A radar false track suppression method in traffic environment

Abstract

The application discloses a false track suppression method in a traffic environment. The method first calculates a characteristic variable correlation threshold of a false track and a real track through prior track data information statistics. Then, the correlation of a temporary track and a formed certain stable track is judged through multi-frame accumulation. After that, whether the temporary track turns into an explicit stable track or an implicit track is determined according to the correlation. If the temporary track turns into an implicit track, the correlation of the implicit track and the certain stable track is judged through multi-frame accumulation. Finally, whether the implicit track is deleted is determined through an implicit stable track gap change trend and whether the number of related frames. The application utilizes the characteristics that the characteristics of the same target track are similar and the characteristics of different target tracks cannot be kept. Through a multi-frame accumulation method for the characteristic differences between the tracks, the false track problem caused by dense vehicles and large vehicle type differences in the traffic environment is effectively solved. The method can be applied to the field of radar track processing.

Description

Technical Field

[0001] This invention belongs to the field of radar data processing technology, and specifically relates to a method for suppressing false radar tracks in traffic environments. Background Technology

[0002] Currently, millimeter-wave radar based on the continuous wave system is increasingly widely used in the transportation field due to its high resolution, all-weather operation, small size, and reasonable price. It can not only achieve real-time positioning and tracking of multiple lanes and multiple targets in high-density scenarios, but also accurately statistically analyze road condition and vehicle information based on radar data, such as average vehicle speed per lane, lane traffic flow, lane occupancy, and queue length. Furthermore, with even higher resolution, it can achieve vehicle identification and classification.

[0003] A crucial prerequisite for achieving the above functions is that the radar can stably track dense vehicle targets in real time, i.e., it has accurate track data. However, millimeter-wave radar has a high resolution, and after CFAR processing, a target will still have multiple echo points. In this case, the number of points and tracks is not one-to-one, and multiple points need to be correctly matched with a certain track. Otherwise, a considerable number of points can easily cause multiple batches of false tracks to be generated.

[0004] To prevent the aforementioned phenomena and reduce subsequent processing pressure, point data aggregation is necessary before track processing. Currently, commonly used aggregation algorithms include sliding window and clustering. Both algorithms require a aggregation threshold; points within this threshold are considered to originate from the same target. However, vehicles on the road are diverse in type and shape. Large trucks and semi-trailers can reach lengths of over 20 meters, while ordinary cars are only 5-6 meters. In radar echo points, large vehicle tracks are more fragmented, with larger spans between points, often exceeding the length of an ordinary car and the distance between vehicles. This presents a challenge in selecting the aggregation threshold. If the threshold is too high, it easily merges parallel or sequentially moving target echoes with similar motion states into one, causing the disappearance of true target tracks and potentially leading to navigation interruptions or track loss during subsequent track processing. If the threshold is too low, it results in excessive processing burden and too many batches of false tracks. This problem has not yet been fully resolved.

[0005] A crucial step in track processing is data association. Among existing data association algorithms, the Joint Probabilistic Data Interconnection (JPDA) algorithm proposed by the Bar-Shalom team and the Multiple Hypothesis Tracking (MHT) method proposed by Reid can achieve good results in tracking dense targets. However, these two algorithms have excessive computational costs. In engineering implementation, due to the limitations of low cost and small size, the track processing platforms selected by millimeter-wave radar currently have limited computing power and storage space. Such small computing platforms are not suitable for the above two algorithms. Therefore, the current association algorithm generally selects the nearest neighbor (NN) method and its variants, which have lower computational costs. However, this association algorithm also cannot solve the problem of multiple batches of false tracks generated by large vehicles.

[0006] In summary, current track processing algorithms applicable to millimeter-wave radar cannot effectively solve the problem of multiple batches of false tracks caused by excessive and scattered track points. Summary of the Invention

[0007] The purpose of this invention is to propose a method for suppressing false radar tracks in traffic environments. The technical problem to be solved is that, considering that each confirmed stable track is initially transformed from a temporary track after the track initiation conditions are met, and multiple batches of false tracks from the same target are no exception, it is too arbitrary and has limited accuracy to judge whether a temporary track is related to a certain confirmed stable track in the frame at the beginning of the temporary track.

[0008] The technical solution adopted in this invention is as follows:

[0009] This invention relates to a method for suppressing false tracks in traffic environments, comprising the following steps:

[0010] S1. Determine the threshold values ​​for the characteristic variables of false and true tracks.

[0011] S2. Determine the correlation between the temporary flight and the established stable flight during the period from the establishment of the temporary flight to the satisfaction of the flight path start conditions based on the relevant thresholds of the characteristic variables.

[0012] S3. Determine whether the temporary flight can normally initiate a visible stable track based on the correlation. If yes, proceed to step S6; otherwise, proceed to step S4.

[0013] S4. Convert the temporary flight path into a hidden track and continue to determine the correlation between the hidden track and the determined stable flight path over the subsequent preset number of frames and weeks.

[0014] S5. Determine whether the hidden track can be converted into a visible stable track based on the degree of relevance. If yes, proceed to step S6; otherwise, proceed to step S7.

[0015] S6. Convert the temporary or hidden track into a visible stable track.

[0016] S7. Delete hidden tracks.

[0017] As a further aspect of the present invention: step S1 includes:

[0018] S11. Based on the current prior track data, calculate the set of absolute differences between the feature data S = {X, Y, V} of multiple tracks from the same target within M frames. Where X and Y are the x and y coordinates of the target in Cartesian coordinates, respectively, and V is the radial velocity. and These are the absolute values ​​of the differences in X, Y, and V between multiple tracks of target t in the k-th frame.

[0019] S12. After several frames, arrange the absolute values ​​of the differences of all the feature data that have appeared in ascending order, count the number of occurrences of each value, build a distribution histogram, calculate the probability of each value, and select the absolute values ​​of the differences of the feature data corresponding to a cumulative distribution function value of 90% as the selected values ​​of the thresholds for each feature variable, thus obtaining the threshold set G. thres ={ΔX,ΔY,ΔV}, and use this set of threshold values ​​as the correlation threshold values ​​between the feature variables of false tracks and true tracks.

[0020] As a further aspect of the present invention: step S2 includes:

[0021] S21. Establishing a new temporary navigation from free points.

[0022] S22, Traverse all established stable navigation routes Calculate the absolute value of the difference between it and the new temporary flight characteristic variable.

[0023] S23, When the absolute value of the difference When the value in the interval is not greater than the relevant threshold value of the feature variable, the corresponding established stable flight path is taken as the temporary stable flight path of the new temporary flight path, the value of the relevant frame number is incremented by 1, and the absolute value of the difference is normalized to obtain the normalized difference value.

[0024] S24. Calculate the overall relevance. The formula is as follows:

[0025]

[0026] In the formula, λ1 and λ2 are both no greater than 1, and are the differences between eigenvalues, respectively. and The proportion of overall relevance.

[0027] As a further aspect of the present invention: In step S23, when the absolute value of the difference... If the value in the value is greater than the threshold value of the feature variable, a new temporary flight is re-established in the free point track of the next frame for judgment, until a new temporary flight meets the track start condition.

[0028] As a further aspect of the present invention: in step S23, the normalization formula is as follows:

[0029]

[0030] As a further aspect of the present invention: step S3 includes:

[0031] S31. When a temporary flight reaches the trajectory start condition, the sum of the relevant frames of all stable flights related to the temporary flight and the sum of the comprehensive correlation are calculated.

[0032] S32. Determine whether the sum of the number of stable flight-related frames reaches the preset frame value and whether the sum of the correlation reaches the preset correlation value. If so, convert it into an invisible flight path for management. Otherwise, convert this temporary flight path into a visible normal stable flight path for tracking and management.

[0033] As a further aspect of the present invention: the preset frame value and preset correlation value are related to the characteristics of the original point trace. When the front-end processing algorithm and radar front-end parameters change, the characteristics of the original point trace change, and the preset frame value and preset correlation value are increased or decreased accordingly.

[0034] As a further aspect of the present invention: In step S23: if multiple stable navigation conditions are satisfied... The two stable flight paths with the highest overall correlation are selected as the relevant stable flight paths for this temporary flight path, and the number of relevant frames is increased by 1 for each.

[0035] As a further aspect of the present invention: step S4 includes:

[0036] After a subsequent preset number of frames, determine whether the feature difference between the stealth track and a certain stable track has an increasing trend, or whether the number of unrelated frames is greater than half of the preset number of frames.

[0037] As a further aspect of the present invention: In step S5, when the feature difference between the stealth track and a certain stable track shows an increasing trend, or the number of unrelated frames is greater than half of the preset frame period, it is determined that the track comes from a real target, and the stealth track is converted into an explicit stable track for management.

[0038] The beneficial effects of this invention are as follows: This invention provides a false track suppression method based on multi-frame accumulation in traffic environments. The method first statistically calculates the correlation threshold between the characteristic variables of false tracks and real tracks based on prior track data. Then, this correlation threshold is set as a preset correlation threshold, and the correlation between temporary voyages and established stable voyages is determined through multi-frame accumulation. Next, based on the correlation, it is determined whether the temporary voyage transforms into an explicit stable voyage or a hidden track. If a hidden track exists, the correlation between the hidden track and the established stable voyage is further determined through multi-frame accumulation. Finally, the trend of the difference between the two and the number of uncorrelated frames are used to determine whether the hidden track should be eliminated. This invention utilizes the characteristic that tracks with the same target have similar features, while tracks with different targets have uncorrelated features. By accumulating the difference between tracks over multiple frames, it effectively solves the problem of false tracks caused by dense vehicle traffic and large differences in vehicle types in traffic environments. Engineering practice has verified that this method is effective, enabling rapid initiation and effectively reducing the false alarm rate. Attached Figure Description

[0039] The invention will now be further described with reference to the accompanying drawings.

[0040] Figure 1 This is a schematic diagram of traffic radar installation.

[0041] Figure 2 This is a flowchart of a method for suppressing false flight paths in a traffic environment.

[0042] Figure 3 Logical diagram of false flight path suppression methods in traffic environments

[0043] Figure 4 This is a flowchart for determining the correlation between temporary stability and flight status.

[0044] Figure 5 This is a diagram showing the result of using the normal track processing algorithm.

[0045] Figure 6 This is a diagram showing the results of track processing after incorporating this method. Detailed Implementation

[0046] To facilitate understanding of the technical content of the present invention by those skilled in the art, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0047] In all embodiments of the present invention, the radar is side-mounted at the intersection to monitor and track vehicle targets in real time within a 30m-250m road segment ahead. For details, please refer to... Figure 1 As shown. Since radar is typically installed at a height of about 7 meters, this radar has a blind zone of about 30 meters, and 250 meters is the farthest range at which the radar spot is stably illuminated. Figure 2 A flowchart of a method for suppressing false tracks in traffic environments is presented, based on... Figure 2As shown, the method for suppressing false tracks in a traffic environment according to the present invention includes the following steps:

[0048] S1: Determine the threshold values ​​for the correlation of characteristic variables between false and true tracks.

[0049] Step S1 includes the following steps S11-S12.

[0050] S11. Based on the current prior track data, calculate the set of absolute differences between the feature data S = {X, Y, V} of multiple tracks from the same target within M frames. Where X and Y are the x and y coordinates of the target in Cartesian coordinates, respectively, and V is the radial velocity. and These are the absolute values ​​of the differences in X, Y, and V between multiple tracks of target t in the k-th frame.

[0051] In practice, based on the current trajectory processing algorithm, 1000 frames of trajectory data are selected, and the differences between various features S = {X, Y, V}, such as X coordinate, Y coordinate, and velocity, among multiple trajectories from the same target during this period are statistically analyzed. And store it.

[0052] S12. After several frames, arrange the absolute values ​​of the differences of all the feature data that have appeared in ascending order, count the number of occurrences of each value, build a distribution histogram, calculate the probability of each value, and select the absolute values ​​of the differences of the feature data corresponding to a cumulative distribution function value of 90% as the selected values ​​of the thresholds for each feature variable, thus obtaining the threshold set G. thres ={ΔX,ΔY,ΔV}, and use this set of threshold values ​​as the correlation threshold values ​​between the feature variables of false tracks and true tracks.

[0053] In practice, after 1000 frames, the differences of all occurrences of each feature value are arranged in ascending order. The frequency of each value is then counted, a distribution histogram is created, and the probability of each value is calculated. The feature value difference corresponding to a cumulative distribution function value of 90% is selected as the threshold value for each feature variable, ultimately yielding the threshold set G. thres ={ΔX thres ,ΔY thres ,ΔV thres This set of threshold values ​​will be used as the preset gate for subsequent transient stabilization navigation.

[0054] S2: Determine the correlation between the temporary flight and the established stable flight during the period from the establishment of the temporary flight to the satisfaction of the flight path start conditions based on the relevant thresholds of the characteristic variables.

[0055] Step S2 includes the following steps S21-S24, as shown in the logic diagram below. Figure 3 As shown.

[0056] S21. Establish a new temporary flight from free points (i.e., points not related to the stable flight or the established temporary flight).

[0057] S22, Traverse all established stable navigation routes Calculate the absolute value of the difference between it and the new temporary flight characteristic variable.

[0058]

[0059] S23, When the absolute value of the difference When the value in the interval is not greater than the relevant threshold value of the feature variable, the corresponding established stable flight path is taken as the temporary stable flight path of the new temporary flight path, the value of the relevant frame number is incremented by 1, and the absolute value of the difference is normalized to obtain the normalized difference value.

[0060] The normalization method is as follows:

[0061]

[0062] It should be noted that, firstly, when the absolute value of the difference... If the value in the value exceeds the threshold value of the feature variable, a new temporary landing is re-established in the free point track of the next frame for judgment, until a new temporary landing meets the track start condition. That is, steps S21-S23 are repeated until a temporary landing meets the track start condition.

[0063] Secondly, if multiple stable flight conditions exist... The two stable flight paths with the highest overall correlation are selected as the relevant stable flight paths for this temporary flight path, and the number of relevant frames is increased by 1 for each.

[0064] S24. Then calculate the overall relevance.

[0065]

[0066] Where λ1 and λ2 are both no greater than 1, and are the differences between eigenvalues ​​respectively. and The proportion of λ1 and λ2 in the overall relevance. In some specific embodiments, the values ​​of λ1 and λ2 are generally set to 0.3.

[0067] S3: Determine whether the temporary flight can normally initiate a visible stable track based on the correlation. If yes, proceed to step S6; otherwise, proceed to step S4.

[0068] Please combine Figure 4 Step S3 includes the following steps S31-S32.

[0069] S31. When a temporary flight reaches the trajectory start condition, the sum of the relevant frames of all stable flights related to the temporary flight and the sum of the comprehensive correlation are calculated.

[0070] In some embodiments of the present invention, when a temporary flight meets the track start condition, the 6 / 8 logic start method based on sliding window is selected. Therefore, the track start is set to require M=8 frames. The sum of the number of relevant frames of all stable flight related to the temporary flight and the sum of the comprehensive correlation are counted within these 8 frames.

[0071] S32. Determine whether the sum of the number of stable flight-related frames reaches the preset frame value and whether the sum of the correlation reaches the preset correlation value. If so, convert it into an invisible flight path for management. Otherwise, convert this temporary flight path into a visible normal stable flight path for tracking and management.

[0072] In the corresponding embodiment, if the sum of the number of stable flight-related frames reaches a preset frame value and the sum of the correlation reaches a preset correlation value, then it is initially considered that the temporary flight path may come from multiple false flight paths of a large target. It is converted into an invisible flight path for management, and the ViewFlag field is set to 0 when the representative sends its information to the subsequent display interface. At the same time, the data information of the invisible flight path is not included in the analysis of functions such as traffic flow statistics and average vehicle speed calculation. If the determination result is that the temporary flight path is not strongly correlated with any stable flight path, then the ViewFlag field is set to 1, and the temporary flight path is converted into an explicit normal stable flight path for tracking and management.

[0073] It should be noted that the preset frame value and preset correlation value are related to the characteristics of the original spot. When the front-end processing algorithm and radar front-end parameters change, the characteristics of the original spot change, and the preset frame value and preset correlation value are increased or decreased accordingly.

[0074] S4. Convert the temporary flight path into a hidden track and continue to determine the correlation between the hidden track and the determined stable flight path over the subsequent preset number of frames and weeks.

[0075] Specifically, after a preset number of frames, it is determined whether the feature difference between the hidden track and a certain stable track has an increasing trend, or whether the number of unrelated frames is greater than half of the preset number of frames.

[0076] S5. Determine whether the hidden track can be converted into a visible stable track based on the degree of relevance. If yes, proceed to step S6; otherwise, proceed to step S7.

[0077] Specifically, when the difference between a hidden track and a stable track shows an increasing trend, or when the number of unrelated frames exceeds half of a preset frame period, the track is determined to originate from a real target, the ViewFlag field is set to 1, and the hidden track is converted into a visible stable track for management. Otherwise, if the two are determined to be correlated and have a high probability of originating from the same target, the hidden track is considered to be multiple false tracks from a single target and is deleted.

[0078] S6. Convert the temporary or hidden track into a visible stable track.

[0079] S7. Delete hidden tracks.

[0080] Please combine Figure 5 and Figure 6 , Figure 5 and Figure 6 The images show a comparison of the results of normal track processing and track processing after incorporating this method. The comparison shows that adding this method to the track processing algorithm can suppress false tracks from the same target.

[0081] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Various modifications and variations can be made to the invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the scope of the claims of the invention.

Claims

1. A method for suppressing false tracks in a traffic environment, characterized in that, Including the following steps: S1. Determine the threshold values ​​for the characteristic variables of false and true flight paths; S2. Determine the correlation between the temporary flight and the established stable flight during the period from the establishment of the temporary flight to the satisfaction of the flight path start conditions based on the relevant thresholds of the characteristic variables. S3. Determine whether the temporary flight can normally start as a visible stable track based on the correlation; if yes, proceed to step S6; otherwise, proceed to step S4. S4. Convert the temporary flight path into a hidden track, and continue to determine the correlation between the hidden track and the determined stable flight path over the subsequent preset number of frames and weeks. S5. Determine whether the hidden track can be converted into a visible stable track based on the degree of correlation; if yes, proceed to step S6; otherwise, proceed to step S7. S6. Convert the temporary or hidden track into a visible and stable track; S7. Delete hidden tracks; Step S1 includes: S11. Based on the current prior track data, calculate the feature data of each of the multiple tracks from the same target within M frames. The set of absolute values ​​of differences Where: X and Y are the x and y coordinates of the target in Cartesian coordinates, respectively, and V is the radial velocity. , and These are the absolute values ​​of the differences in X, Y, and V between multiple tracks of target t in the k-th frame; S12. After several frames, arrange the absolute values ​​of the differences of all the feature data that have appeared in ascending order, count the number of occurrences of each value, build a distribution histogram, calculate the probability of each value, and select the absolute values ​​of the differences of the feature data corresponding to the cumulative distribution function value of 90% as the selected values ​​of the thresholds for each feature variable, thus obtaining the threshold set. This set of threshold values ​​is used as the correlation threshold values ​​between the feature variables of false and true tracks.

2. The method for suppressing false tracks in a traffic environment according to claim 1, characterized in that, Step S2 includes: S21. Establishing a new temporary navigation from free points. ; S22, Traverse all established stable navigation routes Calculate the absolute value of the difference between it and the new temporary flight characteristic variable. ; S23, When the absolute value of the difference When the value in the interval is not greater than the relevant threshold value of the feature variable, the corresponding established stable flight path is taken as the temporary stable flight path of the new temporary flight path, the value of the relevant frame number is incremented by 1, and the absolute value of the difference is normalized to obtain the normalized difference value. ; S24. Calculate the overall relevance. The formula is as follows: In the formula, and All are no greater than 1, representing the differences between eigenvalues. and The proportion of overall relevance.

3. The method for suppressing false tracks in a traffic environment according to claim 2, characterized in that, In step S23, when the absolute value of the difference... If the value in the value is greater than the threshold value of the feature variable, a new temporary flight is re-established in the free point track of the next frame for judgment, until a new temporary flight meets the track start condition.

4. The method for suppressing false tracks in a traffic environment according to claim 2, characterized in that, In step S23, the normalization formula is as follows: 。 5. The method for suppressing false tracks in a traffic environment according to claim 2, characterized in that, Step S3 includes: S31. When a temporary flight meets the trajectory start condition, calculate the sum of the relevant frames of all stable flights related to the temporary flight and the sum of the comprehensive correlation. S32. Determine whether the sum of the number of stable flight-related frames reaches the preset frame value and whether the sum of the correlation reaches the preset correlation value. If yes, convert it into an invisible flight path for management; otherwise, convert this temporary flight path into an explicit normal stable flight path for tracking and management.

6. The method for suppressing false tracks in a traffic environment according to claim 5, characterized in that, The preset frame value and preset correlation value are related to the characteristics of the original spot. When the front-end processing algorithm and radar front-end parameters change, the characteristics of the original spot change, and the preset frame value and preset correlation value are increased or decreased accordingly.

7. The method for suppressing false tracks in a traffic environment according to claim 2, characterized in that, In step S23: If multiple stable navigation conditions exist... The two stable flight paths with the highest overall correlation are selected as the relevant stable flight paths for this temporary flight path, and the number of relevant frames is incremented by 1 for each.

8. The method for suppressing false tracks in a traffic environment according to claim 1, characterized in that, Step S4 includes: After a subsequent preset number of frames, determine whether the feature difference between the stealth track and a certain stable track has an increasing trend, or whether the number of unrelated frames is greater than half of the preset number of frames.

9. The method for suppressing false tracks in a traffic environment according to claim 8, characterized in that, In step S5, when the difference between the stealth track and a certain stable track shows an increasing trend, or the number of uncorrelated frames exceeds half of the preset frame period, it is determined that the stealth track originates from a true track. The objective is to transform the hidden flight path into a visible, stable flight path for management.

Images