Imaging method, system, storage medium based on millimeter wave radar target cluster detection

Abstract

The embodiment of the application provides a kind of based on millimeter wave radar target cluster detection imaging method, system, storage medium, method includes: according to receiving channel information, non-coherent accumulation is carried out to radar cube data, obtains two-dimensional detection data matrix;Two-dimensional noise floor fitting estimation is carried out to two-dimensional detection data matrix in distance dimension and speed dimension respectively, obtains two-dimensional noise floor plane fitting matrix;Two-dimensional detection data matrix and two-dimensional noise floor plane fitting matrix are compared between processing, generate target detection binary matrix, and determine the target cluster of target detection binary matrix;According to the constraint processing of the horizontal angle value, reference distance value, reference speed value and reference peak intensity information corresponding to target cluster of pre-set observation range information, determine target imaging spectrum.The application utilizes two-dimensional noise floor fitting estimation algorithm to detect and image target cluster, reduce the imaging calculation of non-target point, improve target cluster imaging efficiency and imaging quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of millimeter wave radar imaging, and in particular to a millimeter wave radar target cluster detection-based imaging method and system and a storage medium. BACKGROUND

[0002] The development of automotive radar systems has great social and economic benefits. With the development of millimeter wave radar technology, imaging radar has gradually become a research hotspot in the field of autonomous driving. Due to the complex and changeable driving environment of automobiles, the performance of radar target detection will be affected by the presence of background noise and interference clutter. In the field of automotive radar, the commonly used target detection algorithm is the mean-based CFAR algorithm based on peak selection and the target imaging algorithm on the full two-dimensional range-velocity spectrum plane. This type of algorithm can only achieve single-point target detection and parameter estimation for real physical targets, and the point cloud density is sparse, which cannot meet the point cloud detection requirements of imaging radar. Moreover, target imaging on the full two-dimensional range-velocity spectrum plane will occupy a large amount of computing resources for noise points and clutter points, greatly increasing the signal processing time, affecting the efficiency of target imaging, and increasing the demand for system storage resources, thereby increasing the hardware cost. SUMMARY

[0003] The embodiments of the present application provide a millimeter wave radar target cluster detection-based imaging method, system and storage medium, which can effectively improve the efficiency of target imaging and reduce the hardware cost of the automotive radar system.

[0004] In a first aspect, the embodiments of the present application provide a millimeter wave radar target cluster detection-based imaging method, comprising:

[0005] receiving millimeter wave radar echoes and performing Fourier transform on the millimeter wave radar echoes in the range dimension and the velocity dimension respectively to obtain radar cubic data;

[0006] performing non-coherent accumulation on the radar cubic data according to the reception channel information corresponding to the millimeter wave radar echoes to obtain a two-dimensional detection data matrix;

[0007] performing two-dimensional noise floor fitting estimation on the two-dimensional detection data matrix in the range dimension and the velocity dimension respectively to obtain a two-dimensional noise floor plane fitting matrix;

[0008] performing element-by-element comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix to generate a target detection binary matrix, and determining a target cluster of the target detection binary matrix, wherein the target cluster is an element with a value of 1 in the target detection binary matrix;

[0009] determine target coordinate information from the target detection binary matrix, obtain target radar cube data corresponding to the target coordinate information, and the target coordinate information is coordinate information corresponding to the target cluster;

[0010] based on the received channel information, sequentially performing channel correction, Doppler velocity compensation and Fourier transform processing in the angle dimension on the target radar cube data to obtain a horizontal spatial spectrum;

[0011] perform peak value detection on the horizontal spatial spectrum to obtain a target index, the target index being an index of the pitch angle corresponding to the horizontal spatial spectrum, and perform angle conversion processing on the target index to obtain a horizontal angle value;

[0012] perform target parameter conversion on the target coordinate information to obtain reference distance and reference speed values of the target cluster, and calculate reference peak intensity information of the target cluster according to the reference distance and reference speed values;

[0013] perform constraint processing on the horizontal angle value, reference distance value, reference speed value and reference peak intensity information according to preset observation range information to determine a target imaging spectrum.

[0014] In some embodiments, the radar cube data is a three-dimensional complex matrix, and a matrix size of the three-dimensional complex matrix is obtained according to a product of a first Fourier point number, a second Fourier point number and a virtual receiving channel number of a millimeter wave radar array, wherein the first Fourier point number is a Fourier point number of the three-dimensional complex matrix in the distance dimension, and the second Fourier point number is a Fourier point number of the three-dimensional complex matrix in the speed dimension.

[0015] In some embodiments, the two-dimensional noise floor plane fitting matrix is obtained by respectively performing two-dimensional noise floor fitting estimation on the two-dimensional detection data matrix in the distance dimension and the speed dimension, including:

[0016] based on the distance dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth section extraction processing and mean value processing on row data of the two-dimensional detection data matrix to obtain the first two-dimensional noise floor estimation value;

[0017] based on the speed dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth section extraction processing and mean value processing on column data of the two-dimensional detection data matrix to obtain the second two-dimensional noise floor estimation value;

[0018] the two-dimensional noise floor plane fitting matrix is constituted according to the first two-dimensional noise floor estimation value and the second two-dimensional noise floor estimation value.

[0019] In some embodiments, the one-to-one correspondence between the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix, the element alignment and comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix generates a target detection binary matrix, including:

[0020] Determining the target signal-to-noise ratio corresponding to each element in the two-dimensional noise floor fitting matrix, and comparing each target signal-to-noise ratio with the detection matrix value corresponding to the element of the two-dimensional detection data matrix;

[0021] Assigning 1 to the element position of the detection matrix value greater than the target signal-to-noise ratio, and assigning 0 to the element position of the detection matrix value less than or equal to the target signal-to-noise ratio, to obtain the target detection binary matrix.

[0022] In some embodiments, the target coordinate information is determined from the target detection binary matrix, and is obtained according to the following formula:

[0023]

[0024] Where (x, y) is the target coordinate information, w x is the element value of the target detection binary matrix, R is the distance value of the target cluster, and w x The expression of is:

[0025]

[0026] Where k MAX is the target index,

[0027] In some embodiments, before determining the target coordinate information from the target detection binary matrix, the method further includes:

[0028] Initializing the target detection binary matrix according to the observation range information and the target grid spacing, wherein the target grid spacing is obtained according to the distance resolution of the target cluster.

[0029] In a second aspect, the embodiments of the present application provide an imaging system based on millimeter wave radar target cluster detection, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to realize the imaging method based on millimeter wave radar target cluster detection as described in the first aspect.

[0030] In a third aspect, the embodiments of the present application further provide a computer readable storage medium storing computer executable instructions for executing the imaging method based on millimeter wave radar target cluster detection as described in the first aspect.

[0031] The embodiment of the present application provides an imaging method and system based on millimeter wave radar target cluster detection, and a storage medium. The method comprises the following steps: receiving millimeter wave radar echo, and performing Fourier transform on the millimeter wave radar echo in a distance dimension and a speed dimension respectively to obtain radar cubic data; performing non-coherent accumulation on the radar cubic data according to receiving channel information corresponding to the millimeter wave radar echo to obtain a two-dimensional detection data matrix; performing two-dimensional noise floor fitting estimation on the two-dimensional detection data matrix in the distance dimension and the speed dimension respectively to obtain a two-dimensional noise floor plane fitting matrix; performing element-by-element comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix to generate a target detection binary matrix, and determining a target cluster of the target detection binary matrix, wherein the target cluster is an element with a value of 1 in the target detection binary matrix; determining target coordinate information from the target detection binary matrix, obtaining target radar cubic data corresponding to the target coordinate information, and the target coordinate information is coordinate information corresponding to the target cluster; performing channel correction, Doppler velocity compensation and Fourier transform processing in an angle dimension on the target radar cubic data in sequence based on the receiving channel information to obtain a horizontal spatial spectrum; performing peak value detection on the horizontal spatial spectrum to obtain a target index, wherein the target index is an index of a pitch angle corresponding to the horizontal spatial spectrum, and performing angle conversion processing on the target index to obtain a horizontal angle value; performing target parameter conversion on the target coordinate information to obtain a reference distance value and a reference speed value of the target cluster, and calculating reference peak value intensity information of the target cluster according to the reference distance value and the reference speed value; performing constraint processing on the horizontal angle value, the reference distance value, the reference speed value and the reference peak value intensity information according to preset observation range information to determine target imaging spectrum. The embodiment of the present application utilizes a two-dimensional noise floor fitting estimation algorithm to detect and image a target cluster, greatly reduces imaging calculation of non-target points, and improves target cluster imaging efficiency and imaging quality. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 FIG. 1 is a step flowchart of an imaging method based on millimeter wave radar target cluster detection provided by an embodiment of the present application;

[0033] Figure 2 FIG. 3 is a step flowchart of obtaining a two-dimensional noise floor plane fitting matrix provided by another embodiment of the present application;

[0034] Figure 3 FIG. 4 is a step flowchart of generating a target detection binary matrix provided by another embodiment of the present application;

[0035] Figure 4is a step flow chart provided by another embodiment of the present application for initializing a target detection binary matrix;

[0036] Figure 5 is a before-and-after comparison chart provided by another embodiment of the present application for two-dimensional noise floor fitting of a two-dimensional Fourier transform RDM map under uniform noise condition;

[0037] Figure 6 is a before-and-after comparison chart provided by another embodiment of the present application for two-dimensional noise floor fitting of a two-dimensional Fourier transform RDM map under non-uniform noise condition;

[0038] Figure 7 is a real scene image provided by another embodiment of the present application;

[0039] Figure 8 is a step flow chart provided by another embodiment of the present application for Figure 7 is a comparison schematic diagram of radar imaging maps obtained based on a mean value type CFAR algorithm and a two-dimensional noise floor fitting estimation algorithm;

[0040] Figure 9 is a structure diagram of an imaging system based on millimeter wave radar target cluster detection provided by another embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0042] It can be understood that although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flow chart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the order in the flow chart. The terms "first", "second", etc. in the specification, claims or above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0043] Currently, the research and development of automobile radar system has great social and economic benefits, and with the development of millimeter wave radar technology, imaging radar has gradually become a research hotspot in the field of automatic driving. Due to the complex and changeable driving environment of the automobile, the performance of radar target detection will also be affected by the existence of background noise and interference clutter, resulting in a loss of detection performance. In the field of automobile radar, the commonly used target detection algorithm is the mean-based CFAR algorithm based on peak selection and the target imaging algorithm on the full two-dimensional range-velocity spectrum plane. This kind of algorithm can only realize single-point target detection and parameter estimation for real physical targets, and the point cloud density is sparse, which cannot meet the point cloud detection demand of imaging radar. Moreover, the target imaging on the full two-dimensional range-velocity spectrum plane will bring a large amount of noise points and clutter points to occupy the computing power, greatly increase the signal processing time, affect the efficiency of target imaging, and increase the demand for system storage resources, thereby increasing the hardware cost.

[0044] To solve the above problems, the embodiment of the present application provides a kind of based on millimeter wave radar target cluster detection imaging method, system, storage medium, method includes: receiving millimeter wave radar echo, and respectively in distance dimension and speed dimension Fourier transform is carried out to millimeter wave radar echo, obtains radar cubic data;According to the receiving channel information corresponding to the millimeter wave radar echo, the radar cubic data is incoherently accumulated, and two-dimensional detection data matrix is obtained;Respectively in the distance dimension and the speed dimension, two-dimensional noise floor fitting estimation is carried out to the two-dimensional detection data matrix, and two-dimensional noise floor plane fitting matrix is obtained;The two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix are compared between element processing, generate target detection binary matrix, and determine the target cluster of the target detection binary matrix, the target cluster is the element with value 1 in the target detection binary matrix;From the target detection binary matrix, target coordinate information is determined, the target radar cubic data corresponding to the target coordinate information is acquired, and the target coordinate information is the coordinate information corresponding to the target cluster;Based on the receiving channel information, the target radar cubic data is sequentially processed by channel correction, Doppler velocity compensation and Fourier transform in angle dimension, and horizontal spatial spectrum is obtained;The peak value of the horizontal spatial spectrum is detected, and target index is obtained, the target index is the index of the pitch angle corresponding to the horizontal spatial spectrum, and angle conversion processing is carried out to the target index, and horizontal angle value is obtained;Target parameter conversion is carried out to the target coordinate information, and reference distance value and reference speed value of the target cluster are obtained, and reference peak intensity information of the target cluster is calculated according to the reference distance value and the reference speed value;According to the preset observation range information, the horizontal angle value, reference distance value, reference speed value and reference peak intensity information are constrained, and target imaging spectrum is determined.The embodiment of the present application utilizes two-dimensional noise floor fitting estimation algorithm to detect and image target cluster, greatly reduces the imaging calculation of non-target point, improves the imaging efficiency of target cluster and improves the imaging quality, and then reduces the hardware cost of automobile radar system.

[0045] The embodiment of the present application is further described below with reference to the drawings.

[0046] As Figure 1 shown, Figure 1 It is the step flow chart of the imaging method based on millimeter wave radar target cluster detection provided by the embodiment of the present application, the embodiment of the present application provides an imaging method based on millimeter wave radar target cluster detection, the method includes but is not limited to the following steps:

[0047] Step S110, receiving millimeter wave radar echo, and respectively in distance dimension and speed dimension Fourier transform is carried out to millimeter wave radar echo, obtains radar cubic data;

[0048] In step S120, the radar cube data is non-coherently accumulated according to the receiving channel information corresponding to the millimeter wave radar echo, to obtain a two-dimensional detection data matrix;

[0049] In step S130, two-dimensional noise floor fitting estimation is performed on the two-dimensional detection data matrix in the distance dimension and the velocity dimension respectively, to obtain a two-dimensional noise floor plane fitting matrix.

[0050] In step S140, element-by-element comparison processing is performed between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix, to generate a target detection binary matrix, and a target cluster of the target detection binary matrix is determined, the target cluster being an element with a value of 1 in the target detection binary matrix.

[0051] In step S150, target coordinate information is determined from the target detection binary matrix, target radar cube data corresponding to the target coordinate information is obtained, and the target coordinate information is coordinate information corresponding to the target cluster.

[0052] In step S160, based on the receiving channel information, channel correction, Doppler velocity compensation, and Fourier transform processing in the angle dimension are sequentially performed on the target radar cube data, to obtain a horizontal spatial spectrum.

[0053] In step S170, peak value detection is performed on the horizontal spatial spectrum, to obtain a target index, the target index being an index of a pitch angle corresponding to the horizontal spatial spectrum, and angle conversion processing is performed on the target index, to obtain a horizontal angle value.

[0054] In step S180, target parameter conversion is performed on the target coordinate information, to obtain a reference distance value and a reference velocity value of the target cluster, and reference peak value intensity information of the target cluster is calculated according to the reference distance value and the reference velocity value.

[0055] In step S190, the horizontal angle value, the reference distance value, the reference velocity value, and the reference peak value intensity information are constrained according to preset observation range information, to determine a target imaging spectrum.

[0056] It should be noted that the radar cube data in the embodiments of the present application is a three-dimensional complex matrix, the matrix size of the three-dimensional complex matrix is obtained according to the product of a first Fourier point number Nrange, a second Fourier point number Mdoppler, and a virtual receiving channel number Nrx of the millimeter wave radar array, and the expression of the matrix size of the three-dimensional complex matrix is Nrange*Mdoppler*Nrx, wherein the first Fourier point number Nrange is a Fourier point number of the three-dimensional complex matrix in the distance dimension, and the second Fourier point number Mdoppler is a Fourier point number of the three-dimensional complex matrix in the velocity dimension.

[0057] It should be noted that in the step of determining the two-dimensional detection data matrix, some or all channels can be selected for incoherent accumulation based on the characteristics of different receiving channels corresponding to the millimeter-wave radar echo and the computing power constraints of the radar system. No further restrictions will be imposed here.

[0058] It should be noted that the two-dimensional noise floor fitting in the embodiments of this application includes both uniform and non-uniform noise floor cases, as referred to... Figure 5 , Figure 5 This is a comparison image before and after performing two-dimensional noise floor fitting on the RDM image after two-dimensional Fourier transform under uniform noise conditions, provided by another embodiment of this application; it can be seen that the noise floor distribution of the radar RDM image obtained after two-dimensional noise floor fitting is more uniform; Reference Figure 6 , Figure 6 This is a comparison diagram before and after performing two-dimensional noise floor fitting on the RDM map after two-dimensional Fourier transform under non-uniform noise conditions provided by another embodiment of this application. It can be understood that due to radar movement, the presence of strong targets near the Doppler cell corresponding to vehicle speed affects the noise floor, resulting in a rise in the noise floor in the two-dimensional plane. The noise floor after two-dimensional noise floor fitting is consistent with the fluctuation of the RDM plane.

[0059] It should be noted that the target coordinate information is determined from the target detection binary matrix using the following formula:

[0060]

[0061] Where (x,y) represents the target coordinates, w x Let w be the element value of the binary matrix for target detection, R be the distance value of the target cluster, and w be the distance value of the target cluster. x The expression is:

[0062]

[0063] Where, k MAX The target index is the index of the pitch angle corresponding to the horizontal spatial spectrum, that is, the peak index of the angle-dimensional Fourier spectrum. Specifically, the Fourier spectrum is constructed based on the target signal value sequence. In this embodiment, it is assumed that the antenna array of the radar system is an equivalent 8-element linear array with two transmit and four receive antennas. The expression for the target signal value sequence corresponding to the 8 virtual antennas after channel correction and compensation is as follows:

[0064]

[0065] Where A1 and ψ are the initial amplitude and initial phase at the first antenna in the antenna array, respectively, and the Fourier spectrum corresponding to the target signal value sequence will be in w x A peak value is generated at that location.

[0066] It can be understood that the millimeter wave radar echo is received, and the millimeter wave radar echo is subjected to Fourier transform in the distance dimension and the speed dimension respectively to obtain radar cubic data; the radar cubic data is subjected to non-coherent accumulation according to the receiving channel information corresponding to the millimeter wave radar echo to obtain a two-dimensional detection data matrix; the two-dimensional detection data matrix is subjected to two-dimensional noise floor fitting estimation in the distance dimension and the speed dimension respectively to obtain a two-dimensional noise floor plane fitting matrix; an element-by-element comparison process is performed between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix to generate a target detection binary matrix, and a target cluster of the target detection binary matrix is determined, the target cluster being an element with a value of 1 in the target detection binary matrix, target cluster detection can retain the size information of a target, complete fine imaging of the target, and help subsequent realization of target recognition and classification; target coordinate information is determined from the target detection binary matrix, target radar cubic data corresponding to the target coordinate information is acquired, the target coordinate information being coordinate information corresponding to the target cluster, that is, coordinate information corresponding to a non-0 element in the target detection binary matrix; the target radar cubic data is subjected to channel correction, Doppler velocity compensation and Fourier transform processing in the angle dimension in sequence based on the receiving channel information to obtain a horizontal spatial spectrum; peak value detection is performed on the horizontal spatial spectrum to obtain a target index, the target index being an index of a pitch angle corresponding to the horizontal spatial spectrum, and angle conversion processing is performed on the target index to obtain a horizontal angle value; target parameter conversion is performed on the target coordinate information to obtain a reference distance value and a reference speed value of the target cluster, and reference peak value intensity information of the target cluster is calculated according to the reference distance value and the reference speed value; the horizontal angle value, the reference distance value, the reference speed value and the reference peak value intensity information are subjected to constraint processing according to preset observation range information to determine target imaging spectrum, wherein the observation range information can be set according to the index of a radar system, and is not limited herein. The embodiments of the present application utilize a two-dimensional noise floor fitting estimation algorithm to detect and image a target cluster, compared with a mean value CFAR algorithm, the two-dimensional noise floor fitting estimation algorithm can retain more fine features of a real target, while avoiding a large amount of repeated calculation of sliding window processing and reducing signal processing time; compared with a full plane point cloud imaging method, the present application greatly reduces imaging calculation of non-target points, realizes two-dimensional adaptive screening of a real target, guarantees the detection probability of the target, that is, improves the imaging efficiency of the target cluster while improving the imaging quality, and then reduces the hardware cost of a car radar system.

[0067] It can be understood that the reference Figure 7 and Figure 8 , Figure 7 is a measured scene image provided by another embodiment of the present application, Figure 8is provided in another embodiment of the present application Figure 7 Based on the mean-based CFAR algorithm and the two-dimensional noise floor fitting estimation algorithm, the contrast diagram of the radar imaging diagram obtained by the two algorithms is shown in FIG. 6. Based on the mean-based CFAR algorithm, the imaging diagram of the scene is shown in FIG. 7, and based on the two-dimensional noise floor fitting estimation algorithm, the imaging diagram of the scene is shown in FIG. 8. Figure 7 The scene is a row of metal guardrails, and the imaging diagram of the scene is shown in FIG. 9. Figure 8 As can be seen from FIG. 9, the imaging effect diagram of the present application can restore the imaging information of the guardrail target more finely.

[0068] In addition, in some embodiments, the reference Figure 2 , Figure 1 The step S130 in the embodiment shown includes but is not limited to the following steps:

[0069] Step S210, based on the distance dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth section extraction processing and mean value processing on the row data of the two-dimensional detection data matrix, to obtain a first two-dimensional noise floor estimation value;

[0070] Step S220, based on the speed dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth section extraction processing and mean value processing on the column data of the two-dimensional detection data matrix, to obtain a second two-dimensional noise floor estimation value;

[0071] Step S230, constructing a two-dimensional noise floor plane fitting matrix according to the first two-dimensional noise floor estimation value and the second two-dimensional noise floor estimation value.

[0072] It can be understood that the two-dimensional noise floor fitting estimation is to perform numerical sorting processing, smooth section extraction processing, mean value processing and other operations on the row data corresponding to the i-th distance dimension range bin and the column data corresponding to the j-th speed dimension doppler bin, respectively, to obtain the current first two-dimensional noise floor estimation value and the second two-dimensional noise floor estimation value, and to traverse the entire two-dimensional detection data matrix, i.e. the RDM plane, to obtain a two-dimensional noise floor estimation matrix.

[0073] In addition, the elements between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix are one-to-one corresponding, and the reference Figure 3 In an embodiment, Figure 1 The step S140 in the embodiment shown includes but is not limited to the following steps:

[0074] Step S310, determining the target signal-to-noise ratio corresponding to each element in the two-dimensional noise floor plane fitting matrix, and comparing each target signal-to-noise ratio with the detection matrix value corresponding to the element of the two-dimensional detection data matrix;

[0075] Step S320, assigning the element position with a value of 1 if the detection matrix value is greater than the target signal-to-noise ratio, and assigning the element position with a value of 0 if the detection matrix value is less than or equal to the target signal-to-noise ratio, to obtain a target detection binary matrix.

[0076] It should be noted that, since the elements of the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix are one-to-one corresponding, the step of performing element comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix is as follows: superimposing the minimum detectable signal-to-noise ratio SNRmin, i.e., the target signal-to-noise ratio, in the two-dimensional noise floor fitting matrix, and then performing numerical comparison, assigning the element position with a matrix value greater than the target signal-to-noise ratio as 1, and assigning the element position with a detection matrix value less than or equal to the target signal-to-noise ratio as 0, to obtain a target detection binary matrix.

[0077] In addition, before step S150 is performed, the imaging method based on millimeter wave radar target cluster detection of the embodiment of the application further includes but is not limited to the following steps: Figure 1

[0078] In step S410, the target detection binary matrix is initialized according to the observation range information and the target grid spacing, wherein the target grid spacing is obtained according to the distance resolution of the target cluster.

[0079] It can be understood that initializing the target detection binary matrix can improve the generalization ability of the subsequent target detection binary matrix, and thus improve the imaging quality of the millimeter wave radar target cluster detection.

[0080] In addition, with reference to Figure 9 , Figure 9 FIG. 1 is a structural diagram of an imaging system based on millimeter wave radar target cluster detection according to another embodiment of the application. One embodiment of the application further provides an imaging system 500 based on millimeter wave radar target cluster detection, which includes a memory 510, a processor 520, and a computer program stored in the memory 510 and executable on the processor 520.

[0081] The processor 520 and the memory 510 can be connected through a bus or other means.

[0082] The non-transitory software program and instructions required for implementing the imaging method based on millimeter wave radar target cluster detection of the above-mentioned embodiments are stored in the memory 510, and when executed by the processor 520, the imaging method based on millimeter wave radar target cluster detection applied to the imaging system 500 based on millimeter wave radar target cluster detection in the above-mentioned embodiments is executed, for example, the method steps S110 to S190 in Figure 1 , the method steps S210 to S230 in Figure 2 , the method steps S310 to S320 in Figure 3 , and the method step S410 in Figure 4 .

[0083] ​The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0084] Furthermore, one embodiment of this application provides a computer-readable storage medium storing computer-executable instructions that are executed by a processor or controller, for example, by a processor 520 in the above-described embodiment of the millimeter-wave radar target cluster detection imaging system 500. The processor 520 can then execute the millimeter-wave radar target cluster detection imaging method applied to the millimeter-wave radar target cluster detection imaging system 500 as described above, for example, performing the above-described... Figure 1 Method steps S110 to S190, Figure 2 Method steps S210 to S230, Figure 3 Method steps S310 to S320 and Figure 4 Method step S410. Those skilled in the art will understand that all or some of the steps in the methods disclosed above, and the system, can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

Claims

1. An imaging method based on millimeter wave radar target cluster detection, characterized in that, The method comprises the following steps: receiving millimeter wave radar echoes, and performing Fourier transform on the millimeter wave radar echoes in the distance dimension and the speed dimension respectively to obtain radar cube data; performing non-coherent accumulation on the radar cube data according to the receiving channel information corresponding to the millimeter wave radar echoes to obtain a two-dimensional detection data matrix; performing two-dimensional noise floor fitting estimation on the two-dimensional detection data matrix in the distance dimension and the speed dimension respectively to obtain a two-dimensional noise floor plane fitting matrix; performing element-by-element comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor plane fitting matrix to generate a target detection binary matrix, and determining a target cluster of the target detection binary matrix, the target cluster being an element with a value of 1 in the target detection binary matrix; determining target coordinate information from the target detection binary matrix, obtaining target radar cube data corresponding to the target coordinate information, and the target coordinate information being coordinate information corresponding to the target cluster; based on the receiving channel information, sequentially performing channel correction, Doppler velocity compensation and Fourier transform processing in the angle dimension on the target radar cube data to obtain a horizontal spatial spectrum; performing peak value detection on the horizontal spatial spectrum to obtain a target index, the target index being an index of the pitch angle corresponding to the horizontal spatial spectrum, and performing angle conversion processing on the target index to obtain a horizontal angle value; performing target parameter conversion on the target coordinate information to obtain reference distance values and reference speed values of the target cluster, and calculating reference peak intensity information of the target cluster according to the reference distance values and the reference speed values; performing constraint processing on the horizontal angle value, the reference distance value, the reference speed value and the reference peak intensity information according to preset observation range information to determine a target imaging spectrum.

2. The method of imaging based on millimeter wave radar target cluster detection of claim 1, wherein, The radar cube data is a three-dimensional complex matrix, and the matrix size of the three-dimensional complex matrix is obtained according to the product of a first Fourier point number, a second Fourier point number and a virtual receiving channel number of a millimeter wave radar array, wherein the first Fourier point number is the Fourier point number of the three-dimensional complex matrix in the distance dimension, and the second Fourier point number is the Fourier point number of the three-dimensional complex matrix in the speed dimension.

3. The method of imaging based on millimeter wave radar target cluster detection of claim 1, wherein, The two-dimensional noise floor fitting estimation on the two-dimensional detection data matrix in the distance dimension and the speed dimension respectively to obtain a two-dimensional noise floor plane fitting matrix comprises: based on the distance dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth segment extraction processing and mean value processing on the row data of the two-dimensional detection data matrix to obtain a first two-dimensional noise floor estimation value; based on the speed dimension, traversing the two-dimensional detection data matrix, and sequentially performing numerical sorting processing, smooth segment extraction processing and mean value processing on the column data of the two-dimensional detection data matrix to obtain a second two-dimensional noise floor estimation value; the two-dimensional noise floor plane fitting matrix is constructed according to the first two-dimensional noise floor estimation value and the second two-dimensional noise floor estimation value.

4. The method of imaging based on millimeter wave radar target cluster detection of claim 1, wherein, The elements of the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix are one-to-one corresponding, and the alignment element comparison processing between the two-dimensional detection data matrix and the two-dimensional noise floor fitting matrix generates a target detection binary matrix, including: Determining a target signal-to-noise ratio corresponding to each element in the two-dimensional noise floor fitting matrix, and comparing each target signal-to-noise ratio with a detection matrix value corresponding to an element of the two-dimensional detection data matrix; Assigning 1 to the element position of the detection matrix value greater than the target signal-to-noise ratio, and assigning 0 to the element position of the detection matrix value less than or equal to the target signal-to-noise ratio, to obtain the target detection binary matrix.

5. The method of imaging based on millimeter wave radar target cluster detection of claim 1, wherein, The target coordinate information is determined from the target detection binary matrix, and is obtained according to the following formula: ； wherein, is the target coordinate information, is the element value of the target detection binary matrix, is the distance value of the target cluster, the expression of is: ； wherein for the target index, .

6. The method of imaging based on millimeter wave radar target cluster detection of claim 1, wherein, Before the target coordinate information is determined from the target detection binary matrix, the method further includes: Initializing the target detection binary matrix according to the observation range information and a target grid point spacing, wherein the target grid point spacing is obtained according to a distance resolution of the target cluster.

7. An imaging system based on millimeter wave radar target cluster detection, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the imaging method based on millimeter wave radar target cluster detection according to any one of claims 1 to 6 when executing the computer program. 8.A computer readable storage medium storing computer executable instructions for performing the imaging method based on millimeter wave radar target cluster detection according to any one of claims 1 to 6.

Images