Personnel perception quantity statistical method and system based on millimeter wave radar

Abstract

The invention discloses a personnel perception quantity statistical method and system based on a millimeter-wave radar, and relates to the technical field of millimeter-wave radar signal processing. The method comprises the following steps: receiving a detection echo signal by using a millimeter wave radar, performing ADC sampling on the detection echo signal to convert the detection echo signal into a digital signal, and processing the digital signal by using fast Fourier transform to obtain a millimeter wave echo sampling point frequency domain distribution matrix; and then, analyzing the frequency domain distribution matrix of the millimeter wave echo sampling points by using a data processing and analysis algorithm based on artificial intelligence and deep learning, and carrying out clustering analysis based on the frequency domain distribution characteristics of the key signals, so as to take the number of clustering centers in a clustering analysis result as the total number of personnel statistics. According to the invention, the data processing algorithm based on artificial intelligence and deep learning is utilized, the dependence on specific parameter setting can be reduced, and the robustness and reliability of the system are improved.

Description

Technical Field

[0001] The present application relates to the technical field of millimeter wave radar signal processing, and in particular, to a method and system for counting the number of people sensing based on millimeter wave radar. Background Art

[0002] Millimeter wave radar transmits linear frequency modulated continuous wave (LFM) signals and uses the Doppler effect and the time delay of the echo signal to obtain the distance and speed information of the target object. Using linear frequency modulated continuous wave radar to perceive and count the number of people can achieve real-time perception and statistics of people, which is not affected by environmental factors such as light, smoke, and humidity.

[0003] Chinese patent, which can detect the information of people in the room through linear frequency modulation continuous wave radar, and process this information in real time through FPGA to realize the detection and counting of indoor people. This indoor personnel statistical counting method uses linear frequency modulation continuous wave radar to transmit and receive echo signals, and obtains digital signals by sampling echo signals through ADC. The PL end of FPGA performs FFT processing on these signals, and then passes the data to the PS end through DMA for constant false alarm rate detection (CFAR), point classification and condensation processing, and finally realizes target discrimination. The whole system is accurate, fast and real-time, and can report the results to the host computer through Ethernet and UDP protocols. It is suitable for smart offices or conference rooms, and supports functions such as power off when people leave and lights on when people come.

[0004] However, in the above-mentioned FPGA-based millimeter-wave radar indoor personnel counting method, the point classification and agglomeration algorithm rely on specific parameter settings, such as the number of protection units, the number of reference units, etc. The selection of these parameters is crucial to the effectiveness of the algorithm. Inappropriate parameter selection may lead to false alarms or missed alarms. In addition, the target discrimination process depends on the set distance and speed thresholds. If these thresholds are set improperly, misjudgments may occur. For example, if the distance threshold is too narrow, a distant human target may be missed; if the speed threshold is too wide, a non-human target may be mistaken for a human.

[0005] Therefore, an optimized millimeter-wave radar-based personnel perception and counting solution is needed. Summary of the invention

[0006] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0007] In a first aspect, the present application provides a method for counting the number of people sensed based on millimeter wave radar, the method comprising:

[0008] Obtain millimeter wave radar detection echo signal;

[0009] Performing ADC sampling on the millimeter-wave radar detection echo signal to obtain a millimeter-wave echo sampling digital signal;

[0010] Performing a fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a frequency domain distribution matrix of millimeter wave echo sampling points;

[0011] Extracting frequency domain features of the echo signal from the frequency domain distribution matrix of the millimeter wave echo sampling points to obtain a frequency domain distribution feature graph of the millimeter wave echo signal;

[0012] Inputting the millimeter wave echo signal frequency domain distribution feature map into a feature saliency engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain saliency distribution feature map;

[0013] A cluster analysis is performed on the frequency domain significant distribution feature diagram of the millimeter wave echo signal to determine the total number of personnel statistics.

[0014] Optionally, the frequency domain features of the echo signals are extracted from the frequency domain distribution matrix of the millimeter wave echo sampling points to obtain a frequency domain distribution feature map of the millimeter wave echo signals, including: inputting the frequency domain distribution matrix of the millimeter wave echo sampling points into an echo signal frequency domain feature extractor based on a hole convolutional neural network model to obtain the frequency domain distribution feature map of the millimeter wave echo signals.

[0015] Optionally, the frequency domain distribution feature map of the millimeter wave echo signal is input into a feature saliency engine based on feature distribution gradient to obtain a frequency domain saliency distribution feature map of the millimeter wave echo signal, including: calculating the multi-directional gradient value distribution of each position in the frequency domain distribution feature map of the millimeter wave echo signal, and determining the gradient amplitude value of each position in the frequency domain distribution feature map of the millimeter wave echo signal based on the multi-directional gradient value distribution of each position to obtain the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal; calculating the gradient amplitude value of each position in the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal A local description operator of the gradient amplitude at each position is used to obtain a local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal; the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal is input into a gated masker based on the GELU function to obtain a local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal; and the position point multiplication between the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal and the frequency domain distribution feature map of the millimeter-wave echo signal is calculated to obtain the frequency domain significant distribution feature map of the millimeter-wave echo signal.

[0016] Optionally, the multi-directional gradient value distribution of each position in the millimeter-wave echo signal frequency domain distribution characteristic map is calculated, and the gradient amplitude value of each position in the millimeter-wave echo signal frequency domain distribution characteristic map is determined based on the multi-directional gradient value distribution of each position to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic map, including: calculating the gradient amplitude values ​​of the millimeter-wave echo signal frequency domain distribution characteristic map along different directions to obtain the millimeter-wave echo signal frequency domain height gradient distribution characteristic map, the millimeter-wave echo signal frequency domain width gradient distribution characteristic map and the millimeter-wave echo signal frequency domain channel gradient distribution characteristic map; calculating the binary norm between the characteristic values ​​of corresponding positions in the millimeter-wave echo signal frequency domain height gradient distribution characteristic map, the millimeter-wave echo signal frequency domain width gradient distribution characteristic map and the millimeter-wave echo signal frequency domain channel gradient distribution characteristic map to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic map.

[0017] Optionally, the gradient amplitude values ​​of the millimeter-wave echo signal frequency domain distribution characteristic map are calculated along different directions to obtain a millimeter-wave echo signal frequency domain height gradient distribution characteristic map, a millimeter-wave echo signal frequency domain width gradient distribution characteristic map and a millimeter-wave echo signal frequency domain channel gradient distribution characteristic map, including: calculating the gradient amplitude values ​​of the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic map along the height direction to obtain the millimeter-wave echo signal frequency domain height gradient distribution characteristic map; calculating the gradient amplitude values ​​of the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic map along the width direction to obtain the millimeter-wave echo signal frequency domain width gradient distribution characteristic map; calculating the gradient amplitude values ​​of the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic map along the channel direction to obtain the millimeter-wave echo signal frequency domain channel gradient distribution characteristic map.

[0018] Optionally, the local description operator of the gradient amplitude at each position in the frequency domain gradient amplitude distribution feature map of the millimeter-wave echo signal is calculated to obtain the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal, including: determining the local neighborhood range of the gradient amplitude, and calculating a sequence of the frequency domain gradient amplitude distribution neighborhood feature maps of the millimeter-wave echo signal based on the local neighborhood range of the gradient amplitude and the frequency domain gradient amplitude distribution feature map of the millimeter-wave echo signal; counting the number of feature maps in the sequence of the frequency domain gradient amplitude distribution neighborhood feature maps of the millimeter-wave echo signal to obtain the statistical value of the number of frequency domain gradient amplitude distribution neighborhoods of the millimeter-wave echo signal; calculating the frequency domain gradient amplitude distribution feature map of the millimeter-wave echo signal and the millimeter-wave echo signal. The position differences between each millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood feature map in the sequence of meter-wave echo signal frequency domain gradient amplitude distribution neighborhood feature map are divided by the millimeter-wave echo signal frequency domain gradient amplitude distribution feature map by position point to obtain a sequence of millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature maps; the position points are added for each millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature map in the sequence of millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature map, and the resultant addition is divided by the millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood number statistics by position point to obtain the millimeter-wave echo signal frequency domain gradient amplitude local significant distribution feature map.

[0019] Optionally, cluster analysis is performed on the millimeter wave echo signal frequency domain significant distribution feature map to determine the total number of personnel statistics, including: cluster analysis is performed on the millimeter wave echo signal frequency domain significant distribution feature map to obtain a cluster analysis result; and the number of cluster centers in the cluster analysis result is used as the total number of personnel statistics.

[0020] In a second aspect, the present application provides a personnel perception quantity counting system based on millimeter wave radar, the system comprising:

[0021] A radar detection echo signal acquisition module is used to acquire millimeter wave radar detection echo signals;

[0022] An ADC sampling module is used to perform ADC sampling on the millimeter wave radar detection echo signal to obtain a millimeter wave echo sampling digital signal;

[0023] A fast Fourier transform module, used for performing a fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a frequency domain distribution matrix of the millimeter wave echo sampling points;

[0024] An echo signal frequency domain feature extraction module is used to extract the echo signal frequency domain features from the millimeter wave echo sampling point frequency domain distribution matrix to obtain a millimeter wave echo signal frequency domain distribution feature map;

[0025] A feature distribution gradient significant module, used for inputting the millimeter wave echo signal frequency domain distribution feature map into a feature significant engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain significant distribution feature map;

[0026] The cluster analysis module is used to perform cluster analysis on the significant distribution feature diagram of the millimeter wave echo signal frequency domain to determine the total number of personnel statistics.

[0027] Optionally, the echo signal frequency domain feature extraction module is used to: input the millimeter wave echo sampling point frequency domain distribution matrix into an echo signal frequency domain feature extractor based on a hole convolutional neural network model to obtain the millimeter wave echo signal frequency domain distribution feature map.

[0028] Optionally, the feature distribution gradient significance module includes: a multi-directional gradient value calculation unit, which is used to calculate the multi-directional gradient value distribution of each position in the millimeter wave echo signal frequency domain distribution feature map, and determine the gradient amplitude value of each position in the millimeter wave echo signal frequency domain distribution feature map based on the multi-directional gradient value distribution of each position to obtain the millimeter wave echo signal frequency domain gradient amplitude distribution feature map; a gradient amplitude local description operator calculation unit, which is used to calculate the gradient amplitude local description operator of each position in the millimeter wave echo signal frequency domain gradient amplitude distribution feature map to obtain the millimeter wave echo signal frequency domain gradient amplitude distribution feature map. A local significant distribution feature map of the frequency domain gradient amplitude of a millimeter-wave echo signal is obtained; a gated mask unit is used to input the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal into a gated masker based on a GELU function to obtain a local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal; a position point multiplication calculation unit is used to calculate the position point multiplication between the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal and the frequency domain distribution feature map of the millimeter-wave echo signal to obtain the frequency domain significant distribution feature map of the millimeter-wave echo signal.

[0029] The present application has at least the following technical effects: Compared with the prior art, the present application uses a millimeter-wave radar to receive a detection echo signal, performs ADC sampling on the detection echo signal to convert it into a digital signal, and then uses a fast Fourier transform to process it to obtain a frequency domain distribution matrix of the millimeter-wave echo sampling points. Then, a data processing and analysis algorithm based on artificial intelligence and deep learning is used to analyze the frequency domain distribution matrix of the millimeter-wave echo sampling points, and a cluster analysis is performed based on the frequency domain distribution characteristics of the key signal, so that the number of cluster centers in the cluster analysis results is used as the total number of personnel statistics. The present application uses a data processing algorithm based on artificial intelligence and deep learning, which can reduce dependence on specific parameter settings and improve the robustness and reliability of the system.

[0030] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0032] Figure 1 The present invention is a flowchart of a method for counting the number of people sensed based on millimeter-wave radar according to an exemplary embodiment.

[0033] Figure 2 The figure is a block diagram of a personnel sensing and counting system based on millimeter-wave radar according to an exemplary embodiment. DETAILED DESCRIPTION

[0034] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0035] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0036] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0037] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0038] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0039] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0040] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0041] Figure 1 is a flow chart of a method for counting the number of people sensed based on millimeter wave radar according to an exemplary embodiment, such as Figure 1 As shown, the method includes:

[0042] Step S101, obtaining a millimeter wave radar detection echo signal;

[0043] Step S102, performing ADC sampling on the millimeter wave radar detection echo signal to obtain a millimeter wave echo sampling digital signal;

[0044] Step S103, performing fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a millimeter wave echo sampling point frequency domain distribution matrix;

[0045] Step S104, performing echo signal frequency domain feature extraction on the millimeter wave echo sampling point frequency domain distribution matrix to obtain a millimeter wave echo signal frequency domain distribution feature graph;

[0046] Step S105, inputting the millimeter wave echo signal frequency domain distribution feature map into a feature saliency engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain saliency distribution feature map;

[0047] Step S106: performing cluster analysis on the millimeter wave echo signal frequency domain significant distribution feature graph to determine the total number of personnel statistics.

[0048] The above-mentioned method for counting the number of people perceived based on millimeter wave radar can use the millimeter wave radar to receive the detection echo signal, and after ADC sampling the detection echo signal to convert it into a digital signal, it uses the fast Fourier transform to process it to convert the digital signal of the detection echo signal from the time domain to the frequency domain to obtain the frequency domain distribution matrix of the millimeter wave echo sampling point, so as to analyze the frequency components of the signal. Then, at the back end, the data processing and analysis algorithm based on artificial intelligence and deep learning is used to analyze the frequency domain distribution matrix of the millimeter wave echo sampling point, so as to learn and capture the key signal frequency domain distribution characteristics in the frequency domain distribution matrix of the millimeter wave echo sampling point, and cluster analysis is performed based on the key signal frequency domain distribution characteristics, so as to use the number of cluster centers in the cluster analysis results as the total number of people counted, which can realize more intelligent personnel perception quantity statistics, and the data processing algorithm based on artificial intelligence and deep learning can reduce the dependence on specific parameter settings and improve the robustness and reliability of the system.

[0049] Specifically, in the technical solution of the present application, first, a millimeter wave radar detection echo signal is obtained. The millimeter wave radar detection echo signal contains information about the distance and speed of objects perceived by the millimeter wave radar, which is helpful for personnel perception and quantity statistics. Then, the millimeter wave radar detection echo signal is sampled by ADC to convert the analog signal into a millimeter wave echo sampling digital signal.

[0050] Then, the millimeter-wave echo sampling digital signal is subjected to a fast Fourier transform to convert the millimeter-wave echo sampling digital signal from the time domain to the frequency domain, so as to better analyze and extract the frequency component characteristics of the signal, thereby obtaining a frequency domain distribution matrix of the millimeter-wave echo sampling points containing distance and Doppler information.

[0051] In the process of processing the echo signal detected by the millimeter wave radar, the frequency domain distribution matrix of the millimeter wave echo sampling point is the frequency component information of the millimeter wave echo containing distance and Doppler semantics obtained by fast Fourier transform processing, which reflects the distance and speed dynamic information of different target objects. In order to better extract the relationship between the frequency components of the millimeter wave echo signal in the frequency domain distribution matrix of the millimeter wave echo sampling point, so as to help identify the existence and characteristics of different personnel, thereby providing a basis for the subsequent personnel counting task, in the technical solution of the present application, the frequency domain distribution matrix of the millimeter wave echo sampling point is further input into the echo signal frequency domain feature extractor based on the hole convolution neural network model to obtain the frequency domain distribution feature map of the millimeter wave echo signal. Through the processing of the echo signal frequency domain feature extractor based on the hole convolution neural network model, the frequency domain distribution features and the implicit correlation between the frequency components can be extracted from the frequency domain distribution matrix of the millimeter wave echo sampling point, which is helpful to learn and capture the features related to the position and movement of personnel, so as to facilitate the identification of the existence of personnel and the dynamic information of personnel, and provide a basis for the subsequent clustering tasks and personnel counting tasks.

[0052] In one embodiment of the present application, the frequency domain features of the echo signal are extracted from the frequency domain distribution matrix of the millimeter wave echo sampling points to obtain a frequency domain distribution feature map of the millimeter wave echo signal, including: inputting the frequency domain distribution matrix of the millimeter wave echo sampling points into an echo signal frequency domain feature extractor based on a hole convolutional neural network model to obtain the frequency domain distribution feature map of the millimeter wave echo signal.

[0053] Specifically, since the frequency domain distribution feature map of the millimeter wave echo signal contains the correlation between the frequency domain distribution features of the echo signal and the frequency components, it is used to identify the location and dynamic characteristics of personnel. However, the frequency domain distribution feature map of the millimeter wave echo signal not only contains frequency domain distribution features related to personnel detection and counting tasks, but also contains some irrelevant noise interference and redundant information, which contributes less to subsequent personnel statistics and counting tasks. Therefore, in order to enhance the key information expression capability of the frequency domain distribution feature map of the millimeter wave echo signal, in the technical solution of the present application, the frequency domain distribution feature map of the millimeter wave echo signal is further input into a feature saliency engine based on feature distribution gradient to obtain a frequency domain saliency distribution feature map of the millimeter wave echo signal. Specifically, the feature saliency engine based on feature distribution gradient is a technology used to enhance the expressiveness of feature maps in deep learning models. Its main function is to highlight the important frequency domain distribution feature information related to personnel recognition and counting tasks in the frequency domain distribution feature map of the millimeter wave echo signal through feature gradient amplitude measurement and gated mask mechanism, thereby helping to more accurately identify the existence and characteristics of different personnel, and provide a more accurate basis for subsequent personnel counting tasks.

[0054] In one embodiment of the present application, the frequency domain distribution feature map of the millimeter wave echo signal is input into a feature saliency engine based on feature distribution gradient to obtain a frequency domain saliency distribution feature map of the millimeter wave echo signal, including: calculating the multi-directional gradient value distribution of each position in the frequency domain distribution feature map of the millimeter wave echo signal, and determining the gradient amplitude value of each position in the frequency domain distribution feature map of the millimeter wave echo signal based on the multi-directional gradient value distribution of each position to obtain the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal; calculating the frequency domain gradient amplitude distribution feature of the millimeter wave echo signal A local description operator of the gradient amplitude at each position in the feature map is used to obtain a local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal; the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal is input into a gated masker based on the GELU function to obtain a local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal; and the position point multiplication between the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal and the frequency domain distribution feature map of the millimeter-wave echo signal is calculated to obtain the frequency domain significant distribution feature map of the millimeter-wave echo signal.

[0055] Specifically, first, the multi-directional gradient value distribution of each position in the frequency domain distribution feature map of the millimeter wave echo signal is calculated, which helps to understand the spatial relationship between the frequency components in different directions, especially the relative position and motion state of different objects or people in different directions. And, based on the multi-directional gradient value distribution of each position, the gradient amplitude value of each position is determined. The gradient amplitude value reflects the local change intensity between the frequency components at each position in the feature map, which helps to identify which parts of the feature map have more significant changes. Next, the local description operator of the gradient amplitude of each position in the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal is calculated. The local description operator can further highlight the significant areas of the gradient amplitude, that is, those positions with greater change intensity. This local saliency analysis helps to enhance the key features related to personnel target recognition and motion information in the feature map, so that the model can pay more attention to these significant areas, thereby improving the accuracy of recognition and personnel counting tasks. Then, using the nonlinear characteristics of the GELU function, the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter wave echo signal is input into the gated masker. In this way, the local significant distribution feature map can be adaptively weighted. The activation of the GELU function further strengthens the significant change area, while the non-significant area is suppressed. This helps the model focus more on key features related to personnel recognition and dynamic changes, so as to ensure that the key features related to the subsequent personnel counting task are retained, while the non-key features are weakened. Finally, the obtained local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal is multiplied by position with the initial frequency domain distribution feature map of the millimeter wave echo signal. This step is equivalent to weighting and masking the original feature map, so that the significant feature areas in the feature map with higher correlation with the subsequent tasks are retained and enhanced, while the non-significant feature areas are suppressed, which improves the sensitivity and generalization ability of the model to the input data. In particular, the feature saliency engine based on the feature distribution gradient realizes the transformation from the initial feature map to the optimized feature map. The optimized millimeter wave echo signal frequency domain significant distribution feature map has stronger expressive power, because it not only contains the original echo signal frequency domain distribution feature information, but also incorporates the gradient significance information. This helps to improve the model's ability to identify key detection information and personnel, especially in the millimeter wave radar scenario, and can more accurately identify the presence and characteristics of different personnel, providing a more accurate basis for subsequent personnel counting tasks.

[0056] In one embodiment of the present application, the multi-directional gradient value distribution of each position in the millimeter-wave echo signal frequency domain distribution characteristic map is calculated, and the gradient amplitude value of each position in the millimeter-wave echo signal frequency domain distribution characteristic map is determined based on the multi-directional gradient value distribution of each position to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic map, including: calculating the gradient amplitude values ​​of the millimeter-wave echo signal frequency domain distribution characteristic map along different directions to obtain the millimeter-wave echo signal frequency domain height gradient distribution characteristic map, the millimeter-wave echo signal frequency domain width gradient distribution characteristic map and the millimeter-wave echo signal frequency domain channel gradient distribution characteristic map; calculating the binary norm between the characteristic values ​​of corresponding positions in the millimeter-wave echo signal frequency domain height gradient distribution characteristic map, the millimeter-wave echo signal frequency domain width gradient distribution characteristic map and the millimeter-wave echo signal frequency domain channel gradient distribution characteristic map to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic map.

[0057] Furthermore, in one embodiment of the present application, the gradient amplitude values ​​along different directions of the millimeter-wave echo signal frequency domain distribution characteristic diagram are calculated to obtain a millimeter-wave echo signal frequency domain height gradient distribution characteristic diagram, a millimeter-wave echo signal frequency domain width gradient distribution characteristic diagram and a millimeter-wave echo signal frequency domain channel gradient distribution characteristic diagram, including: calculating the gradient amplitude values ​​along the height direction for the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic diagram to obtain the millimeter-wave echo signal frequency domain height gradient distribution characteristic diagram; calculating the gradient amplitude values ​​along the width direction for the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic diagram to obtain the millimeter-wave echo signal frequency domain width gradient distribution characteristic diagram; calculating the gradient amplitude values ​​along the channel direction for the characteristic values ​​at each position in the millimeter-wave echo signal frequency domain distribution characteristic diagram to obtain the millimeter-wave echo signal frequency domain channel gradient distribution characteristic diagram.

[0058] Furthermore, in one embodiment of the present application, the local description operator of the gradient amplitude at each position in the frequency domain gradient amplitude distribution feature map of the millimeter-wave echo signal is calculated to obtain the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal, including: determining the local neighborhood range of the gradient amplitude, and calculating a sequence of the frequency domain gradient amplitude distribution neighborhood feature maps of the millimeter-wave echo signal based on the local neighborhood range of the gradient amplitude and the frequency domain gradient amplitude distribution feature map of the millimeter-wave echo signal; counting the number of feature maps in the sequence of the frequency domain gradient amplitude distribution neighborhood feature maps of the millimeter-wave echo signal to obtain the statistical value of the number of frequency domain gradient amplitude distribution neighborhoods of the millimeter-wave echo signal; calculating the frequency domain gradient amplitude distribution of the millimeter-wave echo signal The positional differences between the characteristic graph and each millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood characteristic graph in the sequence of the millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood characteristic graph are divided by the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic graph at the position point to obtain a sequence of millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood characteristic graphs; the positional addition of each millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood characteristic graph in the sequence of the millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood characteristic graph is divided by the millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood number statistics at the position point to obtain the millimeter-wave echo signal frequency domain gradient amplitude local significant distribution characteristic graph.

[0059] Specifically, the millimeter wave echo signal frequency domain distribution feature map is input into the feature distribution gradient-based feature saliency engine and processed using the following gradient saliency feature enhancement formula to obtain the millimeter wave echo signal frequency domain saliency distribution feature map;

[0060] Among them, the gradient significant feature enhancement formula is:

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068] in, , , , , and They are respectively the first , , , , and The eigenvalues ​​of the positions, is the first value along the height direction in the frequency domain distribution characteristic diagram of the millimeter wave echo signal The gradient amplitude value at each position, is the first value along the width direction in the frequency domain distribution characteristic diagram of the millimeter wave echo signal The gradient amplitude value at each position, is the first The gradient amplitude value at each position, is the first in the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal The eigenvalues ​​at the positions, is the first The eigenvalues ​​at the positions, The range is Neighborhood of express The number of elements in , The first part of the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter wave echo signal is shown in The eigenvalues ​​at the positions, represents the Gaussian error linear gating function, The local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal is represented by The eigenvalues ​​at the positions, is the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal, is the frequency domain distribution characteristic diagram of the millimeter wave echo signal, It means point multiplication by position. It is a frequency domain significant distribution feature diagram of the millimeter wave echo signal.

[0069] Then, cluster analysis is performed on the frequency domain significant distribution feature map of the millimeter wave echo signal to obtain the cluster analysis result. That is to say, the frequency domain distribution characteristics of the millimeter wave radar detection echo signal after the characteristic distribution gradient is significantly enhanced are used for cluster analysis. It is worth mentioning that in the technical solution of the present application, the goal of cluster analysis is to find the cluster center representing the personnel target from the frequency domain significant distribution feature map of the millimeter wave echo signal. That is to say, different people have different positions in space, and their performance in the frequency domain distribution feature map of the echo signal is also different from other environmental parts or objects. Therefore, the personnel target is used as the cluster center in the signal frequency domain distribution feature to distinguish from other parts, which helps to achieve accurate statistics of the number of people. Furthermore, the number of cluster centers in the cluster analysis results is used as the total number of personnel statistics, which can achieve more intelligent personnel perception quantity statistics, and the use of artificial intelligence and deep learning data processing algorithms can reduce the dependence on specific parameter settings and improve the robustness and reliability of the system.

[0070] In one embodiment of the present application, cluster analysis is performed on the millimeter wave echo signal frequency domain significant distribution feature map to determine the total number of personnel statistics, including: cluster analysis is performed on the millimeter wave echo signal frequency domain significant distribution feature map to obtain a cluster analysis result; and the number of cluster centers in the cluster analysis result is used as the total number of personnel statistics.

[0071] In the technical solution of the present application, the frequency domain distribution feature map of the millimeter wave echo signal is used to represent the local semantic association features of the millimeter wave frequency distribution of the frequency domain distribution matrix of the millimeter wave echo sampling points determined by ADC sampling and fast Fourier transform of the millimeter wave detection echo signal. In this way, when the frequency domain distribution feature map of the millimeter wave echo signal is input into the feature distribution gradient mask salient device, the feature distribution gradient mask salient device uses the gradient amplitude value of each position of the frequency domain distribution feature map of the millimeter wave echo signal to construct a class attention space dimension gated mask salient mechanism to highlight the feature distribution significance of the frequency domain distribution feature map of the millimeter wave echo signal in the spatial dimension. However, this will also cause the image semantic clustering alignment conflict in the spatial dimension of the frequency domain significant distribution feature map of the millimeter wave echo signal, resulting in the loss of aggregation key information, affecting the expression effect of the frequency domain significant distribution feature map of the millimeter wave echo signal and further affecting the accuracy of the clustering analysis results of the frequency domain significant distribution feature map of the millimeter wave echo signal.

[0072] In a preferred example, after the millimeter wave echo signal frequency domain distribution feature map is input into a feature distribution gradient mask salient device to obtain a millimeter wave echo signal frequency domain significant distribution feature map, the millimeter wave echo signal frequency domain significant distribution feature map is subjected to feature distribution space optimization, comprising the following steps:

[0073] The absolute value sum of all eigenvalues ​​of the millimeter wave echo signal frequency domain significant distribution feature map and the square root of the square sum are calculated to obtain the first millimeter wave echo signal frequency domain significant distribution spatial structure value and the second millimeter wave echo signal frequency domain significant distribution spatial structure value, that is:

[0074]

[0075]

[0076] in, The first one represents the significant distribution characteristic diagram of the millimeter wave echo signal in the frequency domain eigenvalues, Indicates the spatial structure value of the significant distribution of the first millimeter wave echo signal in the frequency domain, Indicates the significant distribution spatial structure value of the second millimeter wave echo signal in the frequency domain;

[0077] Determine the total number of eigenvalues ​​of all eigenvalues ​​of the frequency domain significant distribution feature map of the millimeter wave echo signal ;

[0078] For each eigenvalue of the millimeter wave echo signal frequency domain significant distribution feature map, calculate the first millimeter wave echo signal frequency domain significant distribution spatial structure value minus the product of the eigenvalue and the total number of eigenvalues ​​to obtain the first millimeter wave echo signal frequency domain significant distribution long-range dependence value ,in, Indicates the spatial structure value of the significant distribution of the first millimeter wave echo signal in the frequency domain, The first one represents the significant distribution characteristic diagram of the millimeter wave echo signal in the frequency domain eigenvalues, The total number of eigenvalues ​​representing all eigenvalues ​​of the significant distribution feature graph of the millimeter wave echo signal in the frequency domain, Indicates the long-range dependence value of the significant distribution of the first millimeter wave echo signal in the frequency domain;

[0079] Calculate the product of the square root of the total number of eigenvalues ​​multiplied by the eigenvalue minus the spatial structure value of the significant distribution of the frequency domain of the second millimeter wave echo signal to obtain the long-range dependence value of the significant distribution of the frequency domain of the second millimeter wave echo signal ,in, Indicates the significant distribution spatial structure value of the second millimeter wave echo signal in the frequency domain, The first one represents the significant distribution characteristic diagram of the millimeter wave echo signal in the frequency domain eigenvalues, The total number of eigenvalues ​​representing all eigenvalues ​​of the significant distribution feature graph of the millimeter wave echo signal in the frequency domain, Indicates the long-range dependence value of the significant distribution of the second millimeter wave echo signal in the frequency domain;

[0080] The index value calculated by taking the first millimeter wave echo signal frequency domain significant distribution long-range dependence value as the exponent of the natural constant and the inverse of the second millimeter wave echo signal frequency domain significant distribution long-range dependence value are weighted summed to obtain the optimized eigenvalue corresponding to each eigenvalue ,in, Indicates the long-range dependence value of the first millimeter wave echo signal in the frequency domain. Indicates the long-range dependence value of the second millimeter wave echo signal in the frequency domain. represents a natural constant, and represents the weighted hyperparameter, Indicates the optimized eigenvalue corresponding to each eigenvalue;

[0081] The optimized characteristic values ​​are combined into an optimized millimeter wave echo signal frequency domain significant distribution characteristic diagram.

[0082] That is, in view of the possible lack of spatial structure of the feature set of the millimeter-wave echo signal frequency domain significant distribution feature map in the high-dimensional space, which leads to the weight of the classifier implicitly inferring the spatial structure information based on the features, resulting in inconsistent convergence. A long-distance feature dependency relationship is established based on the spatial structure representation of the millimeter-wave echo signal frequency domain significant distribution feature map based on the overall feature scale of the millimeter-wave echo signal frequency domain significant distribution feature map, so as to establish the feature local connectivity of the millimeter-wave echo signal frequency domain significant distribution feature map, and to capture the spatial ambiguous information of the object feature value through the unstructured feature value point prediction of the millimeter-wave echo signal frequency domain significant distribution feature map, thereby improving the spatial inductive deviation perception ability of the feature set of the millimeter-wave echo signal frequency domain significant distribution feature map, thereby improving the expression effect of the optimized millimeter-wave echo signal frequency domain significant distribution feature map, thereby improving the accuracy of the clustering analysis results of the millimeter-wave echo signal frequency domain significant distribution feature map, and achieving more accurate personnel perception quantity statistics, thereby improving the robustness and reliability of the system.

[0083] In summary, the above scheme is adopted, the detection echo signal is received by the millimeter wave radar, and the detection echo signal is sampled by ADC to convert it into a digital signal, and then the fast Fourier transform is used to process the digital signal of the detection echo signal from the time domain to the frequency domain to obtain the frequency domain distribution matrix of the millimeter wave echo sampling point, so as to analyze the frequency components of the signal. Then, the data processing and analysis algorithm based on artificial intelligence and deep learning is used at the back end to analyze the frequency domain distribution matrix of the millimeter wave echo sampling point, so as to learn and capture the key signal frequency domain distribution characteristics in the frequency domain distribution matrix of the millimeter wave echo sampling point, and cluster analysis is performed based on the key signal frequency domain distribution characteristics, so as to use the number of cluster centers in the cluster analysis results as the total number of personnel statistics. In this way, a more intelligent number of personnel perception statistics can be achieved, and the data processing algorithm based on artificial intelligence and deep learning can reduce the dependence on specific parameter settings and improve the robustness and reliability of the system.

[0084] Figure 2 FIG. 1 is a block diagram of a system for counting the number of people based on millimeter wave radar according to an exemplary embodiment. Figure 2 As shown, the system 200 includes:

[0085] A radar detection echo signal acquisition module 201 is used to acquire a millimeter wave radar detection echo signal;

[0086] An ADC sampling module 202 is used to perform ADC sampling on the millimeter wave radar detection echo signal to obtain a millimeter wave echo sampling digital signal;

[0087] A fast Fourier transform module 203 is used to perform a fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a frequency domain distribution matrix of millimeter wave echo sampling points;

[0088] The echo signal frequency domain feature extraction module 204 is used to extract the echo signal frequency domain features from the millimeter wave echo sampling point frequency domain distribution matrix to obtain a millimeter wave echo signal frequency domain distribution feature map;

[0089] The feature distribution gradient significant module 205 is used to input the millimeter wave echo signal frequency domain distribution feature map into a feature significant engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain significant distribution feature map;

[0090] The cluster analysis module 206 is used to perform cluster analysis on the significant distribution feature diagram of the millimeter wave echo signal frequency domain to determine the total number of personnel statistics.

[0091] In one embodiment of the present application, the echo signal frequency domain feature extraction module is used to: input the millimeter wave echo sampling point frequency domain distribution matrix into the echo signal frequency domain feature extractor based on the void convolutional neural network model to obtain the millimeter wave echo signal frequency domain distribution feature map.

[0092] In one embodiment of the present application, the feature distribution gradient significance module includes: a multi-directional gradient value calculation unit, which is used to calculate the multi-directional gradient value distribution of each position in the millimeter-wave echo signal frequency domain distribution feature map, and determine the gradient amplitude value of each position in the millimeter-wave echo signal frequency domain distribution feature map based on the multi-directional gradient value distribution of each position to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution feature map; a gradient amplitude local description operator calculation unit, which is used to calculate the gradient amplitude local description operator of each position in the millimeter-wave echo signal frequency domain gradient amplitude distribution feature map. The operator is used to obtain a local significant distribution feature map of the frequency domain gradient amplitude of the millimeter wave echo signal; a gated mask unit is used to input the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter wave echo signal into a gated masker based on the GELU function to obtain a local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal; a position point multiplication calculation unit is used to calculate the position point multiplication between the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal and the frequency domain distribution feature map of the millimeter wave echo signal to obtain the frequency domain significant distribution feature map of the millimeter wave echo signal.

[0093] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0094] The modules involved in the embodiments described in this application can be implemented by software or hardware. The name of the module does not limit the module itself in some cases. For example, the test parameter acquisition module can also be described as a "module for acquiring device test parameters corresponding to the target device".

[0095] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0096] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

[0097] In addition, although each operation is described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or to be performed in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0098] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely exemplary forms of implementation. With respect to the apparatus in the above-mentioned embodiments, the specific manner in which each module performs the operation has been described in detail in the embodiments related to the method, and will not be described in detail here.

Claims

1. A method for counting the number of people sensing based on millimeter wave radar, characterized in that: include: Obtain millimeter wave radar detection echo signal; Performing ADC sampling on the millimeter-wave radar detection echo signal to obtain a millimeter-wave echo sampling digital signal; Performing a fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a frequency domain distribution matrix of millimeter wave echo sampling points; Extracting frequency domain features of the echo signal from the frequency domain distribution matrix of the millimeter wave echo sampling points to obtain a frequency domain distribution feature graph of the millimeter wave echo signal; Inputting the millimeter wave echo signal frequency domain distribution feature map into a feature saliency engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain saliency distribution feature map; A cluster analysis is performed on the frequency domain significant distribution feature diagram of the millimeter wave echo signal to determine the total number of personnel statistics.

2. The method for counting the number of people sensed based on millimeter wave radar according to claim 1, characterized in that: The frequency domain features of the millimeter wave echo sampling points are extracted from the frequency domain features to obtain a frequency domain distribution feature map of the millimeter wave echo signal, including: inputting the frequency domain distribution matrix of the millimeter wave echo sampling points into an echo signal frequency domain feature extractor based on a hole convolutional neural network model to obtain the frequency domain distribution feature map of the millimeter wave echo signal.

3. The method for counting the number of people sensing based on millimeter wave radar according to claim 2 is characterized in that: Inputting the millimeter wave echo signal frequency domain distribution feature map into a feature saliency engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain saliency distribution feature map, including: Calculating the multi-directional gradient value distribution of each position in the millimeter wave echo signal frequency domain distribution characteristic map, and determining the gradient amplitude value of each position in the millimeter wave echo signal frequency domain distribution characteristic map based on the multi-directional gradient value distribution of each position to obtain the millimeter wave echo signal frequency domain gradient amplitude distribution characteristic map; Calculating the local description operator of the gradient amplitude at each position in the frequency domain gradient amplitude distribution characteristic map of the millimeter wave echo signal to obtain a local significant distribution characteristic map of the frequency domain gradient amplitude of the millimeter wave echo signal; Inputting the millimeter wave echo signal frequency domain gradient amplitude local significant distribution feature map into a gated masker based on a GELU function to obtain a millimeter wave echo signal frequency domain gradient amplitude local significant gated mask feature map; The millimeter-wave echo signal frequency domain significant distribution feature map is obtained by calculating the point-by-point multiplication of the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter-wave echo signal and the frequency domain distribution feature map of the millimeter-wave echo signal.

4. The method for counting the number of people sensing based on millimeter wave radar according to claim 3 is characterized in that: Calculating the multidirectional gradient value distribution of each position in the millimeter wave echo signal frequency domain distribution feature map, and determining the gradient amplitude value of each position in the millimeter wave echo signal frequency domain distribution feature map based on the multidirectional gradient value distribution of each position to obtain the millimeter wave echo signal frequency domain gradient amplitude distribution feature map, including: Calculating the gradient amplitude values ​​of the millimeter wave echo signal frequency domain distribution characteristic diagram along different directions to obtain a millimeter wave echo signal frequency domain height gradient distribution characteristic diagram, a millimeter wave echo signal frequency domain width gradient distribution characteristic diagram, and a millimeter wave echo signal frequency domain channel gradient distribution characteristic diagram; The binary norm between the characteristic values ​​at corresponding positions in the millimeter-wave echo signal frequency domain height gradient distribution characteristic diagram, the millimeter-wave echo signal frequency domain width gradient distribution characteristic diagram and the millimeter-wave echo signal frequency domain channel gradient distribution characteristic diagram is calculated to obtain the millimeter-wave echo signal frequency domain gradient amplitude distribution characteristic diagram.

5. The method for counting the number of people sensing based on millimeter wave radar according to claim 4 is characterized in that: The millimeter wave echo signal frequency domain distribution characteristic diagram is subjected to gradient amplitude value calculation along different directions to obtain a millimeter wave echo signal frequency domain height gradient distribution characteristic diagram, a millimeter wave echo signal frequency domain width gradient distribution characteristic diagram, and a millimeter wave echo signal frequency domain channel gradient distribution characteristic diagram, including: Calculating the gradient amplitude value of each position characteristic value in the frequency domain distribution characteristic diagram of the millimeter wave echo signal along the height direction to obtain the frequency domain height gradient distribution characteristic diagram of the millimeter wave echo signal; Calculating the gradient amplitude value of each position characteristic value in the frequency domain distribution characteristic diagram of the millimeter wave echo signal along the width direction to obtain the frequency domain width gradient distribution characteristic diagram of the millimeter wave echo signal; The gradient amplitude value of each position characteristic value in the frequency domain distribution characteristic diagram of the millimeter wave echo signal is calculated along the channel direction to obtain the frequency domain channel gradient distribution characteristic diagram of the millimeter wave echo signal.

6. The method for counting the number of people sensing based on millimeter wave radar according to claim 5 is characterized in that: Calculating the local description operator of the gradient amplitude at each position in the frequency domain gradient amplitude distribution feature map of the millimeter wave echo signal to obtain the local significant distribution feature map of the frequency domain gradient amplitude of the millimeter wave echo signal, including: Determine a local neighborhood range of the gradient amplitude, and calculate a sequence of neighborhood characteristic maps of the frequency domain gradient amplitude distribution of the millimeter wave echo signal based on the local neighborhood range of the gradient amplitude and the frequency domain gradient amplitude distribution characteristic map of the millimeter wave echo signal; Counting the number of feature graphs in the sequence of the neighborhood feature graphs of the frequency domain gradient amplitude distribution of the millimeter wave echo signal to obtain a statistical value of the number of neighborhood feature graphs of the frequency domain gradient amplitude distribution of the millimeter wave echo signal; Calculating the position difference between each millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood feature map in the sequence of the millimeter-wave echo signal frequency domain gradient amplitude distribution feature map and the millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood feature map, and dividing the difference by position point with the millimeter-wave echo signal frequency domain gradient amplitude distribution feature map to obtain a sequence of millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature maps; Each millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature map in the sequence of the millimeter-wave echo signal frequency domain gradient relative amplitude distribution neighborhood feature map is added point by point and then divided by the millimeter-wave echo signal frequency domain gradient amplitude distribution neighborhood statistical value point by point to obtain the millimeter-wave echo signal frequency domain gradient amplitude local significant distribution feature map.

7. The method for counting the number of people sensed based on millimeter wave radar according to claim 6, characterized in that: Performing cluster analysis on the millimeter wave echo signal frequency domain significant distribution feature map to determine the total number of personnel statistics, including: Performing cluster analysis on the significant distribution feature graph of the millimeter wave echo signal in the frequency domain to obtain a cluster analysis result; The number of cluster centers in the cluster analysis result is used as the total number of personnel statistics.

8. A personnel sensing and counting system based on millimeter wave radar, characterized in that: include: A radar detection echo signal acquisition module is used to acquire millimeter wave radar detection echo signals; An ADC sampling module is used to perform ADC sampling on the millimeter wave radar detection echo signal to obtain a millimeter wave echo sampling digital signal; A fast Fourier transform module, used for performing a fast Fourier transform on the millimeter wave echo sampling digital signal to obtain a frequency domain distribution matrix of the millimeter wave echo sampling points; An echo signal frequency domain feature extraction module is used to extract the echo signal frequency domain features from the millimeter wave echo sampling point frequency domain distribution matrix to obtain a millimeter wave echo signal frequency domain distribution feature map; A feature distribution gradient significant module, used for inputting the millimeter wave echo signal frequency domain distribution feature map into a feature significant engine based on feature distribution gradient to obtain a millimeter wave echo signal frequency domain significant distribution feature map; The cluster analysis module is used to perform cluster analysis on the significant distribution feature diagram of the millimeter wave echo signal frequency domain to determine the total number of personnel statistics.

9. The personnel sensing and counting system based on millimeter wave radar according to claim 8, characterized in that: The echo signal frequency domain feature extraction module is used to: input the millimeter wave echo sampling point frequency domain distribution matrix into the echo signal frequency domain feature extractor based on the void convolutional neural network model to obtain the millimeter wave echo signal frequency domain distribution feature map.

10. The personnel sensing and counting system based on millimeter wave radar according to claim 9, characterized in that: The feature distribution gradient significant module includes: a multi-directional gradient value calculation unit, used to calculate the multi-directional gradient value distribution of each position in the millimeter wave echo signal frequency domain distribution characteristic map, and determine the gradient amplitude value of each position in the millimeter wave echo signal frequency domain distribution characteristic map based on the multi-directional gradient value distribution of each position to obtain the millimeter wave echo signal frequency domain gradient amplitude distribution characteristic map; A gradient amplitude local description operator calculation unit, used to calculate the gradient amplitude local description operator of each position in the frequency domain gradient amplitude distribution characteristic map of the millimeter wave echo signal to obtain a frequency domain gradient amplitude local significant distribution characteristic map of the millimeter wave echo signal; A gated mask unit, used for inputting the millimeter wave echo signal frequency domain gradient amplitude local significant distribution feature map into a gated masker based on a GELU function to obtain a millimeter wave echo signal frequency domain gradient amplitude local significant gated mask feature map; The position point multiplication calculation unit is used to calculate the position point multiplication between the local significant gated mask feature map of the frequency domain gradient amplitude of the millimeter wave echo signal and the frequency domain distribution feature map of the millimeter wave echo signal to obtain the frequency domain significant distribution feature map of the millimeter wave echo signal.

Images