Interference filtering method and device for millimeter wave radar detection, storage medium and electronic equipment

Abstract

The invention discloses an interference filtering method and device for millimeter wave radar detection, a storage medium and electronic equipment. The method comprises the following steps: acquiring point cloud data of a detection target through a millimeter wave radar; determining a current category feature and a current target distance of the point cloud based on the point cloud data; based on the current category features and the current target distance, comparing a predetermined relationship between category features of normal targets under each set category of the current scene and the target distance, and determining scores of the detection target belonging to each set category; judging whether the detection target belongs to each set category based on the score, and if the detection target does not belong to any set category, determining the detection target as an interference target; and filtering the point cloud data of the interference target. According to the invention, the interference target in millimeter-wave radar detection can be effectively identified, and the accuracy of millimeter-wave radar detection is improved.

Description

Technical Field

[0001] The present application relates to millimeter wave radar technology, and in particular to an interference filtering method, device, storage medium and electronic device for millimeter wave radar detection. Background Art

[0002] Capturing and perceiving traffic targets such as motor vehicles and non-motor vehicles is the most core processing task in intelligent road traffic business scenarios. Current road monitoring systems usually use a combination of millimeter-wave radar and camera to form a radar-vision tracking system, which combines the rich semantics of vision with the advantages of radar in perceiving moving targets. Through multimodal data fusion, target tracking and detection evidence collection can achieve better road traffic monitoring and tracking effects.

[0003] In business practice, the above-mentioned millimeter-wave radar and camera combined processing method has significant effects, but it has also encountered many problems. One of the more typical problems is the problem of target misdetection caused by radar multipath. The multipath misdetection of millimeter-wave radar is the misdetected point cloud and misdetected target formed by the point cloud after the millimeter wave is reflected once or multiple times by obstacles such as walls and the ground. This will cause great interference to target perception and affect the accuracy of millimeter-wave radar detection. For example, in a tunnel scene, Figure 1a This is an illustration of the position of the radar target and point cloud projected onto the ground. Figure 1b The target track and point cloud position in the radar coordinate system are shown in Figure 2. Figure 1a Two parallel trucks in the middle and near distance, due to multipath Figure 1b There are multiple multipath false detection targets marked by circles in the middle. Figure 1a The corresponding positions in the figure are marked by arrows, and it can be seen that the multipath misdetected targets form interference in the distance. Summary of the invention

[0004] The present application provides an interference filtering method, device, storage medium and electronic device for millimeter-wave radar detection, which can effectively identify interference targets in millimeter-wave radar detection and improve the accuracy of millimeter-wave radar detection.

[0005] To achieve the above objectives, this application adopts the following technical solutions: A method for filtering interference in millimeter wave radar detection, comprising: Acquire point cloud data of the detection target through millimeter wave radar; Determining a current category feature and a current target distance of the point cloud based on the point cloud data; Based on the current category feature and the current target distance, comparing the relationship between the category feature and the target distance of the normal target in each set category of the current scene, the score of the detected target belonging to each set category is determined; Determining whether the detection target belongs to each of the set categories based on the score, and if the detection target does not belong to any of the set categories, determining that the detection target is an interference target; The point cloud data of the interfering target is filtered out.

[0006] Preferably, the category features include: point number features of the target point cloud, signal-to-noise ratio (SNR) features of the target point cloud, velocity consistency features of the target point cloud and / or radar cross section (RCS) features.

[0007] Preferably, the method of determining the relationship between the category feature of the normal target under any set category of the current scene and the target distance includes: Statistical analysis is performed on the category characteristic values ​​of the target point cloud of multiple normal targets under any set category at multiple target distances, and a relationship curve between the category characteristic value of the normal target point cloud and the target distance under any set category is obtained by fitting.

[0008] Preferably, when there are multiple category features, determining the scores of the detection targets belonging to the respective set categories includes: For any category feature, based on the relationship between the any category feature of the normal target under each set category and the target distance, determine the score corresponding to the any category feature of each set category that the detected target belongs to; The scores corresponding to the features of each category are comprehensively considered to determine the scores of the detection target belonging to each set category.

[0009] Preferably, the category feature is a point number feature of the target point cloud; The category feature values ​​include: the mean value of the number of points and the upper and lower boundary values ​​of the number of points of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the mean value of the number of points of the normal target point cloud and the target distance and a relationship curve between the upper and lower boundary values ​​of the number of points of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on a relationship curve between the mean number of points of a normal target point cloud and the target distance under any of the set categories, determining the mean number of points corresponding to the current target distance; Based on the relationship curve between the upper and lower boundary values ​​of the number of points and the target distance of the normal target point cloud under any set category, determine the upper and lower boundary values ​​of the number of points corresponding to the current target distance, and calculate the fluctuation amplitude of the number of points corresponding to the current target distance based on the upper and lower boundary values ​​of the number of points; Compare the current number of points of the detection target point cloud with the average number of points corresponding to the current target distance, and calculate the score corresponding to the point feature of the detection target belonging to any of the set categories based on the comparison result and the fluctuation range of the number of points; The score of the detection target belonging to any of the set categories is determined based on the score corresponding to the point feature.

[0010] Preferably, the category feature is an SNR feature of the target point cloud; The category feature values ​​include: the SNR mean and the SNR upper and lower boundary values ​​of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the SNR mean value of the normal target point cloud and the target distance and a relationship curve between the SNR upper and lower boundary values ​​of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on the relationship curve between the SNR mean value and the target distance of the normal target point cloud under any set category, determine the SNR mean value corresponding to the current target distance; Based on the relationship curve between the SNR upper and lower boundary values ​​and the target distance of the normal target point cloud under any set category, determine the SNR upper boundary value and the lower boundary value corresponding to the current target distance, and calculate the SNR fluctuation amplitude corresponding to the current target distance based on the SNR upper boundary value and the lower boundary value; Compare the current SNR mean of the detection target point cloud with the SNR mean corresponding to the current target distance, and calculate the score corresponding to the SNR feature of the detection target belonging to any of the set categories based on the comparison result and the SNR fluctuation amplitude; The score of the detected target belonging to any of the set categories is determined based on the score corresponding to the SNR feature.

[0011] Preferably, for any one of the set categories, a method of judging whether the detection target belongs to any one of the set categories based on the score that the detection target belongs to the any one of the set categories includes: If the score of the detection target belonging to any of the set categories is less than the score threshold corresponding to any of the set categories, it is determined that the detection target does not belong to any of the set categories; otherwise, it is determined that the detection target belongs to any of the set categories; or, If the score that the detection target belongs to any of the set categories is less than the score threshold corresponding to any of the set categories, the confidence that the detection target belongs to any of the set categories is lowered. After the confidence is lower than the confidence threshold of any of the set categories, it is determined that the detection target does not belong to any of the set categories.

[0012] Preferably, when pre-stating the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the ratio of any radar target trajectory to the same normal target trajectory within the set time reaches a set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; or, When pre-stating the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the corresponding relationship between any radar target trajectory and the normal target trajectory pair meets the preset condition, the binding confidence maintained for the any radar target trajectory and the normal target trajectory pair is increased, otherwise, the binding confidence is reduced; when the binding confidence is greater than or equal to the binding confidence threshold, the any radar target trajectory is bound to the normal target trajectory pair, and the bound radar target trajectory and the normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; wherein the preset condition is: the any radar target trajectory and the normal target trajectory pair are marked as corresponding in the corresponding relationship determined based on the RV calibration result, or the ratio of the number of times the any radar target trajectory and the normal target trajectory pair are marked as corresponding within a set time is greater than the binding threshold.

[0013] An interference filtering device for millimeter wave radar detection includes: a data acquisition unit, a scoring unit and an interference identification unit; The data acquisition unit is used to acquire point cloud data of the detection target through the millimeter wave radar, and determine the current category characteristics of the point cloud and the current target distance based on the point cloud data; The scoring unit is used to determine the scores of the detected target belonging to each set category based on the current category feature and the current target distance, and in comparison with the predetermined relationship between the category feature and the target distance of the normal target under each set category of the current scene; The interference identification unit is used to determine whether the detection target belongs to the set categories based on the score, and if the detection target does not belong to any set category, determine that the detection target is an interference target; and is also used to filter out the point cloud data of the interference target.

[0014] Preferably, the device further comprises a statistical unit for predetermining the relationship between the category characteristics of normal targets and the target distance under each set category of the current scene; The method for determining the relationship between the category feature of a normal target and the target distance under any set category feature of the current scene includes: Statistical analysis is performed on the category characteristic values ​​of the target point cloud of multiple normal targets under any set category at multiple target distances, and a relationship curve between the category characteristic value of the normal target point cloud and the target distance under any set category is obtained by fitting.

[0015] Preferably, in the scoring unit, when there are multiple category features, determining the scores of the detection target belonging to each set category includes: For any category feature, based on the relationship between the any category feature of the normal target under each set category and the target distance, determine the score corresponding to the any category feature of each set category that the detected target belongs to; The scores corresponding to the features of each category are comprehensively considered to determine the scores of the detection target belonging to each set category.

[0016] Preferably, the category feature is a point number feature of the target point cloud; The category feature values ​​include: the mean value of the number of points and the upper and lower boundary values ​​of the number of points of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the mean value of the number of points of the normal target point cloud and the target distance and a relationship curve between the upper and lower boundary values ​​of the number of points of the normal target point cloud and the target distance; In the scoring unit, a method of determining whether the detection target belongs to any of the set categories includes: Based on a relationship curve between the mean number of points of a normal target point cloud and the target distance under any of the set categories, determining the mean number of points corresponding to the current target distance; Based on the relationship curve between the upper and lower boundary values ​​of the number of points and the target distance of the normal target point cloud under any set category, determine the upper and lower boundary values ​​of the number of points corresponding to the current target distance, and calculate the fluctuation amplitude of the number of points corresponding to the current target distance based on the upper and lower boundary values ​​of the number of points; Compare the current number of points of the detection target point cloud with the average number of points corresponding to the current target distance, and calculate the point score of the detection target belonging to any of the set categories based on the comparison result and the fluctuation range of the number of points; A score is determined based on the point score for the detection target belonging to any of the set categories.

[0017] Preferably, the category feature is an SNR feature of the target point cloud; The category feature values ​​include: the SNR mean and the SNR upper and lower boundary values ​​of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the SNR mean value of the normal target point cloud and the target distance and a relationship curve between the SNR upper and lower boundary values ​​of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on the relationship curve between the SNR mean value and the target distance of the normal target point cloud under any set category, determine the SNR mean value corresponding to the current target distance; Based on the relationship curve between the SNR upper and lower boundary values ​​and the target distance of the normal target point cloud under any set category, determine the SNR upper boundary value and the lower boundary value corresponding to the current target distance, and calculate the SNR fluctuation amplitude corresponding to the current target distance based on the SNR upper boundary value and the lower boundary value; Compare the current SNR mean of the detection target point cloud with the SNR mean corresponding to the current target distance, and calculate the SNR score of the detection target belonging to any of the set categories based on the comparison result and the SNR fluctuation amplitude; A score is determined based on the SNR score, indicating that the detected target belongs to any one of the set categories.

[0018] Preferably, in the interference identification unit, for any one of the set categories, a method of judging whether the detection target belongs to any one of the set categories based on the score that the detection target belongs to the any one of the set categories includes: If the score of the detection target belonging to any of the set categories is less than the score threshold corresponding to any of the set categories, it is determined that the detection target does not belong to any of the set categories; otherwise, it is determined that the detection target belongs to any of the set categories; or, If the score that the detection target belongs to any of the set categories is less than the score threshold corresponding to any of the set categories, the confidence that the detection target belongs to any of the set categories is lowered. After the confidence is lower than the confidence threshold of any of the set categories, it is determined that the detection target does not belong to any of the set categories.

[0019] Preferably, when the statistical unit pre-statistics the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the ratio of any radar target trajectory to the same normal target trajectory within the set time reaches a set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; or, When pre-stating the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target track determined based on the RV calibration result and the normal target track in the image acquisition is recorded. If the corresponding relationship between any radar target track and the normal target track pair meets the preset condition, the binding confidence maintained for the any radar target track and the normal target track pair is increased, otherwise, the binding confidence is reduced; when the binding confidence is greater than or equal to the binding confidence threshold, the any radar target track and the normal target track pair are bound, and the bound radar target track and the normal target track are used to perform statistics on the relationship between the target distance and the category characteristics; wherein the preset condition is: the any radar target track and the normal target track pair are marked as corresponding in the corresponding relationship determined based on the RV marking result, or the ratio of the number of times the any radar target track and the normal target track pair are marked as corresponding within a set time is greater than the binding threshold.

[0020] A computer-readable storage medium stores computer instructions, which, when executed by a processor, can implement any of the above-mentioned interference filtering methods for millimeter-wave radar detection.

[0021] An electronic device comprising at least a computer-readable storage medium and a processor; The processor is used to read executable instructions from the computer-readable storage medium and execute the instructions to implement any of the interference filtering methods for millimeter-wave radar detection described above.

[0022] As can be seen from the above technical solution, in this application, the point cloud data of the detection target is obtained by the millimeter wave radar, and the current category features and current target distance of the point cloud are determined based on the point cloud data; thereby, the distance and category feature information for the current detection target can be obtained; next, based on the current category features and current target distance of the detection target, the relationship between the category features and target distance of the normal target under each set category of the current scene is determined in advance, and the score of the detection target belonging to each set category is determined; in this way, for each set category, the current category features and current target distance of the detection target are compared with the corresponding relationship of the normal target to obtain the score of the detection target belonging to the corresponding category, so as to characterize the possibility that the detection target belongs to the corresponding category. Finally, based on the score, it is determined whether the detection target belongs to the corresponding set category. If the detection target does not belong to any set category, the detection target is determined to be an interference target; the point cloud data of the interference target is filtered out. Through the above processing, the relationship between the point cloud data features of the target and the target distance is fully utilized to identify the interference target, so that the interference target in the millimeter wave radar detection can be effectively identified, and the accuracy of the millimeter wave radar detection can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1a A schematic diagram of the location of radar targets and point clouds projected onto the ground; Figure 1b It is a schematic diagram of the target track and point cloud position in the radar coordinate system; Figure 2a This is a schematic diagram of the statistical relationship between the number of radar point clouds of large vehicles and the target distance in traffic scenes; Figure 2b This is a schematic diagram of the statistical relationship between the number of radar point clouds of a car and the target distance in a traffic scene; Figure 3 The figure is a schematic diagram of the statistical relationship between the average SNR of the radar target point cloud of the large vehicle and the small vehicle and the target distance; Figure 4 Schematic diagram of the statistical relationship between the point cloud average SNR and target distance of the false detection target FP and the normal target of the car category; Figure 5 A schematic diagram of the basic flow of the interference filtering method for millimeter wave radar detection in this application; Figure 6 A schematic diagram of a specific process of the interference filtering method for millimeter wave radar detection in a specific embodiment of the present application; Figure 7 It is a diagram showing the relationship between the average value of CFAR points and the upper and lower boundary values ​​of the normal target of the car category and the target distance; Figure 8 It is a basic structural schematic diagram of the interference filtering device in this application; Fig. 9It is a schematic diagram of the basic structure of the electronic device in this application. DETAILED DESCRIPTION

[0024] In order to make the objectives, technical means and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings.

[0025] In the current scenarios where millimeter-wave radar is used for traffic monitoring, radar track information is usually obtained using millimeter-wave radar. In fact, radar can also obtain radar raw point cloud information, such as the coordinates, speed, signal-to-noise ratio (SNR), radar cross section (RCS) and other information of each point in the point cloud. The feature information of the entire point cloud can be obtained from the information of these points.

[0026] The applicant has found through research on radar point cloud information that different types of targets can be distinguished based on the distribution information of point cloud features at different target distances, and at the same time, falsely detected targets and normal targets can be distinguished.

[0027] Specifically, the distribution information of point cloud features at different target distances, that is, the relationship between point cloud features and different target distances. Take the relationship between the number of point clouds and average SNR of the target point cloud and the target distance as an example. Figure 2a is the statistical relationship between the number of radar point clouds and target distance of large vehicles in traffic scenes. Figure 2b The statistical relationship between the number of radar point clouds and target distance of a car in a traffic scene, where the horizontal axis represents the target distance, the vertical axis represents the number of constant false alarm rate (CFAR) points of the target point cloud, and the bar line represents the range of points of multiple targets of the same type at each target distance, and the point on the bar line represents the statistical mean of the points of multiple targets of the same type. Figure 2a and Figure 2b It can be seen from the comparison that large and small vehicle types can be effectively distinguished by CFAR points at various target distances. Figure 3 The statistical relationship between the average SNR and target distance of the radar target point cloud of large and small vehicles. This figure gives an intuitive comparison of two different types of targets. The upper bars and points represent the relationship between the SNR range and average SNR and target distance of large vehicles, and the lower bars and points represent the relationship between the SNR range and average SNR and target distance of small vehicles. Obviously, the SNR range and average SNR of large vehicles are higher than those of small vehicles. Large and small vehicle types can also be effectively distinguished by the average SNR at various target distances. The above situation shows the possibility of classifying different types of targets from the statistical values ​​of these point cloud features at different target distances.

[0028] Through a series of statistical findings by the applicant, other typical point cloud features, such as velocity consistency and RCS, can also effectively distinguish different types of targets. As can be seen from the above, different types of normal targets can be effectively distinguished through the relationship between these point cloud features and target distance. In this application, these point cloud features that can distinguish different types of targets are referred to as category features of point clouds. The normal target here refers to the actual target in radar detection, not the falsely detected target. Similarly, various types of normal targets and falsely detected targets can also be distinguished through the relationship between the category features of the point cloud and the target distance. Still taking the relationship between the average SNR and the target distance as an example, Figure 4 The statistical relationship between the average SNR of the point cloud and the target distance of the false detection target FP and the normal target of the car category is given. Among them, the line 1 of the average SNR and the corresponding column and point of the line correspond to the normal target of the car category, and the line 2 and the corresponding column and point correspond to the false detection target. It can be seen that the average SNR value of the multipath false detection target and the normal vehicle has a significantly different trend with the distance. The SNR value of the multipath false detection target shows an overall increasing trend with the increase of distance, while the SNR value of the normal vehicle target shows a decreasing trend. This illustrates the feasibility of distinguishing normal targets from noise targets. The relationship between the average SNR and the target distance can be used to distinguish false detection targets from normal targets. And through experimental statistics, it is found that for other point cloud category features, false detection targets and normal targets can be distinguished by their relationship with the target distance.

[0029] Based on the above analysis, this application proposes to use the relationship between point cloud category features and target distance to identify misdetected targets. Figure 5 This is a basic flow chart of the interference filtering method for millimeter wave radar detection in this application. Figure 5 As shown, the method includes: Step 501: Acquire point cloud data of the detection target through millimeter wave radar.

[0030] Here, the detection target refers to a target newly detected by the millimeter wave radar, and this method is used to determine whether the detection target is an interference target. The point cloud data can be obtained in frames, for example, the entire method can be performed on a single frame of point cloud data of the detection target.

[0031] Step 502 : determining the current category feature and the current target distance of the point cloud based on the point cloud data.

[0032] The point cloud data of the detected target is analyzed to determine its category features and target distance. To distinguish them from the category features and target distance of subsequent normal targets, the category features and target distance determined based on the point cloud data of the detected target are referred to as current category features and current target distance, respectively. The category features may be pre-set, such as the number of points in the target point cloud, the SNR of the target point cloud, the speed consistency of the target point cloud, or the RCS of the target point cloud. The number of points in the target point cloud may be the number of points in a certain type of point cloud generated after preprocessing the radar raw data, such as the number of CFAR points, etc. There may be a variety of different algorithms for CFAR detection, such as CA-CFAR, GO-CFAR, SO-CFAR, OS-CFAR, etc.

[0033] Step 503 , based on the current category feature and the current target distance, the scores of the detected target belonging to each set category are determined by comparing the relationship between the category feature and the target distance of the normal target in each set category of the current scene that is predetermined.

[0034] In this application, the relationship between the category characteristics and the target distance of the normal target under each set category is determined in advance for the current scene. Among them, each set category can refer to different classification information of the normal target in the current scene. For example, in the road traffic monitoring scene, the set categories of normal targets may include: large vehicles, small vehicles, non-motor vehicles, etc. As mentioned above, the normal target in this application refers to the actual target to distinguish it from the interference target.

[0035] For each set category, the relationship between the category characteristics of normal targets under the set category and the target distance is determined in advance, for example, the relationship between the SNR characteristics of normal targets of large vehicle type and the target distance is determined in advance, and the relationship between the SNR characteristics of normal targets of small vehicle type and the target distance is determined. At the same time, the current category characteristics and the current target distance of the detection target are determined through the aforementioned step 502. In this step, corresponding to each set category, referring to the relationship between the category characteristics of normal targets under the set category and the current target distance, and then combining the current category characteristics and the current target distance of the detection target, the score of the detection target belonging to the corresponding set category is determined, which is also called the score corresponding to the corresponding set category.

[0036] In addition, optionally, when there are multiple category features, such as both SNR features and point features, then for each category feature, the relationship between the category feature of a normal target and the target distance can be used to determine the score corresponding to the category feature belonging to each set category. For example, the score corresponding to the SNR feature and the score corresponding to the point feature belonging to each set category can be determined, and then the final score corresponding to each set category can be determined in combination with the scores corresponding to each category feature.

[0037] Step 504: Based on the score of step 503, it is determined whether the detection target belongs to each set category. If the detection target does not belong to any set category, the detection target is determined to be an interference target.

[0038] For each set category, it is determined whether the detection target belongs to the corresponding set category based on the score corresponding to the set category; for example, the score of the detection target belonging to the large vehicle type is determined by step 503, and this step determines whether the detection target belongs to the large vehicle type based on the score. The above judgment is performed for each set category. If the judgment result is that the detection target does not belong to any set category, that is, the current category feature of the detection target and the current target distance do not conform to the relationship between the category feature and the target distance of a normal target under any set category, then the detection target does not belong to any set category, and it is determined to be an interference target.

[0039] Step 505: filter out the point cloud data of the interfering target.

[0040] For the identified interference targets, their point cloud data are filtered out to remove the interference data, thereby effectively improving the accuracy of millimeter-wave radar detection.

[0041] At this point, the interference filtering method process in this application ends. Through the above method, the interference target in the millimeter wave radar can be effectively identified, and the target detection accuracy of the millimeter wave radar can be improved.

[0042] The specific implementation of this application is described below through specific embodiments. Figure 6 The following is a specific flow chart of the interference filtering method for millimeter wave radar detection in a specific embodiment of the present application. In which, a tunnel scene and the category features are SNR and CFAR points as an example for explanation, such as Figure 6 As shown, the method includes: Step 601 , statistically determine the relationship between the point cloud SNR and CFAR points of normal targets in each set category of the current scene and the target distance.

[0043] In this embodiment, in order to simplify the description, it is assumed that the set categories include large vehicles and small vehicles. For the categories of large vehicles and small vehicles, the specific statistical method for determining the relationship between the category characteristics and the target distance is the same. This embodiment mainly considers the CFAR point count and SNR characteristics of the radar target, and counts its distribution at different distances. Among them, the category characteristics of the CFAR point count can be reflected by the point mean and / or the upper and lower boundary values ​​of the point count of the target point cloud, which are referred to as category characteristic values; the category characteristic value of SNR may include the average SNR and / or the upper and lower boundary values ​​of the SNR of the target point cloud. Specifically in this embodiment, the relationship between the point mean, the upper and lower boundary values ​​of the point count, the average SNR and the upper and lower boundary values ​​of the SNR and the target distance of the target point cloud is statistically analyzed.

[0044] The statistical method can adopt various existing methods. Here, an exemplary statistical method is given. The synchronized radar-camera data is marked, that is, the point cloud data in each frame collected by the radar is marked correspondingly with the image data collected by the camera, and the set category of the target individual (that is, the target is a large car or a small car) is marked from the image data collected by the camera. The point cloud category feature value and point cloud distance of the corresponding target individual are recorded from the radar point cloud data, and the point cloud category feature value-point cloud distance-target type value of the target individual is recorded. Note that in order to remove some extreme values, in this embodiment, the value of each feature value in each distance range is arranged in the middle 20%-80%. Therefore, the upper and lower boundary values ​​of the CFAR points and SNR in the category feature value are not extreme values, but boundary values ​​within the range of 20%-80%. Statistical analysis is performed using the marked data of multiple normal targets at multiple target distances, and the relationship curve between the category feature value and the target distance is fitted. In this embodiment, the statistical results of the average CFAR point count and the upper and lower boundary values ​​of multiple small car targets in a tunnel scene and the target distance are obtained by linear regression fitting. Three curves are obtained, as shown in FIG. Figure 7 As shown, from top to bottom are the relationship curve between the upper boundary value of the CFAR points of the small car and the target distance, the relationship curve between the mean value of the CFAR points of the small car and the target distance, and the relationship curve between the lower boundary value of the CFAR points of the small car and the target distance; at the same time, the statistical results of the average value of the CFAR points of multiple large vehicle targets and the upper and lower boundary values ​​and the target distance are fitted using linear regression to obtain three curves, which are the relationship curve between the upper boundary value of the CFAR points of the large vehicle and the target distance, the relationship curve between the mean value of the CFAR points of the large vehicle and the target distance, and the relationship curve between the lower boundary value of the CFAR points of the large vehicle and the target distance. Similarly, the same method can also be used to obtain three relationship curves about SNR of the large vehicle and the small vehicle, namely, the relationship curve between the upper boundary value of the small vehicle point cloud SNR and the target distance, the relationship curve between the mean value of the small vehicle point cloud SNR and the target distance, the relationship curve between the lower boundary value of the small vehicle point cloud SNR and the target distance, the relationship curve between the upper boundary value of the large vehicle point cloud SNR and the target distance, the relationship curve between the mean value of the large vehicle point cloud SNR and the target distance, and the relationship curve between the lower boundary value of the large vehicle point cloud SNR and the target distance. Therefore, in this embodiment, 3 relationship curves of each of the 2 category features under 2 set categories are statistically obtained, that is, a total of 12 relationship curves. Note Figure 7 The fitting curve in the figure is explained by taking a straight line as an example. In fact, it can also be a polynomial curve or an exponential curve.

[0045] In addition, in the above-mentioned labeling process of comparing the synchronized radar-camera data, the coordinates in the radar coordinate system and the coordinates in the camera image coordinate system can be converted and corresponded to each other through RV calibration. Based on the RV calibration result, the correspondence between the radar target trajectory and the normal target trajectory in the camera image can be determined and labeled manually, but the labor cost and time cost are relatively high. In the present application, when labeling, the correspondence between the radar target trajectory determined based on the RV calibration result and the normal target trajectory determined by the camera for image acquisition can be automatically recorded. If the ratio of any radar target trajectory corresponding to the same normal target trajectory within a set time reaches a set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between target distance and category features (specifically, the corresponding normal target trajectory can be determined based on the binding relationship between the radar target trajectory and the normal target trajectory). The corresponding radar target and normal target are determined by the normal target, and the set category to which the target belongs is determined by using the normal target, and the relationship between the target distance and the category feature under the corresponding set category is determined by using the point cloud data corresponding to the radar target); alternatively, the binding confidence of each radar target trajectory and normal target trajectory pair (usually any radar target trajectory and any normal target trajectory can form a trajectory pair) can be maintained. If the corresponding relationship between a radar target trajectory and a normal target trajectory pair meets the preset condition, the binding confidence is increased, otherwise the binding confidence is reduced. When the binding confidence is greater than or equal to the set binding confidence threshold, the radar target trajectory and the normal target trajectory pair are bound; the preset condition can be that a radar target trajectory and a normal target trajectory pair are marked as corresponding in the corresponding relationship determined based on the RV calibration result, or it can also be that the ratio of the number of times a radar target trajectory and a normal target trajectory pair are marked as corresponding within a set time reaches a set binding threshold. In other words, it is possible to automatically select the matching pairs of RV tracks and image targets that are relatively stable and bound for a relatively long time, and use these radar tracks and corresponding point cloud data to automatically complete the marking of each set category target at each distance and the analysis and curve fitting processing of the corresponding data. Among them, RV calibration is the existing process of automatically mapping radar point cloud data to the camera captured image.

[0046] Step 602: Acquire point cloud data of the detection target through millimeter wave radar.

[0047] The processing of this step is the same as step 501 and will not be repeated here.

[0048] Step 603: Acquire the current category feature and current target distance of the detection target based on the point cloud data of the detection target.

[0049] For any radar detection target in the current frame, determine the distance of the target as the current target distance, and obtain the current CFAR point number of the corresponding point cloud and the current SNR mean , as the current category feature value.

[0050] Step 604, scoring the detected target for each set category.

[0051] This step is used to calculate the score of the detection target for each set category, and the calculation of the score of the detection target for the large vehicle category is used as an example for explanation.

[0052] A radar detection target and its corresponding point cloud data are obtained through step 602, and the current target distance of the detection target from the radar can be obtained through step 603, and the current CFAR points of its point cloud can be counted. And the current SNR mean (i.e. the average SNR of the detected target point cloud) Substituting the current target distance determined in step 603 into the relationship curves between the mean and target distance of two of the six curves of the large vehicle category obtained in step 601, the mean CFAR point number corresponding to the normal target of the large vehicle at this target distance can be obtained. and SNR mean Substitute the current target distance into the relationship curve between the other four upper and lower boundary values ​​of the vehicle and the target distance, and obtain the upper and lower boundary values ​​of the CFAR points and the upper and lower boundary values ​​of the SNR corresponding to the normal target of the vehicle at the current target distance. Difference between the upper and lower boundary values ​​of the CFAR points can estimate the fluctuation range of the CFAR points corresponding to the normal target of the vehicle at the current target distance. By subtracting the upper and lower boundary values ​​of SNR, we can estimate the fluctuation range of SNR corresponding to the current target distance of the normal target of the large vehicle. .

[0053] Next, the score of the detection target belonging to the large vehicle category is calculated based on the above data. In this embodiment, the category features include CFAR point features and SNR features. Therefore, when calculating the score, it is necessary to obtain the scores corresponding to the two category features respectively, and then determine the final score. Specifically, the score corresponding to the CFAR point feature (hereinafter referred to as the CFAR point score) and the score corresponding to the SNR feature (hereinafter referred to as the SNR score) that the detection target belongs to the large vehicle category can be obtained for the CFAR point feature and the SNR feature respectively. For the CFAR point feature, the current point count of the detection target point cloud can be The average number of points corresponding to the current target distance Compare, based on the comparison results and the point fluctuation range , calculate the score of the detected target belonging to the large vehicle category. A more detailed score calculation method can be designed according to the actual scene requirements. For the SNR feature, the current SNR mean of the detected target point cloud can be The mean SNR value corresponding to the current target distance Compare and compare based on the comparison results and SNR fluctuations , calculate the SNR score of the detected target belonging to the large vehicle category. Finally, the final score of the detected target belonging to the large vehicle category is determined based on the CFAR point score and the SNR score.

[0054] The specific calculation rules of CFAR score, SNR score and final score can be designed according to the actual scenario requirements. In this embodiment, an exemplary method is given. First, the CFAR score of the target is calculated respectively. and SNR score , the score ranges from 0 to 1, and the specific score calculation method is as follows: , .

[0055] Among them, α and β are fluctuation threshold coefficients. Generally, the fluctuation threshold can be slightly larger than the actual range, so α and β are less than 1 and greater than 0. Qualitatively, the closer the current actual CFAR points or SNR of the detected target is to the mean of the normal target, the closer its score is to 1. The current true value is allowed to differ from the average value of the normal target by at most α and β fluctuation amplitudes at the same distance. If it exceeds this amplitude, the score is set to 0. In other words, the score represents the degree of consistency between the current category feature value and the mean category feature value within the fluctuation range of the normal target at the same target distance, that is, the similarity between the detected target and the normal target.

[0056] In this embodiment, CFAR point features and SNR features are used as category features, so the final detection target belongs to the vehicle category score Can be scored based on CFAR points and SNR score Determine, for example, Similarly, using the statistical curve of the car's point cloud category features and the target distance, we can calculate the score of the detected target as a car category. .

[0057] Step 605: determine whether the detection target belongs to each set category based on the score of each set category. If it does not belong to any set category, determine that the detection target is an interference target.

[0058] When determining whether a detection target belongs to a certain set category, the following two methods can be used: Method 1: If the score of the detection target belonging to a certain set category is less than the score threshold corresponding to the set category, it is determined that the detection target does not belong to the set category; otherwise, it is determined that the detection target belongs to the set category; if the detection target does not belong to any set category, it is determined that the detection target is an interference target.

[0059] In this way, single-frame point cloud data can be directly used to determine whether the detected target is an interference target, and the identification of interference targets can meet real-time requirements.

[0060] Method 2: For each set category, maintain a confidence level that the detection target belongs to the category. The initial value of the confidence level can be pre-set based on experience. If the score of the detection target belonging to a set category is less than the score threshold corresponding to the set category, the confidence level that the detection target belongs to the set category is reduced. After the confidence level is lower than the confidence threshold of the corresponding set category, it is determined that the detection target does not belong to the set category; if the detection target does not belong to any set category, it is determined that the detection target is an interference target.

[0061] In this way, it may be necessary to use multi-frame point cloud data to determine whether the detected target is an interference target, but the identification of the interference target can be more reliable.

[0062] Through tests in actual scenarios, it is found that the interference target identification accuracy of method one can meet the general requirements and can also effectively meet the real-time requirements. Therefore, method one can be selected to determine the interference target in ordinary scenarios; for scenarios with very high requirements for the reliability of interference identification, method two can be considered to determine the interference target.

[0063] In addition, in the above judgment process, if the detection target belongs to multiple set categories, the set category with the highest score is selected as the final target category. In this way, the classification of the detection target can be realized simultaneously in the process of interference identification.

[0064] Step 606: Filter out the point cloud data of the interfering target.

[0065] The processing of this step is the same as step 505, and will not be repeated here.

[0066] So far, Figure 6 The method flow shown ends. It should be noted that, in addition to the target CFAR point count and SNR features in the above specific embodiments, other point cloud velocity consistency, RCS and other features can also be used as category features for interference target identification. These category features can be used alone for interference target identification, or multiple category features can be used simultaneously (such as using two category features simultaneously in the specific embodiment) to classify the detection target more accurately. Through the above Figure 6In the interference identification stage, each newly started track and corresponding detection target are scored according to the above steps to determine whether it is a false detection. If it is a false detection, the point cloud data corresponding to the track is filtered out, and the track is no longer included in the subsequent output of the filtered radar signal. In this way, interference filtering in millimeter-wave radar can be effectively realized, the accuracy of target detection can be improved, and it can be operated online in real time, completing target identification and filtering in the initial frame where the multipath target appears.

[0067] The above-mentioned method of the present application can be applied in a variety of scenarios, especially in tunnels or road traffic monitoring scenarios with high-rise buildings and obstructions on both sides.

[0068] The above is a specific implementation of the interference filtering method for millimeter wave radar detection in the present application. The present application also provides an interference filtering device, which can be used to implement the above interference filtering method. Figure 8 Schematic diagram of the basic structure of the interference filtering device in this application. Figure 8 As shown, the device includes: a data acquisition unit, a scoring unit and an interference identification unit.

[0069] The data acquisition unit is used to acquire point cloud data of the detection target through the millimeter wave radar, and determine the current category characteristics of the point cloud and the current target distance based on the point cloud data; A scoring unit, for determining the scores of the detected targets belonging to the respective set categories based on the current category features and the current target distance and comparing the predetermined relationship between the category features and the target distance of the normal targets in the respective set categories of the current scene; The interference identification unit is used to determine whether the detection target belongs to each set category based on the score. If the detection target does not belong to any set category, the detection target is determined to be an interference target; it is also used to filter out the point cloud data of the interference target.

[0070] Optionally, the device further includes a statistical unit for predetermining the relationship between the category characteristics of normal targets under each set category of the current scene and the target distance; The method for determining the relationship between the category feature of a normal target and the target distance under any set category feature of the current scene includes: The category characteristic values ​​of the target point cloud of multiple normal targets under any set category at various target distances are statistically analyzed, and the relationship curve between the category characteristic value of the normal target point cloud and the target distance under any set category is fitted.

[0071] Optionally, the category feature is a point count feature of the target point cloud; The category feature values ​​include: the mean value of the number of points and the upper and lower boundary values ​​of the number of points of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: the relationship curve between the mean value of the number of points of the normal target point cloud and the target distance and the relationship curve between the upper and lower boundary values ​​of the number of points of the normal target point cloud and the target distance; In the scoring unit, a method for determining whether the detection target belongs to any of the set categories includes: Based on the relationship curve between the mean number of points of the normal target point cloud and the target distance under any set category, determine the mean number of points corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of the points of the normal target point cloud and the target distance under any set category, the upper and lower boundary values ​​of the points corresponding to the current target distance are determined, and the fluctuation amplitude of the points corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of the points ; The current number of points in the target point cloud will be detected The average number of points corresponding to the current target distance Compare, based on the comparison results and the said point fluctuation range , calculating the score of the detected target belonging to any of the set categories; A score is determined based on the point score for the detection target belonging to any of the set categories.

[0072] Optionally, the category feature is an SNR feature of the target point cloud; The category feature values ​​include: the SNR mean and the SNR upper and lower boundary values ​​of the target point cloud; The relationship curve between the category feature value of the normal target point cloud and the target distance includes: the relationship curve between the SNR mean value of the normal target point cloud and the target distance and the relationship curve between the SNR upper and lower boundary values ​​of the normal target point cloud and the target distance; The method of determining whether the detection target belongs to any of the set categories includes: Based on the relationship curve between the SNR mean of the normal target point cloud and the target distance under any set category, determine the SNR mean corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of SNR and the target distance of the normal target point cloud under any set category, the upper and lower boundary values ​​of SNR corresponding to the current target distance are determined, and the SNR fluctuation amplitude corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of SNR ; The current SNR mean of the target point cloud will be detected The mean SNR value corresponding to the current target distance Compare and compare based on the comparison results and SNR fluctuations , calculate the SNR score of the detected target belonging to any of the set categories; The score of the detected target belonging to any set category is determined based on the SNR score.

[0073] Optionally, in the interference identification unit, for any set category in each set category, a method of judging whether the detection target belongs to any set category based on a score that the detection target belongs to any set category includes: If the score of the detection target belonging to any set category is less than the score threshold corresponding to any set category, it is determined that the detection target does not belong to any set category; otherwise, it is determined that the detection target belongs to any set category; or, If the score of the detection target belonging to any of the set categories is less than the score threshold corresponding to any of the set categories, the confidence that the detection target belongs to any of the set categories is lowered. After the confidence is lower than the confidence threshold of any of the set categories, it is determined that the detection target does not belong to any of the set categories.

[0074] Optionally, when the statistical unit pre-statistics the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory calibrated by the RV and the normal target trajectory in the image acquisition is recorded. If the ratio of any radar target trajectory corresponding to the same normal target trajectory within the set time reaches a set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; or, When pre-stating the relationship between the category characteristics and target distance of normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory calibrated by RV and the normal target trajectory is recorded. If the corresponding relationship between any radar target trajectory and the normal target trajectory pair meets the preset condition, the binding confidence maintained for any radar target trajectory and the normal target trajectory pair is increased, otherwise, the binding confidence is reduced; when the binding confidence is greater than or equal to the binding confidence threshold, any radar target trajectory is bound to the normal target trajectory pair, and the bound radar target trajectory and the normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; wherein the preset condition is: any radar target trajectory and the normal target trajectory pair are marked as corresponding in the RV mark, or the ratio of the number of corresponding times between any radar target trajectory and the normal target trajectory pair within the set time is greater than the binding threshold.

[0075] The present application also provides a computer-readable storage medium that stores instructions, which, when executed by a processor, can execute the steps in the interference filtering method for implementing millimeter-wave radar detection as described above. In practical applications, the computer-readable medium can be included in each device / apparatus / system in the above-mentioned embodiments, or it can exist independently without being assembled into the device / apparatus / system. Instructions are stored in a computer-readable storage medium, and the instructions stored therein can execute the steps in the interference filtering method for millimeter-wave radar detection as described above when executed by a processor.

[0076] According to the embodiments disclosed in the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof, but it is not used to limit the scope of protection of the present application. In the embodiments disclosed in the present application, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0077] Fig. 9 This application also provides an electronic device. Fig. 9 As shown, it shows a schematic diagram of the structure of the electronic device involved in the embodiment of the present application, specifically: The electronic device may include a processor 901 with one or more processing cores, a memory 902 with one or more computer-readable storage media, and a computer program stored in the memory and executable on the processor. When executing the program in the memory 902, an interference filtering method for millimeter wave radar detection may be implemented.

[0078] Specifically, in practical applications, the electronic device may further include components such as a power supply 903, an input / output unit 904, etc. Those skilled in the art may understand that Fig. 9 The structure of the electronic device shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. Processor 901 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. It executes various functions of the server and processes data by running or executing software programs and / or modules stored in memory 902 and calling data stored in memory 902, thereby monitoring the electronic device as a whole.

[0079] The memory 902 can be used to store software programs and modules, that is, the above-mentioned computer-readable storage medium. The processor 901 executes various functional applications and data processing by running the software programs and modules stored in the memory 902. The memory 902 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the server, etc. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 902 may also include a memory controller to provide the processor 901 with access to the memory 902.

[0080] The electronic device also includes a power supply 903 for supplying power to various components, which can be logically connected to the processor 901 through a power management system, so as to manage charging, discharging, power consumption and other functions through the power management system. The power supply 903 can also include any components such as one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, and power status indicators.

[0081] The electronic device may also include an input and output unit 904, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, and optical signal inputs related to user settings and function control. The input unit output 904 may also be used to display information input by the user or information provided to the user and various graphical user interfaces, which may be composed of graphics, text, icons, videos, and any combination thereof.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for filtering interference in millimeter wave radar detection, characterized in that: include: Acquire point cloud data of the detection target through millimeter wave radar; Determining a current category feature and a current target distance of the point cloud based on the point cloud data; Based on the current category feature and the current target distance, comparing the relationship between the category feature and the target distance of the normal target in each set category of the current scene, the score of the detected target belonging to each set category is determined; Determining whether the detection target belongs to each of the set categories based on the score, and if the detection target does not belong to any of the set categories, determining that the detection target is an interference target; The point cloud data of the interfering target is filtered out.

2. The method according to claim 1, characterized in that The category features include: point number features of the target point cloud, signal-to-noise ratio (SNR) features of the target point cloud, velocity consistency features of the target point cloud and / or radar cross section (RCS) features.

3. The method according to claim 1, characterized in that The method of determining the relationship between the category feature of a normal target under any set category of the current scene and the target distance includes: Statistical analysis is performed on the category characteristic values ​​of the target point cloud of multiple normal targets under any set category at multiple target distances, and a relationship curve between the category characteristic value of the normal target point cloud and the target distance under any set category is obtained by fitting.

4. The method according to any one of claims 1 to 3, characterized in that: When there are multiple category features, determining the scores of the detection target belonging to each set category includes: For any category feature, based on the relationship between the any category feature of the normal target under each set category and the target distance, determine the score corresponding to the any category feature of each set category that the detected target belongs to; The scores corresponding to the features of each category are comprehensively considered to determine the scores of the detection target belonging to each set category.

5. The method according to claim 3, characterized in that: The category feature is a point number feature of the target point cloud; The category feature values ​​include: the mean value of the number of points and the upper and lower boundary values ​​of the number of points of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the mean value of the number of points of the normal target point cloud and the target distance and a relationship curve between the upper and lower boundary values ​​of the number of points of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on the relationship curve between the mean number of points of the normal target point cloud and the target distance under any set category, determine the mean number of points corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of the points of the normal target point cloud under any set category and the target distance, the upper and lower boundary values ​​of the points corresponding to the current target distance are determined, and the fluctuation amplitude of the points corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of the points ; The current number of points in the detection target point cloud The mean of the number of points corresponding to the current target distance Compare, based on the comparison results and the said point fluctuation range , calculating the score corresponding to the point feature of the detection target belonging to any of the set categories; The score of the detection target belonging to any of the set categories is determined based on the score corresponding to the point feature.

6. The method according to claim 3 or 5, characterized in that: The category feature is the SNR feature of the target point cloud; The category feature values ​​include: the SNR mean and the SNR upper and lower boundary values ​​of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the SNR mean value of the normal target point cloud and the target distance and a relationship curve between the SNR upper and lower boundary values ​​of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on the relationship curve between the SNR mean of the normal target point cloud and the target distance under any set category, determine the SNR mean corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of the SNR of the normal target point cloud under any set category and the target distance, the upper and lower boundary values ​​of the SNR corresponding to the current target distance are determined, and the SNR fluctuation amplitude corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of the SNR ; The current SNR mean of the detected target point cloud The mean SNR value corresponding to the current target distance The comparison is based on the comparison result and the SNR fluctuation amplitude. , calculating the score corresponding to the SNR feature of the detected target belonging to any of the set categories; The score of the detected target belonging to any of the set categories is determined based on the score corresponding to the SNR feature.

7. The method according to claim 1, characterized in that For any one of the set categories, a method of judging whether the detection target belongs to any one of the set categories based on the score that the detection target belongs to the any one of the set categories includes: If the score of the detection target belonging to any of the set categories is less than the score threshold corresponding to any of the set categories, it is determined that the detection target does not belong to any of the set categories; otherwise, it is determined that the detection target belongs to any of the set categories; or, If the score that the detection target belongs to any of the set categories is less than the score threshold corresponding to any of the set categories, the confidence that the detection target belongs to any of the set categories is lowered. After the confidence is lower than the confidence threshold of any of the set categories, it is determined that the detection target does not belong to any of the set categories.

8. The method according to claim 1, characterized in that When pre-stating the relationship between the category characteristics and target distance of normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the ratio of any radar target trajectory to the same normal target trajectory within the set time reaches the set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; or, When pre-stating the relationship between the category characteristics and target distance of normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the corresponding relationship between any radar target trajectory and the normal target trajectory pair meets the preset conditions, the binding confidence maintained for any radar target trajectory and the normal target trajectory pair is increased, otherwise, the binding confidence is reduced; When the binding confidence is greater than or equal to the binding confidence threshold, any one of the radar target trajectories is bound to the normal target trajectory pair, and the bound radar target trajectory and the normal target trajectory are used to perform statistics on the relationship between the target distance and the category feature; wherein the preset condition is: any one of the radar target trajectory and the normal target trajectory pair is marked as corresponding in the corresponding relationship determined based on the RV calibration result, or, within a set time, the ratio of the number of times that any one of the radar target trajectory and the normal target trajectory pair is marked as corresponding is greater than the binding threshold.

9. An interference filtering device for millimeter wave radar detection, characterized in that: include: data acquisition unit, scoring unit and interference identification unit; The data acquisition unit is used to acquire point cloud data of the detection target through the millimeter wave radar, and determine the current category characteristics of the point cloud and the current target distance based on the point cloud data; The scoring unit is used to determine the scores of the detected target belonging to each set category based on the current category feature and the current target distance, and in comparison with the predetermined relationship between the category feature and the target distance of the normal target under each set category of the current scene; The interference identification unit is used to determine whether the detection target belongs to the set categories based on the score, and if the detection target does not belong to any set category, determine that the detection target is an interference target; and is also used to filter out the point cloud data of the interference target.

10. The device according to claim 9, characterized in that The device further includes a statistical unit for predetermining the relationship between the category characteristics of normal targets under each set category of the current scene and the target distance; The method for determining the relationship between the category feature of a normal target and the target distance under any set category feature of the current scene includes: Statistical analysis is performed on the category characteristic values ​​of the target point cloud of multiple normal targets under any set category at multiple target distances, and a relationship curve between the category characteristic value of the normal target point cloud and the target distance under any set category is obtained by fitting.

11. The device according to claim 9 or 10, characterized in that In the scoring unit, when there are multiple category features, determining the scores of the detection target belonging to each set category includes: For any category feature, based on the relationship between the any category feature of the normal target under each set category and the target distance, determine the score corresponding to the any category feature of each set category that the detected target belongs to; The scores corresponding to the features of each category are comprehensively considered to determine the scores of the detection target belonging to each set category.

12. The device according to claim 10, characterized in that The category feature is a point number feature of the target point cloud; The category feature values ​​include: the mean value of the number of points and the upper and lower boundary values ​​of the number of points of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the mean value of the number of points of the normal target point cloud and the target distance and a relationship curve between the upper and lower boundary values ​​of the number of points of the normal target point cloud and the target distance; In the scoring unit, a method of determining whether the detection target belongs to any of the set categories includes: Based on the relationship curve between the mean number of points of the normal target point cloud and the target distance under any set category, determine the mean number of points corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of the points of the normal target point cloud under any set category and the target distance, the upper and lower boundary values ​​of the points corresponding to the current target distance are determined, and the fluctuation amplitude of the points corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of the points ; The current number of points in the detection target point cloud The mean of the number of points corresponding to the current target distance Compare, based on the comparison results and the said point fluctuation range , calculating the score of the detected target belonging to any of the set categories; A score is determined based on the point score for the detection target belonging to any of the set categories.

13. The device according to claim 10 or 12, characterized in that The category feature is the SNR feature of the target point cloud; The category feature values ​​include: the SNR mean and the SNR upper and lower boundary values ​​of the target point cloud; The relationship curve between the category characteristic value of the normal target point cloud and the target distance includes: a relationship curve between the SNR mean value of the normal target point cloud and the target distance and a relationship curve between the SNR upper and lower boundary values ​​of the normal target point cloud and the target distance; The method of determining the score of the detection target belonging to any of the set categories includes: Based on the relationship curve between the SNR mean of the normal target point cloud and the target distance under any set category, determine the SNR mean corresponding to the current target distance ; Based on the relationship curve between the upper and lower boundary values ​​of the SNR of the normal target point cloud under any set category and the target distance, the upper and lower boundary values ​​of the SNR corresponding to the current target distance are determined, and the SNR fluctuation amplitude corresponding to the current target distance is calculated based on the upper and lower boundary values ​​of the SNR ; The current SNR mean of the detected target point cloud The mean SNR value corresponding to the current target distance The comparison is based on the comparison result and the SNR fluctuation amplitude. , calculating the SNR score of the detected target belonging to any of the set categories; A score is determined based on the SNR score, indicating that the detected target belongs to any one of the set categories.

14. The device according to claim 9, characterized in that In the interference identification unit, for any one of the set categories, a method of judging whether the detection target belongs to any one of the set categories based on the score that the detection target belongs to the any one of the set categories includes: If the score of the detection target belonging to any of the set categories is less than the score threshold corresponding to any of the set categories, it is determined that the detection target does not belong to any of the set categories; otherwise, it is determined that the detection target belongs to any of the set categories; or, If the score that the detection target belongs to any of the set categories is less than the score threshold corresponding to any of the set categories, the confidence that the detection target belongs to any of the set categories is lowered. After the confidence is lower than the confidence threshold of any of the set categories, it is determined that the detection target does not belong to any of the set categories.

15. The device according to claim 9, characterized in that When the statistical unit pre-statistics the relationship between the category characteristics and the target distance of the normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the ratio of any radar target trajectory to the same normal target trajectory within a set time reaches a set binding threshold, the radar target trajectory is bound to the same normal target trajectory, and the bound radar target trajectory and normal target trajectory are used to perform statistics on the relationship between the target distance and the category characteristics; or, When pre-stating the relationship between the category characteristics and target distance of normal targets under each set category of the current scene, the corresponding relationship between the radar target trajectory determined based on the RV calibration result and the normal target trajectory in the image acquisition is recorded. If the corresponding relationship between any radar target trajectory and the normal target trajectory pair meets the preset conditions, the binding confidence maintained for any radar target trajectory and the normal target trajectory pair is increased, otherwise, the binding confidence is reduced; When the binding confidence is greater than or equal to the binding confidence threshold, any one of the radar target trajectories is bound to the normal target trajectory pair, and the bound radar target trajectory and the normal target trajectory are used to perform statistics on the relationship between the target distance and the category feature; wherein the preset condition is: any one of the radar target trajectory and the normal target trajectory pair is marked as corresponding in the corresponding relationship determined based on the RV marking result, or, within a set time, the ratio of the number of times that any one of the radar target trajectory and the normal target trajectory pair is marked as corresponding is greater than the binding threshold.

16. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by the processor, the interference filtering method for millimeter wave radar detection described in any one of claims 1 to 8 can be implemented.

17. An electronic device, characterized in that: The electronic device includes at least a computer-readable storage medium and also includes a processor; The processor is used to read executable instructions from the computer-readable storage medium and execute the instructions to implement the interference filtering method for millimeter-wave radar detection as described in any one of claims 1 to 8 above.

Images