A method for judging new targets in millimeter wave radar personnel tracking

Abstract

The application discloses a kind of millimeter wave radar personnel tracking in new target generation judging method, comprising the following steps: to be measured space range is launched electromagnetic wave signal by millimeter wave radar, and the signal reflected back in the space range to be measured is received by millimeter wave radar;The received signal is processed, and the point cloud information of current frame is obtained;The point cloud left after being associated with existing target is analyzed, to judge whether it meets the condition of regenerating a new target;If the point cloud of last frame meets the condition of generating new target, an observation target is generated, and then the state of the observation target and the associated point cloud are recorded frame by frame within the set time;After reaching the set time, whether it becomes a new target is judged in combination with the continuity of the observation target within this period of time and the reliability of associated point cloud.The application considers the information of point cloud and target in time dimension to judge whether new target is generated, which significantly reduces the possibility of false target generation.

Description

TECHNICAL FIELD

[0001] The application relates to a millimeter wave radar signal processing technology, and particularly relates to a method for judging new target generation in millimeter wave radar personnel tracking. BACKGROUND

[0002] Personnel detection and quantity statistics are indispensable data in many existing scenes, such as airports, shopping malls, chain stores, stations, museums, conference rooms and the like. Personnel discovery, counting and data estimation play an important role in safety management and improvement of overall service level. Common personnel tracking detection technologies mainly use single identification methods of images or infrareds. However, images have certain influence on privacy and have limited application range. Infrareds have high resolution and high cost, and are easily disturbed.

[0003] In existing personnel detection and quantity statistics applications, mainly include camera recognition, traditional manual counting, automatic counting mechanical equipment, RFID access control systems and the like. These methods are limited by cost, accuracy, implementation difficulty and the like, so that personnel detection and quantity statistics application is difficult to achieve balance in economy and accuracy.

[0004] In existing personnel detection and quantity statistics methods, the new target generation method is mainly based on point clouds generated by signal processing on received signals of a current frame, and whether a new target track is generated in the region is judged according to the number, Doppler velocity and signal-to-noise ratio of the point clouds. However, external factors are easily disturbed to produce false targets. SUMMARY

[0005] In order to solve the above technical problems, the application provides a method for judging new target generation in millimeter wave radar personnel tracking.

[0006] The technical scheme adopted by the application is as follows: a method for judging new target generation in millimeter wave radar personnel tracking is provided, comprising the following steps:

[0007] a. transmitting electromagnetic wave signals to a to-be-measured space range through a millimeter wave radar, and receiving signals reflected back in the to-be-measured space range through the millimeter wave radar;

[0008] b. processing the received signals to obtain point cloud information of a current frame;

[0009] c. analyzing the point clouds left after association with existing targets to judge whether the point clouds meet the condition of generating a new target;

[0010] d. If the point cloud in the previous frame meets the conditions for generating a new target, instead of generating a new target directly, an observation target is generated, and then the state of the observation target and the associated point cloud are recorded frame by frame within a set time.

[0011] e. After the set time is reached, determine whether the observed target has become a new target based on the continuity of the observed target during this period and the reliability of the associated point cloud.

[0012] As a preferred method for determining the generation of new targets in millimeter-wave radar personnel tracking, step b includes:

[0013] The acquired signals are subjected to a fast Fourier transform, transforming them from the time domain to the frequency domain. The acquired radar signals are then analyzed based on their frequency domain characteristics to obtain a point cloud containing information on range, azimuth, elevation, Doppler velocity, and signal-to-noise ratio.

[0014] As a preferred method for determining the generation of new targets in millimeter-wave radar personnel tracking, step c includes:

[0015] The generated point clouds are clustered into point clusters. Each point cluster contains multiple point clouds that are close to each other. After associating the point clusters with nearby existing targets, the remaining point clusters that have not been associated are analyzed.

[0016] As a preferred method for determining the generation of new targets in millimeter-wave radar personnel tracking, step d includes:

[0017] When the P of the nth point heap n ≥P0, S n ≥S0、d n ≥d0、N n When ≤N0, the pile at that point is considered the observation target, and the above P n Let d be the number of points in the nth point heap, Sn be the average snr of all points in the heap, and d be the number of points in the nth point heap. n To minimize the distance to nearby targets, N n The number of surrounding targets; P0, S0, d0, and N0 are set values;

[0018] Within a set time period, at frame i, the cluster of points associated with the observed target K satisfies P. iK ≥P0, S iK ≥S0, and the average Doppler velocity of the point pile |dop in If |≥dop0, where dop0 is a set value, then the observed target K is considered to meet the condition in the current frame, and the total number of frames meeting the condition is C. K =C K +1; If the current observed target K is occluded by other targets, then the number of occluded frames is O. K =O K+1.

[0019] As a preferred way of the judgment method for new target generation in the personnel tracking of the millimeter wave radar, the set value P0 is 6-15, S0 is 20-50, d0 is 0.8-1.5 meters, N0 is 1 or 2, and dop0 is 0.05-0.1 m / s.

[0020] As a preferred way of the judgment method for new target generation in the personnel tracking of the millimeter wave radar, the step e comprises:

[0021] After a certain number of frames of the observed target are generated, it is judged whether the observed target is a false target or a new target; suppose that the score S of the observed target K in the i-th frame is i =a*P iK +b*S iK -c*D iK , wherein D iK represents the distance measured in the i-th frame of the observed target K, wherein a, b, and c are weight values; after m frames of the observed target K are generated, the score S of the observed target K is , wherein d and e are weight values; finally, whether the score S of the target K is greater than a set score is judged to determine whether the observed target is converted into a new target track. K

[0022] The beneficial effects of the present application are: in view of the fact that many false targets are easily generated in the application of the millimeter wave radar in personnel tracking and positioning, a new target judgment method is proposed, which comprehensively considers the information of the point cloud and the target in the time dimension to judge whether a new target is generated, and significantly reduces the possibility of false target generation. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 The flow chart of the judgment method for new target generation in the personnel tracking of the millimeter wave radar disclosed by the present application.

[0024] Figure 2 is a point cloud schematic diagram obtained after signal processing of the millimeter wave radar signal. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings, but the embodiments of the present application are not limited thereto.

[0026] ​The existing new target generation method is mostly based on the point cloud generated by signal processing on the received signal of the current frame, and whether a new target track is generated in the region is judged according to the number of point clouds, Doppler velocity and signal-to-noise ratio, but false targets are easily generated due to external factor interference, and the application proposes a method for generating a new target in millimeter wave radar personnel tracking, which combines the divergence of the point cloud and the point cloud information obtained in different frames and the correlation degree, analyzes the continuity and correlation of the point cloud in time, and obviously reduces the probability of false target generation, effectively improves the tracking performance of the millimeter wave radar.

[0027] Reference Figure 1 The specific scheme of the application is: a judgment method for generating a new target in millimeter wave radar personnel tracking, comprising the following steps:

[0028] a. Transmitting an electromagnetic wave signal to the space range to be measured by a millimeter wave radar, and receiving the signal reflected back in the space range to be measured by the millimeter wave radar.

[0029] Specifically, the electromagnetic wave signal is transmitted to the space to be measured, the millimeter wave radar uses a 60GHz millimeter wave broadband radar, the radar transmission and reception parameters are adjusted, and the space to be measured can be adjusted from 10m*10m to 100m*100m, which can be applied to small range application scenarios such as indoor conference rooms, and can also be used for outdoor application scenarios such as scenic spots.

[0030] b. Process the received signal to obtain the point cloud information of the current frame.

[0031] Specifically, the collected signal is subjected to fast Fourier transform (FFT), and the collected signal is transformed from time domain to frequency domain, and the collected radar signal is analyzed from frequency domain characteristics to obtain point clouds containing distance, azimuth angle, pitch angle, Doppler speed and signal-to-noise ratio information. The point cloud distribution is shown by the small dots in Figure 2 , and Figure 2 The circles represent the target position.

[0032] c. Analyze the remaining point clouds associated with the existing target to determine whether they meet the condition for generating a new target.

[0033] Specifically, the generated point clouds are clustered into point piles, each point pile contains multiple point clouds with close positions, and after associating the point piles with the existing targets nearby, the remaining point piles that are not associated are analyzed.

[0034] d. If the point cloud of the previous frame meets the condition for generating a new target, a new target is not directly generated, but an observation target is generated, and then the state of the observation target and the associated point cloud are recorded frame by frame within a set time.

[0035] Specifically, when the Pn ≥P0, S n ≥S0、d n ≥d0、N n When ≥N0, the point pile is identified as the observation target, and the above P n Let d be the number of points in the nth point heap, Sn be the average snr of all points in the heap, and d be the number of points in the nth point heap. n To minimize the distance to nearby targets, N n Let P0 be the number of surrounding targets; S0, d0, and N0 are set values. Within a set time period, at frame i, the cluster of points associated with the observed target K satisfies P... iK ≥P0, S iK ≥S0, and the average Doppler velocity of the point pile |dop in If |≥dop0, where dop0 is a set value, then the observed target K is considered to meet the condition in the current frame, and the total number of frames meeting the condition is C. K =C K +1; If the current observed target K is occluded by other targets, then the number of occluded frames is O. K =O K +1. In actual operation, the number of points P0 in the set point cluster can be 6-15, the average snr value S0 of the set point cluster can be 20-50, the minimum distance d0 to nearby targets can be 0.8-1.5 meters, and the number of surrounding targets N0 can be 1 or 2. The average Doppler velocity dop0 of the set point cluster can be 0.05-0.1 m / s.

[0036] e. After the set time is reached, determine whether the observed target has become a new target based on the continuity of the observed target during this period and the reliability of the associated point cloud.

[0037] Specifically, after a certain number of frames have passed since the observed target was generated, it is determined whether the observed target is a false target or a new target; assuming the score S of the observed target K in the i-th frame is... i =a*P iK +b*S iK -c*D iK D iK This represents the distance measured by the observed target K in the i-th frame, where a, b, and c are weight values; the score of the observed target K is calculated after m frames are generated. Where d and e are weight values; finally, based on the target K score S K The system determines whether the observed target has changed into a new target track based on whether the score is greater than a set value, and updates the number of people in the current frame accordingly.

[0038] The above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalent features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining generation of a new target in millimeter wave radar personnel tracking, characterized by, The method comprises the following steps: S1. Transmitting electromagnetic wave signals to a space range to be measured by a millimeter wave radar, and receiving signals reflected back in the space range to be measured by the millimeter wave radar; S2. Processing the received signals to obtain point cloud information of a current frame; S3. Analyzing the point cloud remaining after being associated with an existing target to determine whether the point cloud meets a condition for regenerating an observation target; S4. If the point cloud of the previous frame meets the condition for generating an observation target, generating an observation target, and then recording the state of the observation target and the associated point cloud frame by frame within a set time; The step S4 comprises: When the nth point cluster is , , , , the point cluster is determined as the observation target, and the above is the number of points of the nth point cluster, Sn is the average value of snr of all points in the point cluster, is the minimum distance from the nearby target, is the number of surrounding targets; is a set value; At a set time, at the i-th frame, the point pile associated with the observation target K satisfies , , and the average Doppler velocity of the point pile , is a set value, it is considered that the observation target K satisfies the condition at the current frame, the total number of frames satisfying the condition ; if the observation target K is currently blocked by other targets, the number of blocked frames ; S5. After the set time is reached, determining whether the observation target becomes a new target in combination with the continuity of the observation target within the time and the reliability of the associated point cloud; The step S5 comprises: After a certain number of frames have passed since the observed target was generated, it is determined whether the observed target is a false target or a new target; the score of observed target K in the i-th frame. , where P iK S is the number of points in the point stack of the observation target K in the i-th frame. iK Let SNR be the average value of all points in the point heap, where This represents the distance measured by the observed target K in the i-th frame, where a, b, and c are weight values; the score of the observed target K is calculated after m frames are generated. , where C K To satisfy the conditions, O K The number of frames occluded is given, where d and e are weight values; finally, the target K score is calculated. Whether the observed target has changed into a new target track is determined by whether it exceeds a set score.

2. The method of claim 1, wherein the determination of the generation of the new target in the millimeter wave radar personnel tracking is characterized by, The step S2 comprises: Performing fast Fourier transform on the collected signals, transforming the collected signals from the time domain to the frequency domain, and analyzing the collected radar signals from the frequency domain characteristics to obtain point cloud containing distance, azimuth angle, elevation angle, Doppler velocity and signal-to-noise ratio information.

3. The method of claim 2, wherein the determination of the generation of the new target in the millimeter wave radar personnel tracking is characterized by, The step S3 comprises: Clustering the generated point cloud into point piles, each point pile containing multiple point clouds with close positions, associating the point piles with nearby existing targets, and then analyzing the point piles that are not associated.

4. The method of claim 1, wherein the determination of the new target generation in the millimeter wave radar personnel tracking is characterized by, Setpoint Take 6-15, Take 20-50, Take 0.8-1.5 meters, Take 1 or 2, Take 0.05-0.1 m / s.

Images