic diagram of an optional electronic device provided in an embodiment of this application. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0027] The terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0028] According to one aspect of the embodiments of this application, a method for separating the scattering center of a precessing target based on a three-dimensional radar data cube is provided. The following is in conjunction with... Figure 1 This application describes a method for separating the scattering center of a precessing target based on a three-dimensional radar data cube, as provided in the embodiments of this application.

[0029] Figure 1 This is a flowchart illustrating an optional method for separating the scattering center of a precessing target based on a three-dimensional radar data cube, as provided in an embodiment of this application. Figure 1 As shown, the process of this method may include the following steps: S102 processes the radar echo pulse signal of a conical precession target and establishes a precession model of the target to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. S104, perform two-dimensional range-Doppler imaging on the radar echo signal, and within a preset imaging time, arrange the obtained multi-frame range-Doppler images in chronological order to construct a range-Doppler image sequence; S106, a feature enhancement method based on binary mask automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, extracts the coordinate information of each scattering center, and generates a three-dimensional radar data cube containing distance, frequency, and time dimensions; S108, using the three-dimensional optimal path selection method, the sliding scattering center signal in the three-dimensional radar data cube is tracked and separated.

[0030] This application provides a method for separating the scattering center of a precessing target based on a three-dimensional radar data cube. By distinguishing the characteristics of different types of scattering centers, it achieves effective separation of radar echo signals of precessing targets in the three-dimensional signal domain.

[0031] Combination Figure 1 and Figure 2 As shown, the method includes the following steps: Step 1: Process the target echo pulse to obtain the precession model of the cone-shaped target, and perform signal analysis on the target echo by distinguishing different scattering centers.

[0032] Assuming accurate compensation is made for the target's translational motion, the target echo can be described as follows:

[0033] In the formula, Indicates the first The scattering coefficient of each scattering center Representing the The instantaneous Doppler frequency of each scattering center Represents time.

[0034] According to the scattering center theory, the fixed point at the top of a conical target can be assumed to be the local scattering center. The micro-motion distance of this local scattering center is then... and micro-motion frequency It can be represented as

[0035] In the formula, It is the signal wavelength. It is the distance from the top of the cone to the center of mass. It is an instantaneous time. The precession angle is the angle between the cone rotation axis and the target axis of symmetry. Represents the radar line-of-sight angle. Let be the cone rotation angular velocity, and the target depression angle under initial conditions be . . This represents the average field of view angle.

[0036] Therefore, it can be seen that the micro-motion distance and micro-motion frequency of the local scattering center at the cone apex both exhibit sinusoidal form in the time domain.

[0037] Unlike the local scattering center at the apex of the cone, the sliding scattering center generated by diffraction at the bottom edge of the cone is modulated by both the target's spin and the cone's shape, making it a scattering center that slides at the bottom. Assuming we denote the sliding scattering center as points B and C, the micro-motion distance at point B is... and micro-motion frequency The micro-movement distance of point C and micro-motion frequency It can be represented as

[0038]

[0039] in It is the distance from the center of mass to the bottom of the cone. This represents the radius of the base of the cone.

[0040] It can be seen that the micro-Doppler at the sliding scattering center exhibits a complex non-sinusoidal form. Compared to the local scattering center, it increases... This modulation term results in a motion pattern of the sliding scattering center that differs significantly from the precession of the cone target itself.

[0041] The second step is to perform two-dimensional range-Doppler imaging on the echo signal. Over a period of time, the obtained multi-frame RD (range-Doppler) images can be combined into an RD sequence.

[0042] The radar transmits a linear frequency modulated (LFM) signal, from the... The first scattering center received the first Frame radar echo signal Its defining formula is

[0043] in , These represent fast time and slow time, respectively. Indicates the linear frequency modulation slope. For radar carrier frequency, It is the pulse duration. Representing the The scattering coefficient of each scattering center Represents slow time Time radar and the first The distance between the scattering centers Represents the speed of light. This represents a rectangular function.

[0044] The reference signal can be written as:

[0045] in This indicates the distance between the radar and the reference point.

[0046] After frequency modulation removal, the echo signal can be represented as:

[0047] in This represents the difference between the target distance and the reference distance.

[0048] Use process parameters replace Regarding By performing a Fourier transform, a high-resolution range image of the target can be obtained:

[0049] In the formula, Indicates frequency.

[0050] The target's RD image can be obtained by performing a Fast Fourier Transform on the above equation. Over a period of time, the obtained multiple RD images can be combined into an RD sequence.

[0051] The third step is to use a feature enhancement method based on a binary mask to automatically separate local scattering center data from sliding scattering center data, and extract the target scattering center coordinate information from the RD sequence to obtain a three-dimensional range-frequency-time radar data cube.

[0052] Specifically, the local mean and variance of the image are first estimated, and a binary mask is created using multi-level thresholds obtained through Otsu's method. Local scattering center information is extracted to correlate these centers. The binary mask is then reconstructed to extract sliding scattering center information.

[0053] Step 4: Separate the sliding scattering centers using the three-dimensional optimal path selection method.

[0054] Through the above steps, the radar echo pulse signal of the conical precessing target is processed to establish a precession model of the target, thereby distinguishing the different signal characteristics of the local scattering center and the sliding scattering center. Two-dimensional range-Doppler imaging is performed on the radar echo signal, and within a preset imaging time, the obtained multi-frame range-Doppler images are arranged in chronological order to construct a range-Doppler image sequence. Based on a feature enhancement method using a binary mask, the data of the local scattering center and the sliding scattering center are automatically separated from the range-Doppler image sequence, and the coordinate information of each scattering center is extracted to generate a three-dimensional radar data cube containing range, frequency, and time dimensions. Using a three-dimensional optimal path selection method, the sliding scattering center signal in the three-dimensional radar data cube is tracked and separated. This solves the problems of existing methods, such as difficulty in effectively distinguishing scattering centers whose trajectories overlap on a two-dimensional plane, excessive reliance on complex signal processing tools, and insufficient robustness. It achieves efficient, automatic, and accurate separation of the echo signals of each scattering center on the precessing target, providing a reliable data foundation for subsequent micro-motion parameter estimation.

[0055] In an exemplary embodiment, the micro-motion distance and micro-motion frequency of the local scattering center are sinusoidal in the time domain, while the micro-motion distance and micro-motion frequency of the sliding scattering center are non-sinusoidal, modulated by both target precession and spin.

[0056] In an exemplary embodiment, the feature enhancement method based on a binary mask automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, extracts the coordinate information of each scattering center, and generates a three-dimensional radar data cube containing range, frequency, and time dimensions, including: Pixel-level adaptive Wiener filtering is applied to each frame of distance-Doppler image to suppress noise and preserve image features; The Otsu thresholding algorithm is applied to the filtered image matrix to calculate multiple hierarchical thresholds and generate a first-level binary mask based on the highest threshold. The first-level binary mask is subjected to Hadamard product operation with the original range-Doppler image to extract the image matrix containing only local scattering center information, and the coordinates of local scattering centers in all frames are associated to realize three-dimensional trajectory association; After performing a logical NOT operation on the first-level binary mask, a Hadamard product operation is performed with the original range-Doppler image to obtain an intermediate image matrix that does not contain local scattering centers. This intermediate matrix is ​​then filtered and thresholded again to generate a second-level binary mask to extract an image matrix containing information about sliding scattering centers.

[0057] Optionally, a feature enhancement method based on a binary mask is designed to automatically separate local scattering center data from sliding scattering center data, and extract the target scattering center coordinate information from the RD sequence to obtain a three-dimensional range-frequency-time radar data cube.

[0058] Assume the reconstructed RD image is , , These represent the number of units in the distance and frequency dimensions, respectively. The local mean and variance of each pixel are estimated using the following formula:

[0059]

[0060] In the formula, express middle Local neighborhood of size Represents the local mean. This represents local variance. Represents the pixels in the reconstructed RD image, and , .

[0061] A pixel-level adaptive Wiener filter can be expressed as:

[0062] In the formula, This represents the noise variance.

[0063] Pixel-level adaptive Wiener filters can suppress noise while preserving useful properties of the image matrix. Let the RD image matrix after the pixel-level adaptive Wiener filter be defined as... The matrix was calculated using Otsu's method. The three thresholds are ordered from smallest to largest as follows: Optionally, the first-level binary mask The generation method is as follows:

[0064] in, For matrix The pixels in This is the image matrix after Wiener filtering. This represents the highest threshold obtained using the Otsu algorithm.

[0065] First-level binary mask The elements in satisfy That is, it is retained, and the rest are set to 0. The elements in satisfy This indicates that the pixel belongs to the background or noise of the image; The elements in satisfy This indicates that the pixel belongs to a sliding scattering center or a strong noise point; The elements in satisfy This indicates that the pixel belongs to the local scattering center.

[0066] Image matrix containing only local scattering center coordinate information The following formula can be used to extract it:

[0067] in, , These represent the binary mask matrix and the RD image matrix, respectively. The representative is the Hadamard product.

[0068] Assume the number of frames in the RD sequence is Then, by repeating the above operation, the coordinate information of the local scattering center in each frame of the RD image can be obtained.

[0069] Image matrix Extract using the following formula:

[0070] in Represents the logical NOT operation.

[0071] This means the image matrix It does not contain the signal component from the local scattering center. Similarly, we will After passing through a pixel-level adaptive Wiener filter, the resulting image matrix is ​​defined as follows: Multi-level thresholds were obtained using Otsu's method. Similarly, the second-level binary mask can be represented as .

[0072] Image matrix containing information about the sliding scattering center Extract using the following formula:

[0073] The coordinates of the scattering centers in the RD sequence form a range-frequency-time radar data cube. The method of forming the range-frequency-time radar data cube from the coordinates of the scattering centers will not be elaborated here.

[0074] In one optional example, the three-dimensional optimal path selection method is specifically as follows: A path is defined as the sequence of position coordinates of the scattering center in each frame of the range-Doppler image sequence. The optimal path is found by an optimization algorithm. The optimal path is the path with the minimum total motion cost over the entire time series. The total motion cost is composed of the sum of the cost of changes in the distance dimension and the cost of changes in the frequency dimension.

[0075] This embodiment proposes a three-dimensional optimal path selection method, assuming that the... The RD image matrix of the frame is , The number of RD sequence frames, Represents the set of real numbers. This represents the number of distance cells after reconstruction. This represents the number of frequency units after reconstruction. Assume the... The location of a scattering center in the RD image of the frame is , , Assume the locations of the scattering centers form a sequence. for:

[0076] It can be seen that the sequence The elements in the sequence represent the positions of scattering centers in a specific frame of the RD sequence; therefore, this sequence is called the path of the scattering centers. Each frame contains a single scattering center position. There are several possibilities, so the entire path corresponding to a single scattering center has A choice.

[0077] The optimal path must satisfy:

[0078] in, The variable value represents the minimum value of the objective function.

[0079] Optionally, the cost of the change in the distance dimension is calculated as follows: When the absolute value of the change in the distance unit coordinates of the scattering center between two adjacent frames is not greater than the preset distance change threshold, the cost of the change in the distance dimension is considered to be zero. When the absolute value of the change in the coordinates of the distance unit is greater than the distance change threshold, the cost of the change in the distance dimension is proportional to the square of the portion of the change that exceeds the threshold, and is adjusted by a distance penalty factor.

[0080] The distance penalty function represents the distance from the first... A certain element in a frame up to the first The distance cost of a certain element in a frame can be defined as the function:

[0081] in, Defined as the threshold of the distance penalty function; This represents the distance penalty factor. It can be seen that the greater the difference between the distance units corresponding to two time intervals, the larger the value of the corresponding distance penalty function.

[0082] In an exemplary embodiment, the cost of the change in the frequency dimension is calculated as follows: When the absolute value of the change in the frequency unit coordinates of the scattering center between two adjacent frames is not greater than the preset frequency change threshold, the cost of the frequency dimension change is considered to be zero. When the absolute value of the change in the frequency unit coordinate is greater than the frequency change threshold, the cost of the change in the frequency dimension is proportional to the square of the portion of the change that exceeds the threshold, and is adjusted by a frequency penalty factor.

[0083] The distance penalty function represents the distance from the first... A certain element in a frame up to the first The distance cost of a certain element in a frame can be defined as the function:

[0084] in, Defined as the threshold of the distance penalty function; This represents the distance penalty factor. It can be seen that the greater the difference between the distance units corresponding to two time intervals, the larger the value of the corresponding distance penalty function.

[0085] The frequency penalty function represents the frequency penalty function from the first... A certain element in a frame up to the first The frequency cost of a specific element in a frame. This function can be defined as:

[0086] in, The threshold is defined as the frequency penalty function; This represents the frequency penalty factor. It can be seen that the greater the difference between the frequency units corresponding to two time intervals, the larger the value of the corresponding frequency penalty function.

[0087] As can be seen from the cost function of the optimal path, the optimal path takes into account the distance and frequency of the scattering center in the radar data cube, and can guarantee that the position of the scattering center sequence that satisfies the optimal path will not change abruptly.

[0088] In another alternative embodiment: simulation experiment of a precession target scattering center separation method based on a distance-frequency-time data cube.

[0089] Simulation parameter settings: It is assumed that the translational motion of the conical precession target has been compensated. It is assumed that the radar transmission carrier frequency... ,bandwidth Pulse repetition frequency The linear frequency modulated signal, the radar observation time is The precession angle of the cone is... Precession angular velocity .

[0090] like Figure 3 and Figure 4 As shown, a geometric model of the target is constructed, where the distance from the target's center of mass to the top is... Distance from the center of mass to the bottom Base radius The cone's apex radius is 7.5 cm. Electromagnetic data of the target were acquired using FEKO software and physical optics methods, such as... Figure 5 As shown.

[0091] Two-dimensional range-Doppler imaging was performed on the echo signal. The RD sequence had 100 frames, with a time interval of 0.02 s between each RD frame. Arranging the resulting RD sequence along the time dimension yielded... Figure 6 It can be seen that each RD image has three scattering centers of varying intensities. Each RD image can be processed using Wiener filtering, and the resulting two-dimensional matrix can be quantized for the first time using the multi-level thresholds obtained through the first Otsu's method. At this point, the image matrix containing only local scattering center information can be obtained. Extract and process the data to obtain the coordinates of the local scattering centers. Performing the above processing on each frame of the RD image yields the coordinates of the local scattering centers at each time step. Correlating these coordinates allows for the automatic association of the local scattering centers. Figure 7 As shown.

[0092] Next, the proposed 3D optimal path selection method is used to process the radar data cube, with the following parameters: , , , .

[0093] The final scattering center separation results are as follows Figure 8 As shown in the figure, curve separation can be achieved better in the three-dimensional domain. To better observe the separation effect of scattering centers, the three-dimensional image can be projected onto a two-dimensional plane, such as... Figure 9 , Figure 10 , Figure 11 As shown.

[0094] This embodiment provides a method for separating the scattering centers of precessing targets based on a range-frequency-time data cube. The method processes the target echo pulse to derive a precession model of a conical target, distinguishes different scattering centers, and performs signal analysis on the target echo. Two-dimensional range-Doppler imaging is performed on the echo signal, and the two-dimensional RD images obtained at different times are arranged in the time dimension to obtain an RD sequence. A feature enhancement method based on a binary mask is designed to automatically separate local scattering center data from sliding scattering center data, and the target scattering center coordinate information is extracted from the RD sequence, thus obtaining a three-dimensional range-frequency-time radar data cube. The sliding scattering center is separated using a three-dimensional optimal path selection method. This enables rapid and accurate separation of the echo signal of a conical precessing target, laying the foundation for further micro-Doppler characteristic parameter estimation.

[0095] According to another aspect of the embodiments of this application, a precession target scattering center separation system for implementing the above-described method for separating precession target scattering centers based on a three-dimensional radar data cube is also provided. This system may include: The modeling unit is used to process the radar echo pulse signal of a cone-shaped precessing target and establish a precession model of the target to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. The construction unit is used to perform two-dimensional range-Doppler imaging on radar echo signals, and within a preset imaging time, arranges the obtained multi-frame range-Doppler images in chronological order to construct a range-Doppler image sequence. The preliminary separation unit is used to automatically separate the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence based on the feature enhancement method of binary mask, and extract the coordinate information of each scattering center to generate a three-dimensional radar data cube containing the distance, frequency and time dimensions. The precision separation unit is used to track and separate the sliding scattering center signal in the three-dimensional radar data cube using a three-dimensional optimal path selection method.

[0096] Through the aforementioned modules, the radar echo pulse signal of a conical precessing target is processed to establish a precession model of the target, thereby distinguishing the different signal characteristics of local scattering centers and sliding scattering centers. Two-dimensional range-Doppler imaging is performed on the radar echo signal, and within a preset imaging time, multiple frames of range-Doppler images are arranged chronologically to construct a range-Doppler image sequence. Based on a feature enhancement method using a binary mask, the data of local scattering centers and sliding scattering centers are automatically separated from the range-Doppler image sequence, and the coordinate information of each scattering center is extracted to generate a three-dimensional radar data cube containing range, frequency, and time dimensions. Using a three-dimensional optimal path selection method, the sliding scattering center signal in the three-dimensional radar data cube is tracked and separated. This solves the problems of existing methods, such as difficulty in effectively distinguishing scattering centers whose trajectories overlap on a two-dimensional plane, excessive reliance on complex signal processing tools, and insufficient robustness. It achieves efficient, automatic, and accurate separation of echo signals from various scattering centers on a precessing target, providing a reliable data foundation for subsequent micro-motion parameter estimation.

[0097] According to another aspect of the embodiments of this application, a storage medium is also provided. Optionally, in this embodiment, the storage medium can be used to execute the program code of any of the above-described methods for separating the precession target scattering centers based on a three-dimensional radar data cube in the embodiments of this application.

[0098] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: S1 processes the radar echo pulse signal of the conical precession target and establishes the target's precession model to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. S2, perform two-dimensional range-Doppler imaging on the radar echo signal, and within a preset imaging time, arrange the obtained multi-frame range-Doppler images in chronological order to construct a range-Doppler image sequence; S3, a feature enhancement method based on binary masks, automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, extracts the coordinate information of each scattering center, and generates a three-dimensional radar data cube containing range, frequency, and time dimensions; S4. Using the three-dimensional optimal path selection method, the sliding scattering center signal in the three-dimensional radar data cube is tracked and separated.

[0099] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated in this embodiment.

[0100] The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0101] According to another aspect of the embodiments of this application, an electronic device for implementing the above-described method for separating the scattering center of a precessing target based on a three-dimensional radar data cube is also provided. The electronic device may be a server, a terminal, or a combination thereof.

[0102] Figure 12 This is a schematic diagram of the structure of an optional electronic device according to an embodiment of this application, such as... Figure 12 As shown, it includes a processor 1202, a communication interface 1204, a memory 1206, and a communication bus 1208. The processor 1202, communication interface 1204, and memory 1206 communicate with each other via the communication bus 1208. Memory 1206 is used to store computer programs; When processor 1202 executes a computer program stored in memory 1206, it performs the following steps: S1 processes the radar echo pulse signal of the conical precession target and establishes the target's precession model to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. S2, perform two-dimensional range-Doppler imaging on the radar echo signal, and within a preset imaging time, arrange the obtained multi-frame range-Doppler images in chronological order to construct a range-Doppler image sequence; S3, a feature enhancement method based on binary masks, automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, extracts the coordinate information of each scattering center, and generates a three-dimensional radar data cube containing range, frequency, and time dimensions; S4. Using the three-dimensional optimal path selection method, the sliding scattering center signal in the three-dimensional radar data cube is tracked and separated.

[0103] Optionally, the communication bus can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 12 The symbol is represented by a single thick line, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned electronic device and other devices.

[0104] The memory may include RAM, or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0105] The processors mentioned above can be general-purpose processors, including but not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; they can also be DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0106] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0107] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0108] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0109] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0110] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of embodiments of this disclosure upon considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

[0111] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are 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.

[0112] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for separating the scattering center of a precessing target based on a three-dimensional radar data cube, characterized in that, include: The radar echo pulse signal of the conical precession target is processed to establish the precession model of the target in order to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. Two-dimensional range-Doppler imaging is performed on the radar echo signal, and within a preset imaging time, the obtained multi-frame range-Doppler images are arranged in chronological order to construct a range-Doppler image sequence; The feature enhancement method based on binary masks automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, and extracts the coordinate information of each scattering center to generate a three-dimensional radar data cube containing range, frequency and time dimensions. The three-dimensional optimal path selection method is used to track and separate the sliding scattering center signal in the three-dimensional radar data cube.

2. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 1, characterized in that, The micro-motion distance and micro-motion frequency of the local scattering center are sinusoidal in the time domain, while the micro-motion distance and micro-motion frequency of the sliding scattering center are non-sinusoidal, modulated by both target precession and spin.

3. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 1, characterized in that, The feature enhancement method based on binary masks automatically separates the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence, extracts the coordinate information of each scattering center, and generates a three-dimensional radar data cube containing range, frequency, and time dimensions, including: Pixel-level adaptive Wiener filtering is applied to each frame of distance-Doppler image to suppress noise and preserve image features; The Otsu thresholding algorithm is applied to the filtered image matrix to calculate multiple hierarchical thresholds and generate a first-level binary mask based on the highest threshold. The first-level binary mask is subjected to Hadamard product operation with the original range-Doppler image to extract the image matrix containing only local scattering center information, and the coordinates of local scattering centers in all frames are associated to realize three-dimensional trajectory association; After performing a logical NOT operation on the first-level binary mask, a Hadamard product operation is performed with the original range-Doppler image to obtain an intermediate image matrix that does not contain local scattering centers. This intermediate matrix is ​​then filtered and thresholded again to generate a second-level binary mask to extract an image matrix containing information about sliding scattering centers.

4. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 3, characterized in that, First-level binary mask The generation method is as follows: in, For matrix The pixels in This is the image matrix after Wiener filtering. This represents the highest threshold obtained using the Otsu algorithm.

5. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 1, characterized in that, The three-dimensional optimal path selection method is as follows: A path is defined as the sequence of position coordinates of the scattering center in each frame of the range-Doppler image sequence. The optimal path is found by an optimization algorithm. The optimal path is the path with the minimum total motion cost over the entire time series. The total motion cost is composed of the sum of the cost of changes in the distance dimension and the cost of changes in the frequency dimension.

6. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 5, characterized in that, The cost of the change in the distance dimension is calculated as follows: When the absolute value of the change in the distance unit coordinates of the scattering center between two adjacent frames is not greater than the preset distance change threshold, the cost of the change in the distance dimension is considered to be zero. When the absolute value of the change in the coordinates of the distance unit is greater than the distance change threshold, the cost of the change in the distance dimension is proportional to the square of the portion of the change that exceeds the threshold, and is adjusted by a distance penalty factor.

7. The method for separating the scattering center of a precessing target based on a three-dimensional radar data cube as described in claim 5, characterized in that, The cost of the change in the frequency dimension is calculated as follows: When the absolute value of the change in the frequency unit coordinates of the scattering center between two adjacent frames is not greater than the preset frequency change threshold, the cost of the frequency dimension change is considered to be zero. When the absolute value of the change in the frequency unit coordinate is greater than the frequency change threshold, the cost of the change in the frequency dimension is proportional to the square of the portion of the change that exceeds the threshold, and is adjusted by a frequency penalty factor.

8. A precession target scattering center separation system based on a three-dimensional radar data cube, characterized in that, include: The modeling unit is used to process the radar echo pulse signal of a cone-shaped precessing target and establish a precession model of the target to distinguish the different signal characteristics of the local scattering center and the sliding scattering center. The construction unit is used to perform two-dimensional range-Doppler imaging on radar echo signals, and within a preset imaging time, arranges the obtained multi-frame range-Doppler images in chronological order to construct a range-Doppler image sequence. The preliminary separation unit is used to automatically separate the data of local scattering centers and sliding scattering centers from the range-Doppler image sequence based on the feature enhancement method of binary mask, and extract the coordinate information of each scattering center to generate a three-dimensional radar data cube containing the distance, frequency and time dimensions. The precision separation unit is used to track and separate the sliding scattering center signal in the three-dimensional radar data cube using a three-dimensional optimal path selection method.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 7.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method of any one of claims 1 to 7 through the computer program.

Images