A radar data processing method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a radar data processing method and device, electronic equipment and storage medium, the method comprises: obtaining radar data to be processed, the radar data to be processed includes radar data detected by multiple radars on at least one object in a preset period; according to the radar data to be processed, determine the ghost object in the preset time period, and delete the radar data belonging to the ghost object in the radar data to be processed. Therefore, embodiments of the present application can solve the problem that the removal accuracy of ghost data in the radar data of multiple radars is low in the prior art, thereby making the radar trajectory accuracy displayed low.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a radar data processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] Currently, radar equipment is increasingly being used in industries such as smart intersections and autonomous driving. Based on the reflection characteristics of waves, radar equipment senses the speed and size of dynamic objects on the road within a certain range, thereby reconstructing the continuous traffic flow state of the sensing area. Through a single processing algorithm, real-time state data of the target objects is generated from the raw scanned sector data (including judgments of angle and position).

[0003] Because the original radar data is affected by different equipment angles, lighting, weather, and interference from multiple factors such as overlapping, obstruction, shape features, and height of the target, the actual detection effect presented after processing the original vehicle trajectory data based on physical detection differs greatly from the real situation. This can result in significant vehicle shaking, vehicle body overlap, and trajectory loss, among other anomalies.

[0004] Currently, to improve radar data coverage, it is common practice to deploy multiple radar devices at different locations within the same detection area. This can significantly improve the completeness of the radar trajectory of the same target object, but it can also lead to repeated detection of the same target object by multiple radar devices, causing ghosting of radar trajectories. Furthermore, because the ID of the same target object differs depending on which radar device detects it, it is difficult to accurately correlate the radar trajectory data of the same target object detected by different radar devices.

[0005] Therefore, in the prior art, when multiple radar devices detect the same target object, the radar trajectory cannot be accurately judged when ghosting occurs, and the multiple radar trajectory data of the same target object cannot be accurately correlated, resulting in low accuracy of the displayed radar trajectory. Summary of the Invention

[0006] The purpose of this invention is to provide a radar data processing method, apparatus, electronic device, and storage medium to solve the problem in the prior art where the accuracy of removing ghost data in radar data from multiple radars is low, resulting in low accuracy of the displayed radar trajectory.

[0007] In a first aspect of the present invention, a radar data processing method is provided, the method comprising:

[0008] Acquire radar data to be processed, which includes radar data from multiple radars detecting at least one object within a preset segment;

[0009] Based on the radar data to be processed, identify the ghosting objects within the preset time period, and delete the radar data belonging to the ghosting objects from the radar data to be processed.

[0010] In a second aspect of the present invention, a radar data processing apparatus is provided, the apparatus comprising:

[0011] The data acquisition module is used to acquire radar data to be processed, which includes radar data of multiple radars detecting at least one object within a preset segment.

[0012] The ghosting processing module is used to determine the ghosting objects within the preset time period based on the radar data to be processed, and to delete the radar data belonging to the ghosting objects from the radar data to be processed.

[0013] In a third aspect of the present invention, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0014] Memory, used to store computer programs;

[0015] The processor, when executing a program stored in memory, implements the radar data processing method described in any of the above embodiments.

[0016] In a fourth aspect of the invention, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform any of the radar data processing methods described above.

[0017] In a fifth aspect of the invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform any of the radar data processing methods described above.

[0018] The radar data processing method provided in this invention can acquire radar data from multiple radars detecting at least one object within a preset time period. Based on this data, it identifies ghosting objects within the preset time period and deletes radar data belonging to these ghosting objects. Therefore, in this embodiment of the invention, parsing radar data segments to remove ghosting data from those segments yields more information than real-time data parsing, effectively improving the reliability of ghosting removal and thus enhancing the accuracy of the displayed radar trajectory. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0020] Figure 1 This is a flowchart of the radar data processing method according to an embodiment of the present invention;

[0021] Figure 2 This is a flowchart illustrating a specific implementation of the radar data processing method according to an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of set W in an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram illustrating the real-time data iteration principle in an embodiment of the present invention;

[0024] Figure 5 This is a schematic diagram of long-term ghosting deletion in an embodiment of the present invention;

[0025] Figure 6 This is a schematic diagram of the original radar trajectory of an intersection in an embodiment of the present invention;

[0026] Figure 7 This is a path-time diagram of the original radar data of a vehicle in a first time segment in an embodiment of the present invention;

[0027] Figure 8 This is a path-time diagram of the radar data after cleaning for the first time segment in this embodiment of the invention;

[0028] Figure 9 This is a path-time diagram of the original radar data of a vehicle in a second time segment in an embodiment of the present invention;

[0029] Figure 10 This is a path-time diagram of radar data cleaned within the second time segment in this embodiment of the invention;

[0030] Figure 11 This is a path-time diagram of the original radar data of multiple vehicles in a third time segment in an embodiment of the present invention;

[0031] Figure 12 This is a path-time diagram of radar data cleaned within the third time segment in this embodiment of the invention;

[0032] Figure 13 This is a schematic diagram illustrating vehicle identification errors in the queuing area according to an embodiment of the present invention;

[0033] Figure 14 This is a schematic diagram of the original following distance limit result in the embodiment of the present invention;

[0034] Figure 15 This is a schematic diagram of the queuing area following distance limitation results in an embodiment of the present invention;

[0035] Figure 16 This is a structural block diagram of the radar data processing device provided in the embodiments of the present invention;

[0036] Figure 17 This is a structural block diagram of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0039] In various embodiments of the present invention, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0040] Figure 1 This invention provides a radar data processing method that can be applied to electronic devices, such as computers and servers. Figure 1 As shown, the radar data acquisition method may include the following steps 101 to 102:

[0041] Step 101: Acquire radar data to be processed.

[0042] The radar data to be processed includes radar data from multiple radars detecting at least one object within a preset segment. These multiple radars are, for example, radars installed at intersections, capable of detecting vehicles at the intersections.

[0043] In addition, different radars may identify the same object differently; the radar data mentioned above may include the object's type, location (e.g., latitude and longitude information), speed, and direction of movement (e.g., azimuth information).

[0044] Step 102: Based on the radar data to be processed, determine the ghosting objects within the preset time period, and delete the radar data belonging to the ghosting objects from the radar data to be processed.

[0045] In this invention, the duration of the preset time period is longer than the duration of a single radar data frame. Therefore, in the embodiments of the invention, ghosting is removed based on a radar data segment of a certain duration. Compared to real-time data analysis, ghosting removal is based on a larger amount of data, which allows for more accurate ghosting removal.

[0046] As described above, the radar data processing method provided in this embodiment of the invention can acquire radar data from multiple radars detecting at least one object within a preset time period. Based on this data, it can determine the ghosting object within the preset time period and delete the radar data belonging to the ghosting object. Therefore, in this embodiment of the invention, parsing radar data segments to remove ghosting data from those segments yields more information than real-time data parsing, effectively improving the reliability of ghosting removal and thus enhancing the accuracy of the displayed radar trajectory.

[0047] Optionally, the step of determining the ghosting objects within the preset time period based on the radar data to be processed, and deleting the radar data belonging to the ghosting objects from the radar data to be processed, includes:

[0048] Based on the radar data to be processed, identify the first ghost object whose trajectory duration is greater than the first preset value within the preset time period, and delete the radar data belonging to the first ghost object from the radar data to be processed.

[0049] Based on the radar data to be processed, a second ghost object with a trajectory duration less than or equal to the first preset value within the preset time period is identified, and radar data belonging to the second ghost object in the radar data to be processed is deleted.

[0050] The trajectory duration of an object is the total duration of the time frames in which the object exists continuously. For example, if an object exists in the 1st to 40th time frames and the duration of a time frame is 0.1 seconds, then the trajectory duration of the object is 4 seconds.

[0051] In addition, the first ghost object whose trajectory duration is greater than the first preset value within the aforementioned preset time period can be called a long-duration ghost object; the second ghost object whose trajectory duration is less than or equal to the first preset value within the aforementioned preset time period can be called a short-duration ghost object.

[0052] As can be seen from the above, in the embodiments of the present invention, long-term ghosting and short-term ghosting in the radar data to be processed can be identified, and long-term ghosting and short-term ghosting can be deleted respectively, thereby improving the accuracy of ghosting deletion.

[0053] Optionally, the step of determining, based on the radar data to be processed, a first ghosting object whose trajectory duration is greater than a first preset value within the preset time period, and deleting radar data belonging to the first ghosting object from the radar data to be processed, includes:

[0054] When i takes the integer value from 1 to M, the following process is executed:

[0055] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0056] Obtain the fourth identifier that meets the third preset condition in the radar data of the i-th period, and store the fourth identifier in the first set, wherein the third preset condition includes that it exists in the (X-1)-th frame, the X-th frame, and the Y+1-th frame of the radar data of the i-th period;

[0057] Obtain the fifth identifier that meets the fourth preset condition in the radar data of the i-th period, and store the fifth identifier in the second set, wherein the fourth preset condition includes that it exists in the (X-1)-th frame, and does not exist in the X-th frame and the Y+1-th frame;

[0058] Obtain a first target identifier from the first set and a second target identifier from the second set, wherein the distance between the objects represented by the first target identifier and the second target identifier in the (X-1)th frame is less than a first preset value;

[0059] Delete the radar data corresponding to the second target identifier in the radar data from frame X-1 to frame Y+1 of the i-th cycle.

[0060] Among them, the objects represented by the fourth identifier that meet the third preset condition and the objects represented by the fifth identifier that meet the fourth preset condition are the first ghost objects whose trajectory duration is greater than the first preset value within the preset time period, which are long-duration ghost objects.

[0061] For example, if the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute one cycle), and X = 10 and Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0062] For radar data in one cycle, the following two sets of IDs are obtained based on certain conditions:

[0063] Set ①: In frame T9, T 10 Frame and T 50 ID present in the frame;

[0064] Set ②: Exists in frame T9, in T 10 Frame and T 50 ID not found in the frame;

[0065] Find objects A1 and A2 in set ① and set ② respectively. Objects A1 and A2 satisfy that the distance between their latitude and longitude coordinates at frame T9 is less than d1, where d1 is the first preset value mentioned above (for example, if the radar detects a vehicle, then d1 is the minimum distance between the coordinates of the two vehicles in normal following conditions).

[0066] Then, delete object A2 in this cycle from frame T9 to T... 50 Record all records between frames, and record the ID of object A2 as 'Original ID' and the ID of object A1 as 'Change ID' in table S1.

[0067] Among them, compared with the IDs in set ① and set ②, the IDs in set ① are in T. 50 If the image still exists in the frame, then the IDs in these two sets are relative. Set ① represents the trajectory ID that disappeared after the ghosting occurred, and set ② represents the trajectory ID that disappeared first when the ghosting occurred. Therefore, object A2 is the object that disappeared first when the ghosting occurred relative to object A1, so the ID of object A1 should be recorded as 'Change ID', and the ID of object A2 should be recorded as 'Original ID'.

[0068] Furthermore, based on the latitude and longitude coordinates of objects A1 and A2, the formula for calculating the distance between them is as follows:

[0069]

[0070] Where R = 6371 km, Represents the latitude and longitude coordinates of object A1. Represents the latitude and longitude coordinates of object A2.

[0071] The above steps complete the removal of long-duration ghosting from the aforementioned 10 seconds of radar data. Specifically, the steps removed the ghosting from "object A2 within one cycle of radar data, from frame T9 to T..." 50"All records between frames", where a frame is 0.1 seconds long, then 4 seconds of ghosting will be removed in each cycle. In this way, each cycle is formed, and 4 seconds of ghosting is continuously removed in that cycle until the 4 seconds of ghosting in the last cycle are removed, ending the long-term ghosting removal work. At this point, even if there is ghosting in the obtained 10 seconds of radar data, its duration will be less than 4 seconds.

[0072] Optionally, the radar data may include object identifiers;

[0073] The step of determining, based on the radar data to be processed, a second ghost object whose trajectory duration within the preset time period is less than or equal to the first preset value, and deleting radar data belonging to the second ghost object from the radar data to be processed, includes:

[0074] When i takes the integer value from 1 to M, the following process is executed:

[0075] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0076] Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1;

[0077] Obtain a second identifier from the first identifier that meets the first preset condition and the second preset condition, wherein the first preset condition includes that it exists in the (X+1)th frame of the i-th period radar data and does not exist in the Y-th and Y+1th frames of the i-th period radar data, and the second preset condition includes that it does not exist in the (X-1)th frame of the i-th period radar data, where Y = N-1;

[0078] Delete the radar data corresponding to the second identifier in the radar data of the i-th cycle.

[0079] Among them, the objects represented by the second identifier that meet the first preset condition and the second preset condition are the second ghost objects whose trajectory duration within the preset time period is less than or equal to the first preset value, which are short-term ghost objects.

[0080] For example, if the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute one cycle), and X = 10 and Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0081] For T 10Filter all IDs in the frame and select IDs that meet both of the following conditions: ① ID in T 11 Frame exists, in T 49 Frame and T 50 ① The ID does not exist in frame T9; ② The ID does not exist in frame T9; where T k This represents the radar data in the k-th frame of a period.

[0082] IDs that meet the above two conditions represent brief ghosting events in the radar trajectory; that is, these IDs appear briefly in T... 11 It exists in, but T9, T 49 T 50 If it does not exist, it means that the maximum duration of the trajectory for this type of ID is also less than 4 seconds, which is a short-duration trajectory. Therefore, this type of ID needs to be deleted from the radar data of this period.

[0083] Optionally, the radar data may include object identifiers;

[0084] The step of determining a second ghosting object whose trajectory duration within the preset time period is less than or equal to the first preset value based on the radar data to be processed, and deleting radar data belonging to the second ghosting object from the radar data to be processed, further includes:

[0085] Obtain the sixth identifier that meets the fifth preset condition in the radar data of the i-th cycle, and store the sixth identifier in the third set. The fifth preset condition includes that it exists in both the (X-1)th frame and the Xth frame in the radar data of the i-th cycle, and does not exist in both the Yth frame and the Y+1th frame in the radar data of the i-th cycle, where X is an integer greater than 1 and Y = N-1.

[0086] Obtain the seventh identifier that meets the sixth preset condition in the radar data of the i-th period, and store the seventh identifier in the fourth set, wherein the sixth preset condition includes not existing in the (X-1)-th frame and existing in the X-th frame;

[0087] A third target identifier is obtained from the third set, and a fourth target identifier is obtained from the fourth set, wherein the distance between the objects represented by the third target identifier and the fourth target identifier in the (X-1)th frame is less than a first preset value, and the distance between the objects represented by the third target identifier and the fourth target identifier in the target frame is less than the first preset value, and the target frame is the last frame in which the third target identifier appears in the radar data of the i-th cycle;

[0088] Delete the radar data corresponding to the third target identifier in the target frame from the (X-1)th frame to the target frame in the i-th period radar data.

[0089] Among them, the objects represented by the sixth identifier that meet the fifth preset condition, and the objects represented by the seventh identifier that meet the sixth preset condition, are the second ghost objects whose trajectory duration within the preset time period is less than or equal to the first preset value, which are short-term ghost objects.

[0090] Additionally, it should be noted that if the object represented by the fourth target identifier does not exist in the target frame, then by default, the latitude and longitude coordinates of the object represented by the fourth target identifier in the target frame are taken as the latitude and longitude coordinates of the origin. Therefore, the distance between the third target identifier and the object represented by the fourth target identifier in the target frame is less than the first preset value. Similarly, if the object represented by the fourth target identifier does not exist in frame X-1, then by default, the latitude and longitude coordinates of the object represented by the fourth target identifier in frame X-1 are taken as the latitude and longitude coordinates of the origin. Therefore, the distance between the third target identifier and the object represented by the fourth target identifier in frame X-1 is less than the first preset value.

[0091] For example, if the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute one cycle), and X = 10 and Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0092] For radar data in one cycle, the following two sets of IDs are obtained based on certain conditions:

[0093] Set ③: In frame T9, T 10 It exists in all frames, in T 49 T 50 IDs that do not exist in any frame;

[0094] Set ④: Not present in frame T9, but present in frame T... 10 Existing in the frame ID ;

[0095] In sets ③ and ④, find objects A3 and A4 respectively, where objects A3 and A4 satisfy the condition that the distance between their latitude and longitude coordinates at frame T9 is less than d1, and at frame T... A The distance between the latitude and longitude coordinates of the frame is less than d1, where d1 is the first preset value (for example, if the radar detects a vehicle, then d1 is the minimum distance between the coordinates of the two vehicles in normal following conditions), T A The last frame in which object A3 appears;

[0096] Then, remove object A3 from frame T9 to T... A Record all records between frames, and record the ID of object A4 as 'Original ID' and the ID of object A3 as 'Change ID' in table S1.

[0097] In this context, set ③ represents the trajectory ID that disappears first when ghosting occurs, and set ④ represents the trajectory ID that disappears later when ghosting occurs. Therefore, object A3 is the object that disappears first when ghosting occurs relative to object A4, so the ID of object A3 should be recorded as 'Change ID', and the ID of object A4 should be recorded as 'Original ID'.

[0098] Optionally, the radar data may include object identifiers;

[0099] Before determining the first ghosting object with a trajectory duration greater than a first preset value within the preset time period based on the radar data to be processed, and deleting the radar data belonging to the first ghosting object from the radar data to be processed, the method further includes:

[0100] When i takes the integer value from 1 to M, the following process is executed:

[0101] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0102] Obtain the eighth identifier in the i-th period radar data that meets the seventh preset condition, wherein the seventh preset condition includes that it exists in both the Y-1 and Y+1 frames in the i-th period radar data, and does not exist in the Y-th frame in the i-th period radar data, Y = N-1;

[0103] Based on the radar data of the object represented by the eighth identifier in the (Y-1)th frame and the radar data in the (Y+1)th frame, determine the radar data of the object represented by the eighth identifier in the Yth frame.

[0104] For example, if the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute one cycle), and X = 10 and Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0105] For T 48 All records in the frame are iterated and checked, and the ID of the record is found in T. 49 Frames and T 50 Does the frame exist? If the record ID is in T 48 Frames and T 50 It appears in all frames, but in T 49 In the case of a missing frame, the record ID is used to locate the missing frame in T. 48 Frames and T 50 Radar data in the frame, determining the record ID in T49 Radar data in the frame, i.e., the record ID in T 49 Data completion is performed within the frame.

[0106] As can be seen from the above, in the embodiments of the present invention, data completion of the radar data to be processed before ghost removal can further improve the accuracy of subsequent ghost identification, thereby improving the accuracy of ghost removal and thus improving the accuracy of the displayed radar trajectory.

[0107] Optionally, radar data includes object type, object size, object number information, velocity, and latitude and longitude coordinates;

[0108] The step of determining the radar data of the object represented by the eighth identifier in the Y-1 frame based on the radar data of the object represented by the eighth identifier in the Y-1 frame and the radar data in the Y+1 frame includes:

[0109] The object type, object size, and number information of the object represented by the eighth identifier in the (Y-1)th frame are respectively determined as the object type, object size, and number information of the object represented by the eighth identifier in the Yth frame;

[0110] The speed of the object represented by the eighth identifier in the (Y-1)th frame and the speed of the object represented by the eighth identifier in the (Y+1)th frame are taken as the speed of the object represented by the eighth identifier in the Yth frame.

[0111] The latitude and longitude coordinates of the object represented by the eighth identifier in the (Y-1)th frame and the average latitude and longitude coordinates of the object represented by the eighth identifier in the (Y+1)th frame are determined as the latitude and longitude coordinates of the object represented by the eighth identifier in the Yth frame.

[0112] For example, in the example above, based on a certain record ID in T 48 Frames and T 50 Radar data in the frame, determining the record ID in T 49 When dealing with radar data in a frame, for example, if the record ID is k, then k can be placed in T. 49 The frame's object type, object size, and detection radar number information are related to k in T. 48 The information in the frames remains consistent, and this k is in T 49 frame rate latitude and longitude coordinates in, Indicates k in T 48 Frame rate, Indicates that k is in Frame rate, Indicates k in T48 The latitude and longitude coordinates of the frame Indicates k in T 50 The latitude and longitude coordinates of the frame.

[0113] Optionally, radar data may also include object identification;

[0114] The step of adjusting the latitude and longitude coordinates of the object in the radar data to be processed based on the positional relationship between the detected object of the radar and the boundary line of a pre-determined target area includes:

[0115] When i takes the integer value from 1 to M, the following process is executed:

[0116] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0117] Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1;

[0118] Obtain the identifiers whose latitude and longitude coordinates are within the boundary line of the target area from the first identifier, and use them as the ninth identifier;

[0119] The latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are adjusted to the first target coordinates, wherein the straight line containing the first point and the second point is perpendicular to the center line of the boundary line of the target area, and the first point is located on the center line. The first point is the point represented by the first target coordinates, and the second point is the point represented by the original latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame.

[0120] For example, the target area mentioned above could be a vehicle queuing area, the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute a cycle), X = 10, Y = 49. Then, for these 10 seconds of radar data, each frame forward forms a cycle of radar data, and the following process is performed on the radar data of this cycle:

[0121] In T 10 Within a frame, given the curve formulas L1 and L2 of the two lane lines of lane D1, the lane centerline L3, and the coordinates of the beginning and end points, vehicles with latitude and longitude coordinates located between L1 and L2 can have their lane attributes assigned a value of D1. For vehicles with lane attributes that are not empty, their latitude and longitude coordinates are modified. For example, if an ID with a lane attribute is p, then the modified latitude and longitude coordinates of that ID represent the original latitude and longitude coordinates. The position of the foot of the perpendicular from the lane centerline L3 of the lane it belongs to. For example, assuming the lane centerline is the line y = k1x + b1, the perpendicular line is... The position of the foot of the perpendicular is y = k1x + b1 and The intersection point.

[0122] As described above, the ghosting removal process largely eliminates object ghosting caused by repeated detection by multiple radar devices, but the object jitter problem remains unresolved. For example, when a vehicle enters a queuing area, because the lane lines are solid, the vehicle cannot change lanes in the queuing area without violating traffic regulations. However, vehicle jitter can cause the vehicle to jump across lanes in the radar trajectory along the solid line segment. Embodiments of this invention can constrain the lateral position of the lane, adjusting the latitude and longitude coordinates of the ID located within the lane lines to the center line of the lane lines, thereby solving the vehicle jitter problem to some extent.

[0123] Optionally, after adjusting the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame to the first target coordinates, the method further includes:

[0124] Based on the slope of the centerline, determine the first driving direction angle of the object represented by the ninth identifier;

[0125] Based on the latitude and longitude coordinates of the object represented by the ninth identifier in the (X+2)th and (X+3)th frames of the radar data in the i-th cycle, determine the second driving direction angle of the object represented by the ninth identifier;

[0126] When the signs of the first driving direction angle and the second driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the ninth identifier in the X+1th frame.

[0127] The determination of the first driving direction angle of the object represented by the ninth identifier based on the slope of the centerline includes:

[0128] According to the first preset formula Determine the first driving direction angle θ1, where α represents the slope of the centerline.

[0129] The above-mentioned determination of the second driving direction angle of the object represented by the ninth identifier based on the latitude and longitude coordinates of the object in the (X+2)th and (X+3)th frames of the radar data in the i-th period includes:

[0130] According to the second preset formula Determine the second driving direction angle θ2, where, These represent the latitude and longitude coordinates of the object represented by the lth ninth identifier in frames X+2 and X+3, respectively.

[0131] Optionally, the method further includes:

[0132] Obtain the identifier in the first identifier whose latitude and longitude coordinates are outside the boundary line of the target area, and use it as the tenth identifier;

[0133] Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the first and Xth frames of the i-th cycle radar data, determine the third driving direction angle of the object represented by the tenth identifier;

[0134] Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the (X+2)th and (X+3)th frames of the radar data in the i-th cycle, determine the fourth driving direction angle of the object represented by the tenth identifier;

[0135] When the signs of the third driving direction angle and the fourth driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the tenth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the tenth identifier in the X+1th frame.

[0136] The calculation methods for the third and fourth driving direction angles are the same as those for the second driving direction angle, and will not be repeated here.

[0137] For example, if the target area is a vehicle queuing area, the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute a cycle), and X = 10, Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0138] For T 10 The system iterates through vehicle IDs with lane number attributes (i.e., latitude and longitude coordinates located within lane lines L1 and L2) within a frame. For any ID, it obtains the first driving direction angle θ1 based on the slope α of the centerline of the lane to which the queuing area belongs. Then, it uses this ID to determine the driving direction angle in time T. 12 Frame and T 13 The latitude and longitude coordinates of the frame are used to obtain the ID in T. 12 Frame and T 13 The second driving direction angle θ2 within the frame, if θ1 and θ2 are in opposite directions, then let this ID be in T. 11 The latitude and longitude coordinates of the frame and T 10 The frames remain consistent.

[0139] For T 10Iterate through all vehicle IDs in the frame that have no lane attribute (i.e., latitude and longitude coordinates located outside lane lines L1 and L2). For any ID, determine its position in T1 and T2. 10 Given the latitude and longitude coordinates of the frame, obtain the third driving direction angle θ3 of the ID in the previous second, and then determine the T-axis position based on the ID. 12 Frame and T 13 The latitude and longitude coordinates of the frame are used to obtain the ID in T. 12 Frame and T 13 If the fourth driving direction angle θ4 within the frame is opposite in direction to θ3 (i.e., θ3 and θ4 have opposite signs), then let this ID be in T. 11 The latitude and longitude coordinates of the frame and T 10 The frames remain consistent.

[0140] As can be seen from the above, in the embodiments of the present invention, the radar data to be processed can also be processed according to the no-back-away rule, thereby improving the accuracy of the radar data and thus improving the accuracy of displaying the radar trajectory.

[0141] Optionally, after determining the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame as the latitude and longitude coordinates of the object represented by the ninth identifier in the (X+1)th frame, the method further includes:

[0142] From the ninth identifier, select the identifier that does not exist in the (X-1)th frame of the i-th period radar data, and use it as the eleventh identifier;

[0143] Determine the acceleration of the object represented by the eleventh identifier in the (X-1)th frame;

[0144] Based on the acceleration and velocity of the object represented by the eleventh identifier in the (X-1)th frame, determine the target displacement of the object represented by the eleventh identifier, wherein the target displacement is the displacement of the object represented by the eleventh identifier from its position in the (X-1)th frame within one frame time.

[0145] Based on the target displacement, determine the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame, and use them as the second target coordinates;

[0146] If the distance between the second target coordinates and the first reference coordinates is greater than or equal to a first preset value, and the distance between the second target coordinates and the second reference coordinates is greater than or equal to the first preset value, the eleventh identifier is added to the Xth frame, and the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame are determined to be the second target coordinates.

[0147] Wherein, the first reference coordinate and the second reference coordinate are the latitude and longitude coordinates of two objects located before and after the second target coordinate in the Xth frame, respectively.

[0148] When the radar data includes object type, object size, and object number information, the object represented by the eleventh identifier has the same object type, object size, and object number information in the Xth frame as in the (X-1)th frame.

[0149] For example, if the target area is a vehicle queuing area, the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute a cycle), and X = 10, Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0150] Perform a traversal check on vehicle IDs with lane attributes (i.e., latitude and longitude coordinates located within lane lines L1 and L2) in frame T9. If the ID is in T... 10 If it does not exist within the frame, then in T 10 Add a new radar data entry with this ID within the frame, and set this ID in T. 10 The lane attributes, vehicle type, length, width, and detection radar ID information of the frame are consistent with the information of that ID in frame T9; and the acceleration 'a' of that ID in frame T9 is determined, thereby determining the acceleration 'a' and the speed of that ID in frame T9. Determine the displacement that occurred in this frame. Then, based on this ID in T 10 The first driving direction angle θ1 in the frame and its displacement Obtain the ID in T 10 Latitude and longitude coordinates of the frame and This type of newly added record has an added identification attribute to determine whether it is a missing vehicle.

[0151] If an ID with the aforementioned identification attributes is added, in T 10 If the distance between the vehicles in front and behind within a frame is less than d1 (i.e., the aforementioned first preset value), then it indicates that the aforementioned distance in T... 10 The location corresponding to the ID added to the frame does not contain a vehicle; therefore, the added vehicle is invalid and needs to be removed from T. 10 The radar data with that ID is added to the frame; if an ID with the above identification attributes is added, in T 10 If the distance between the vehicles in front and behind within a frame is greater than or equal to d1 (i.e., the aforementioned first preset value), then it indicates that the aforementioned distance between vehicles in T is greater than or equal to d1. 10If a vehicle exists at the location corresponding to the ID added to the frame, meaning the added vehicle is valid, then the aforementioned vehicle in T is retained. 10 Radar data for that ID is added to the frame.

[0152] In addition, displacement The calculation formula is as follows:

[0153] Longitude calculation formula:

[0154] Latitude calculation formula:

[0155] First driving direction angle:

[0156] α is the slope of the aforementioned center line.

[0157] Among them, the above This represents the target displacement of the object represented by the r-th eleventh identifier. This indicates the velocity of the object represented by the r-th eleventh identifier in the (X-1)-th frame. These represent the latitude and longitude coordinates of the object represented by the r-th eleventh identifier in the (X-1)th and Xth frames, respectively.

[0158] Optionally, determining the acceleration of the object represented by the eleventh identifier in the (X-1)th frame includes:

[0159] If the object represented by the eleventh identifier is obtained as the following object in the X-1 frame, the acceleration of the object represented by the eleventh identifier in the X-1 frame is determined based on the speed and latitude / longitude coordinates of the object represented by the eleventh identifier in the X-1 frame, the speed and latitude / longitude coordinates of the following object in the X-1 frame, and a predetermined following model.

[0160] If the following object is not obtained, the acceleration of the object represented by the eleventh identifier in the (X-1)th frame is determined as the preset acceleration.

[0161] As can be seen from the above, in the embodiments of the present invention, when the above-mentioned car-following object is obtained, the acceleration of the object represented by the eleventh identifier in the X-1 frame can be determined by the car-following model; when the car-following object is not obtained, the acceleration of the object represented by the eleventh identifier in the X-1 frame can be set to a preset acceleration.

[0162] For example, if the object mentioned above is a vehicle, the calculation formula for the car-following model is as follows:

[0163]

[0164] in: In addition, the relevant explanations of the parameters involved in the calculation formula of the car-following model are shown in Table 1.

[0165] Table 1 Car-following model parameters

[0166] Model parameters meaning The instantaneous acceleration a of the vehicle It is generally calculated dynamically every 0.1 seconds. Desired acceleration a′ (Custom settings) <![CDATA[Current vehicle speed v a > vehicle's current speed <![CDATA[Desired speed v0]]> Expected speed of vehicles on this road Speed ​​of the vehicle in front v Speed ​​of following vehicles Speed ​​difference △v between the front and rear vehicles Speed ​​difference between the following vehicle and the current vehicle Safe following distance T 1.5s Expected deceleration b (Custom settings) <![CDATA[The distance s between the front and rear vehicles α > The distance between the front and rear of the following vehicle <![CDATA[Stopping distance s0]]> 2m ε The default value is 4 β The default value is 2

[0167] The aforementioned safe following distance is the ratio of the longitudinal distance between two vehicles to the speed of the following vehicle. It generally reflects the safety of the automatic cruise control function. It can be seen as the time required for the following vehicle to collide with the vehicle in front if the vehicle in front brakes to a stop and the following vehicle does not slow down, for example, 1.5 seconds.

[0168] Furthermore, if a certain ID does not have a preceding vehicle in frame T9, then this vehicle is the first vehicle in the queue. The driving status of this vehicle will be determined according to the movement status of vehicles in the adjacent lanes. If the direction is red and prohibited, the acceleration of this ID in frame T9 will be 0. If the direction is green and the acceleration of this ID in frame T9 will be a fixed value a′, where a′ is the expected acceleration of the current road segment.

[0169] As can be seen from the above, in the embodiments of the present invention, after processing the radar data to be processed according to the no-back-off rule, the Xth frame of the radar data in each cycle can also be processed to complete the trajectory of the vanishing object, thereby improving the accuracy of the radar data and thus improving the accuracy of displaying the radar trajectory.

[0170] Optionally, after adding the eleventh identifier to the Xth frame and determining that the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame are the second target coordinates, the method further includes:

[0171] Obtain the sorted position of the object represented by the ninth identifier in the Xth frame;

[0172] When j takes the value 1 K, the following process is executed:

[0173] Calculate the distance between the j-th object and the (j-1)-th object in the position sorting, and if the distance is less than a first preset value, determine the latitude and longitude coordinates of the j-th object in the X-th frame as the third target coordinates, wherein the third target coordinates are located on the center line, the distance between the third target coordinates and the latitude and longitude coordinates of the (j-1)-th object in the X-th frame is equal to the first preset value, and K is the number of the ninth identifiers.

[0174] For example, if the target area is a vehicle queuing area, the duration of the radar data to be processed is 10 seconds (then M = 50), and N = 49 (i.e., every 50 consecutive frames of radar data constitute a cycle), and X = 10, Y = 49, then for these 10 seconds of radar data, each frame moved forward forms a cycle of radar data. The following process is then performed on the radar data of this cycle:

[0175] In T 10 For a vehicle ID in the queue area with lane attributes (latitude and longitude coordinates located within lane lines L1 and L2) in a frame, the first vehicle, the second vehicle, and so on up to the last vehicle in the queue are determined based on the lane driving direction angle θ1 of the queue area.

[0176] If θ1 = 0, then the vehicle with the largest longitude is the first vehicle, and the vehicle with the smallest longitude is the last vehicle.

[0177] If θ1 = π, then the vehicle with the smallest longitude is the first vehicle, and the vehicle with the largest longitude is the last vehicle.

[0178] If θ1∈[0,π), then the vehicle with the largest latitude is the first vehicle, and the vehicle with the smallest latitude is the last vehicle.

[0179] If θ1∈(π,2π), then the vehicle with the smallest latitude is the first vehicle, and the vehicle with the largest latitude is the last vehicle.

[0180] Then, for each vehicle in each lane, the distance between the vehicle and the vehicle in front is calculated in the order from the second vehicle to the last vehicle. If the distance is less than d1, the position of the vehicle is modified, that is, the latitude and longitude coordinates of the vehicle are changed to the position d1 behind the vehicle in front on the center line of the lane. During this process, the order of the vehicles remains unchanged.

[0181] For example, if the current vehicle is located in front of the vehicle that is nominally in front of it, such as... Figure 13 If the original following distance limit is applied at this time, then it becomes Figure 14 The current situation, with one car missing, does not reflect reality. However, after special optimizations to the following distance restrictions in the queuing area, car ② will not disappear; the current car ① will return to a position d1 behind the nominally preceding car ②, as shown below. Figure 15 As shown.

[0182] As can be seen from the above, in the implementation of the present invention, the distance limit of adjacent objects can also be applied to the radar data of the Xth frame in each period of the radar data to be processed, so as to further improve the accuracy of displaying the radar trajectory.

[0183] Furthermore, it is understood that the above embodiments can be combined with each other. Optionally, one combination method is to combine all the above embodiments together, and the specific implementation method can be as follows: Figure 2 As shown.

[0184] like Figure 2 As shown, the data stream is parsed into a dataset, and a time segment of duration t0 is selected from this dataset. Specifically, the radar data set from the current time and the past t0 seconds is named W. The nth frame in W is named T. n Generally, the 10th frame in W is selected, i.e., T. 10 The target frame for data cleaning and trajectory reconstruction is called the current processing frame. For example... Figure 3 As shown, the dataset W includes frames i to i+49, where frame i+49 is the current frame and frame i+9 is the current processing frame.

[0185] The output T is obtained after three steps: time frame completion, ghosting removal, and optimization of vehicle behavior in the queuing area. 10 Then, the current time is advanced by one frame, thus enabling the algorithm to iterate continuously. The overall iterative process is as follows: Figure 4 As shown. During this process, using... The speed of the record with ID i in the nth frame of dataset W is referred to as... This refers to the longitude and latitude of the record with ID i in the nth frame of dataset M.

[0186] Specifically, refer to Figure 2 The process of "time frame completion, ghosting removal, and optimization of vehicle behavior in the queuing area" is described below:

[0187] 1. Time frame completion:

[0188] For T in dataset W 48 All records in the frame are iterated and checked, and the ID of the record is found in T. 49 Frames and T 50 Does the frame exist? If the record ID is in T 48 Frames and T 50 It appears in all frames, but in T 49 If a frame is missing, time frame completion is performed. That is, the record ID in T... 49 The vehicle type, length, width, and detection radar number information of the frame are compared with the ID in T. 48 The information in the frames remains consistent; for example, the record with ID k in T... 49 frame rate latitude and longitude coordinates

[0189] 2. Ghosting removal

[0190] 2.1 After completing the frame interpolation, for T 10 Filter all IDs in the frame and select IDs that meet both of the following conditions: ① The ID is in T 11 Frame exists, in T49 Frame and T 50 ① The ID does not exist in frame T9. Delete all radar data represented by IDs that meet both conditions ① and ② from dataset M.

[0191] 2.2 Filter the entire dataset W according to the following conditions to obtain the following two sets of IDs:

[0192] Set ①: In frame T9, T 10 Frame and T 50 ID present in the frame;

[0193] Set ②: Exists in frame T9, in T 10 Frame and T 50 ID not found in the frame;

[0194] Find vehicles A1 and A2 in sets ① and ② respectively, where the distance between their latitude and longitude coordinates at frame T9 is less than d1, where d1 is the minimum distance between the coordinates of the two vehicles in normal following mode. Delete vehicle A2 from dataset W between frames T9 and T... 50 All radar data between frames are processed, and the ID of vehicle A2 is recorded as 'Original ID' and the ID of vehicle A1 is recorded as 'Change ID' in Table S1.

[0195] For example Figure 5 As shown, the first trajectory 501 represents the vehicle between T9 and T1. 50 The vehicle represented by the second trajectory 502 exists in all frames, and also exists in frame T9. 10 Frame and T 50 It does not exist in the frame, and in T 11 Frame to T 49 If the trajectories between frames are discontinuous, then the vehicle represented by the first trajectory 501 belongs to the vehicles in set ① above, and the vehicle represented by the second trajectory 502 belongs to the vehicles in set ② above. Therefore, if the distance between the latitude and longitude coordinates of these two vehicles at frame T9 is less than d1, then the section from T9 to T in the second trajectory 502 needs to be deleted. 50 The trajectory in the frame.

[0196] 2.3 Filter the entire dataset W according to the following conditions to obtain the following two sets of IDs:

[0197] Set ③: In frame T9, T 10 It exists in all frames, in T 49 T 50 IDs that do not exist in any frame;

[0198] Set ④: Not present in frame T9, but present in frame T... 10 The ID present in the frame.

[0199] Find vehicles A3 and A4 in sets ③ and ④ respectively, where vehicle A3 and vehicle A4 satisfy the condition that the distance between their latitude and longitude coordinates at frame T9 is less than d1. Here, d1 is the minimum distance between the coordinates of the two vehicles under normal following conditions. A The last frame in which vehicle A3 appears; remove vehicle A3 from frame T9 to T... A All radar data between frames are recorded, and the ID of vehicle A4 is recorded as 'Original ID' and the ID of vehicle A3 is recorded as 'Change ID' in Table S1.

[0200] Since A4 does not exist in frame T9, when calculating the distance between A3 and A4, the latitude and longitude coordinates of A4 in frame T9 are assumed to be the origin coordinates; similarly, when A4 is in frame T9... A When A4 is not present in the frame, it is in T. A The latitude and longitude coordinates in the frame are defaulted to the origin coordinates.

[0201] In addition, after completely deleting overlapping video segments, the trajectories of the two ends of the same vehicle that broke apart can be spliced ​​together based on driving behavior.

[0202] 3. Optimization of vehicle behavior in queuing areas

[0203] By removing ghosting, the vehicle ghosting caused by repeated detection by multiple radar devices has been largely eliminated. However, issues such as vehicle shaking and overlapping still remain in large quantities in the intersection queuing area. Therefore, the next step after ghosting removal is to optimize vehicle behavior in the queuing area.

[0204] Optimizing vehicle behavior in queuing areas requires prior collection of road features at the intersections where the algorithm is applied, including the following three items: ① number of lanes at the intersection, ② lane turning function, and ③ position and orientation of lane lines and center lines.

[0205] For example, the original radar trajectory point map of a certain intersection is as follows: Figure 6 As shown, the east-west approach trajectories at this intersection are clearly divided into multiple clusters according to lanes, providing excellent conditions for lane lateral position constraints. Therefore, after removing ghosting from the original radar data of this intersection, the radar data of this intersection can be processed according to the contents of sections 3.1 to 3.4 below.

[0206] 3.1 Lateral position constraints of the queuing area

[0207] In T 10 Within a frame, given the curve formulas L1 and L2 of the two lane lines of lane D1, the lane centerline L3, and the coordinates of the beginning and end points, vehicles with latitude and longitude coordinates located between L1 and L2 can have their lane attributes set to D1. For vehicles with non-empty lane attributes, their latitude and longitude coordinates are modified, specifically to the original trajectory points of that vehicle. The position of the foot of the perpendicular to the lane centerline L3 of the lane to which it belongs.

[0208] 3.2 The principle of prohibiting backward movement in the queuing area

[0209] The no-back-out rule is divided into global rules and queuing area rules. Within the entire intersection area, it applies to T. 10 Iterate through all vehicle IDs in the frame that have no lane attribute (i.e., latitude and longitude coordinates located outside L1 and L2). For any ID, based on that ID in frame T1 and T2... 10 The frame trajectory obtains the third driving direction angle θ3 of the ID in the previous second, and according to T 12 Frame and T 13 The frame trajectory is obtained by acquiring the ID in T. 12 Frame and T 13 Let T be the fourth driving direction angle θ4. If the signs of θ3 and θ4 are opposite, then let T be... 11 The latitude and longitude coordinates of the frame trajectory and T 10 The frames remain consistent.

[0210] The no-backward rule within the queuing area differs from the global rule. For vehicle IDs with lane number attributes, the first driving direction angle θ1 is obtained based on the slope α of the centerline of the lane to which the queuing area belongs; based on T... 12 Frame and T 13 The frame trajectory is obtained by acquiring the ID in T. 12 Frame and T 13 The second driving direction angle θ2; if θ1 and θ2 are in opposite directions, then let T 11 The latitude and longitude coordinates of the frame trajectory and T 10 The frames remain consistent.

[0211] The specific calculation methods for θ1, θ2, θ3, and θ4 are detailed above and will not be repeated here.

[0212] 3.3 The queue area disappears and vehicles fill in.

[0213] Perform a traversal check on vehicle IDs with lane attributes in frame T9. If the ID is in frame T... 10 If it does not exist within the frame, then in T 10 A new trajectory record for this ID is added within the frame. The lane attributes, vehicle type, length, width, and the ID number of the detection radar in this trajectory record are consistent with the information of this ID in frame T9. The acceleration 'a' is determined, thereby determining the displacement that occurred in this frame. Get T 10 Latitude and longitude coordinates of the frame and This type of newly added record has an added identification attribute to determine whether it is a missing vehicle.

[0214] If an ID with the aforementioned identification attributes is added, in T 10 If the distance between the preceding and following vehicles in a frame is less than d1, then delete the aforementioned vehicle in frame T. 10 The radar data with that ID is added to the frame; if an ID with the above identification attributes is added, in T 10 If the distance between the preceding and following vehicles in a frame is greater than or equal to d1, then the aforementioned condition in T is retained. 10 Radar data for that ID is added to the frame.

[0215] Additionally, among the vehicles with lane attributes in the aforementioned T9 frame, if a vehicle does not have a car in front and the direction of the lane where the vehicle is located is red (no entry), then the acceleration of that ID in the T9 frame is 0. If the direction of the lane where the vehicle is located is green (green light), the acceleration of that ID in the T9 frame is a fixed value a', where a' is the maximum acceleration of the current road segment. If a vehicle has a car in front, then the acceleration a is determined according to the car-following model.

[0216] 3.4 Queuing area following distance limit

[0217] In T 10 For a queuing area vehicle ID with lane attributes in a frame, the first vehicle, the second vehicle, and so on up to the last vehicle in the queue are determined based on the queuing area lane driving direction angle θ1.

[0218] If θ1 = 0, the vehicle with the largest longitude is the first vehicle, and the vehicle with the smallest longitude is the last vehicle.

[0219] If θ1 = π, the vehicle with the smallest longitude is the first vehicle, and the vehicle with the largest longitude is the last vehicle.

[0220] If θ1∈[0,π), the vehicle with the largest latitude is the first vehicle, and the vehicle with the smallest latitude is the last vehicle.

[0221] If θ1∈(π,2π), the vehicle with the smallest latitude is the first vehicle, and the vehicle with the largest latitude is the last vehicle.

[0222] Then, for each vehicle in each lane, the distance between the vehicle and the vehicle in front is calculated in the order from the second vehicle to the last vehicle. If the distance is less than d1, the position of the vehicle is modified, and the latitude and longitude coordinates of the vehicle are changed to the position d1 behind the vehicle in front on the center line of the lane. During this process, the order of the vehicles remains unchanged.

[0223] Furthermore, in queuing areas and prohibited lanes, acceleration and speed are inherently zero. In this situation, if the first vehicle moves forward or backward, the trajectories of subsequent vehicles will also change, causing vehicle shaking and overlap. In this embodiment of the invention, a virtual vehicle can be added to the first vehicle in the lane within the parking area. The radar data of the added virtual vehicle remains unchanged, thus the trajectories of subsequent vehicles also remain unchanged, further resolving the problems of vehicle shaking and overlap. Specifically, the virtual vehicle is added in the last frame of the previous radar data cycle and deleted in the last frame of the current radar data cycle.

[0224] Furthermore, the verification of the radar data processing method of the present invention embodiments is as follows:

[0225] For example, 12 radar devices are deployed around an intersection. The radar detects vehicle information once every 0.1 seconds, and each frame of radar data is equal to 0.1 seconds. In this scenario, when using the radar data processing method of this embodiment, the duration of a single acquisition can be 5 seconds (i.e., 50 frames). The 10th frame is taken as the current processing frame, and the output of real-time data is delayed by 4 seconds (i.e., 40 frames, to determine whether the ghosting is short or long). The following distance d1 is taken as 7.5 meters, and the processed radar trajectory length is 10 minutes.

[0226] Firstly, regarding a single vehicle, such as Figure 7 The image shows the radar trajectory path-time diagram of vehicle A5 in the first time segment of the original data. During this time segment, for vehicle A5, ghosting trajectories appear at times t1, t2, and t3, with ghost trajectory IDs of 280869, 325573, and 360015, respectively. At time t1, ghosting cannot be removed based on existing information alone, and the same applies at times t2 and t3. Therefore, the radar data processing method of this embodiment can be used to achieve real-time data cleaning and reconstruction with a 4-second delay.

[0227] like Figure 8 The image shows the trajectory result of vehicle A5 after radar data cleaning in the first time segment. Multiple trajectories were merged into one, indicating that the trajectory cleaning effect was good. Furthermore... Figure 9 and Figure 10 These are radar trajectory distance-time maps of the aforementioned vehicle A5 before and after radar data cleaning in the second time segment. Although there are differences in the ghosting situation, the cleaning effect is equally good.

[0228] Secondly, regarding multiple vehicles at intersections, such as Figure 11 The image shows the original radar trajectory path-time plots of multiple vehicles within the third time segment, as shown below. Figure 12The radar trajectory path-time diagram after removing ghosting from the radar data in the third time segment shows that the ghosting trajectory removal achieved good results.

[0229] In addition, the following indicators can be introduced to evaluate the verification results of the radar data processing method of the present invention embodiments:

[0230]

[0231] Frequency of sudden disappearance / appearance = |(Total number of detected IDs - Number of complete trajectory IDs)| / Actual number of IDs;

[0232] Vehicle trajectory completeness ratio = number of complete trajectory IDs / actual number of IDs;

[0233] Average number of ghost removal attempts per vehicle = Cumulative number of ghost removal attempts / Actual number of IDs;

[0234] Average number of stitches = Number of stitched pairs / Actual number of IDs;

[0235] Among them, the complete trajectory ID refers to the trajectory of the ID that completely includes the entire path from the entrance lane to the exit lane; the total number of detected IDs refers to the number of all IDs that appear in the radar data; and the actual number of IDs refers to the number of vehicles that actually passed through the intersection during that period, which is obtained through actual video data.

[0236] In addition, the comparison results of the indicators before and after processing the radar data of the above-mentioned intersection using the radar data processing method of the present invention are shown in Table 2 below.

[0237] Table 2 Evaluation of radar data indicators before and after application

[0238]

[0239]

[0240] In summary, the radar data processing method of this invention employs a real-time data iteration method with a delay of several seconds. This is because the real-time data contains relatively little information, insufficient to complete tasks such as ghost trajectory identification, deletion, and queuing area vehicle behavior optimization. A certain number of time frames are needed to acquire sufficient radar trajectory information. Therefore, the radar data processing method of this invention is a real-time iterative algorithm with a delay of several seconds. That is, it is a method that acquires the current processing frame and radar data from the past 3-5 seconds in real time, performs cleaning and reconstruction operations on a fixed frame of data within the time segment, and then outputs the data.

[0241] In this invention, based on radar data from short time segments of 3-5 seconds in the past, vehicles with ghosting in the original radar data are identified. Different deletion strategies are applied according to the characteristics of the ghosting. Based on a microscopic driving behavior model, and after correcting vehicle motion behavior, the broken trajectories are stitched together with high reliability, enabling continuous multi-device tracking of the target vehicle. Therefore, the radar data processing method of this embodiment uses radar data segments delayed by several seconds for analysis. Compared to real-time data analysis, this method obtains a larger amount of information and effectively improves the reliability of the cleaning and reconstruction algorithms.

[0242] like Figure 16 As shown, an embodiment of the present invention also provides a radar data processing apparatus, the radar data processing apparatus 1600 comprising:

[0243] Data acquisition module 1601 is used to acquire radar data to be processed, the radar data to be processed includes radar data of multiple radars detecting at least one object within a preset segment;

[0244] The ghosting processing module 1602 is used to determine the ghosting objects within the preset time period based on the radar data to be processed, and to delete the radar data belonging to the ghosting objects from the radar data to be processed.

[0245] Optionally, the ghosting processing module includes:

[0246] The long-duration ghosting deletion submodule is used to determine, based on the radar data to be processed, the first ghosting object whose trajectory duration is greater than a first preset value within the preset time period, and delete the radar data belonging to the first ghosting object from the radar data to be processed;

[0247] The short-term ghosting deletion submodule is used to determine, based on the radar data to be processed, a second ghosting object whose trajectory duration is less than or equal to the first preset value within the preset time period, and to delete the radar data belonging to the second ghosting object from the radar data to be processed.

[0248] Optionally, the long-term ghosting removal submodule is specifically used for:

[0249] When i takes the integer value from 1 to M, the following process is executed:

[0250] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0251] Obtain the fourth identifier that meets the third preset condition in the radar data of the i-th period, and store the fourth identifier in the first set, wherein the third preset condition includes that it exists in the (X-1)-th frame, the X-th frame, and the Y+1-th frame of the radar data of the i-th period;

[0252] Obtain the fifth identifier that meets the fourth preset condition in the radar data of the i-th period, and store the fifth identifier in the second set, wherein the fourth preset condition includes that it exists in the (X-1)-th frame, and does not exist in the X-th frame and the Y+1-th frame;

[0253] Obtain a first target identifier from the first set and a second target identifier from the second set, wherein the distance between the objects represented by the first target identifier and the second target identifier in the (X-1)th frame is less than a first preset value;

[0254] Delete the radar data corresponding to the second target identifier in the radar data from frame X-1 to frame Y+1 of the i-th cycle.

[0255] Optionally, the radar data includes object identifiers; the short-term ghosting removal module is specifically used for:

[0256] When i takes the integer value from 1 to M, the following process is executed:

[0257] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0258] Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1;

[0259] Obtain a second identifier from the first identifier that meets the first preset condition and the second preset condition, wherein the first preset condition includes that it exists in the (X+1)th frame of the i-th period radar data and does not exist in the Y-th and Y+1th frames of the i-th period radar data, and the second preset condition includes that it does not exist in the (X-1)th frame of the i-th period radar data, where Y = N-1;

[0260] Delete the radar data corresponding to the second identifier in the radar data of the i-th cycle.

[0261] Optionally, the radar data includes object identifiers; the short-term ghosting removal submodule is also used for:

[0262] Obtain the sixth identifier that meets the fifth preset condition in the radar data of the i-th cycle, and store the sixth identifier in the third set. The fifth preset condition includes that it exists in both the (X-1)th frame and the Xth frame in the radar data of the i-th cycle, and does not exist in both the Yth frame and the Y+1th frame in the radar data of the i-th cycle, where X is an integer greater than 1 and Y = N-1.

[0263] Obtain the seventh identifier that meets the sixth preset condition in the radar data of the i-th period, and store the seventh identifier in the fourth set, wherein the sixth preset condition includes not existing in the (X-1)-th frame and existing in the X-th frame;

[0264] A third target identifier is obtained from the third set, and a fourth target identifier is obtained from the fourth set, wherein the distance between the objects represented by the third target identifier and the fourth target identifier in the (X-1)th frame is less than a first preset value, and the distance between the objects represented by the third target identifier and the fourth target identifier in the target frame is less than the first preset value, and the target frame is the last frame in which the third target identifier appears in the radar data of the i-th cycle;

[0265] Delete the radar data corresponding to the third target identifier in the target frame from the (X-1)th frame to the target frame in the i-th period radar data.

[0266] Optionally, the radar data includes object identifiers; the device further includes a frame completion module, wherein the frame completion module is specifically used for:

[0267] When i takes the integer value from 1 to M, the following process is executed:

[0268] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0269] Obtain the eighth identifier in the i-th period radar data that meets the seventh preset condition, wherein the seventh preset condition includes that it exists in both the Y-1 and Y+1 frames in the i-th period radar data, and does not exist in the Y-th frame in the i-th period radar data, Y = N-1;

[0270] Based on the radar data of the object represented by the eighth identifier in the (Y-1)th frame and the radar data in the (Y+1)th frame, determine the radar data of the object represented by the eighth identifier in the Yth frame.

[0271] Optionally, the radar data includes object type, object size, object number information, velocity, and latitude and longitude coordinates; when the frame completion module determines the radar data of the object represented by the eighth identifier in the Y-th frame based on the radar data of the object represented by the eighth identifier in the Y-1 frame and the radar data in the Y+1 frame, it is specifically used for:

[0272] The object type, object size, and number information of the object represented by the eighth identifier in the (Y-1)th frame are respectively determined as the object type, object size, and number information of the object represented by the eighth identifier in the Yth frame;

[0273] The speed of the object represented by the eighth identifier in the (Y-1)th frame and the speed of the object represented by the eighth identifier in the (Y+1)th frame are taken as the speed of the object represented by the eighth identifier in the Yth frame.

[0274] The latitude and longitude coordinates of the object represented by the eighth identifier in the (Y-1)th frame and the average latitude and longitude coordinates of the object represented by the eighth identifier in the (Y+1)th frame are determined as the latitude and longitude coordinates of the object represented by the eighth identifier in the Yth frame.

[0275] Optionally, the radar data includes the latitude and longitude coordinates of the object; the device further includes:

[0276] The first constraint module is used to adjust the latitude and longitude coordinates of the object in the radar data to be processed based on the positional relationship between the detected object of the radar and the boundary line of a pre-determined target area.

[0277] Optionally, the radar data may also include object identifiers; the first constraint module is specifically used for:

[0278] When i takes the integer value from 1 to M, the following process is executed:

[0279] The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1.

[0280] Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1;

[0281] Obtain the identifiers whose latitude and longitude coordinates are within the boundary line of the target area from the first identifier, and use them as the ninth identifier;

[0282] The latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are adjusted to the first target coordinates, wherein the straight line containing the first point and the second point is perpendicular to the center line of the boundary line of the target area, and the first point is located on the center line. The first point is the point represented by the first target coordinates, and the second point is the point represented by the original latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame.

[0283] Optionally, the device further includes: a second constraint module;

[0284] The second constraint module is used for:

[0285] Based on the slope of the centerline, determine the first driving direction angle of the object represented by the ninth identifier in the Xth frame and the X+1th frame of the i-th period radar data;

[0286] Based on the latitude and longitude coordinates of the object represented by the ninth identifier in the X+2 and X+3 frames of the i-th cycle radar data, determine the second driving direction angle of the object represented by the ninth identifier in the X+2 and X+3 frames;

[0287] When the signs of the first driving direction angle and the second driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the ninth identifier in the X+1th frame.

[0288] Optionally, the second constraint module is further configured to:

[0289] Obtain the identifier in the first identifier whose latitude and longitude coordinates are outside the boundary line of the target area, and use it as the tenth identifier;

[0290] Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the first and Xth frames of the i-th cycle radar data, determine the third driving direction angle of the object represented by the tenth identifier in the (X-1)th and Xth frames;

[0291] Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the X+2 and X+3 frames of the i-th cycle radar data, determine the fourth driving direction angle of the object represented by the tenth identifier in the X+2 and X+3 frames;

[0292] When the signs of the third driving direction angle and the fourth driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the tenth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the tenth identifier in the X+1th frame.

[0293] Optionally, the device further includes: a third constraint module;

[0294] The third constraint module is used for:

[0295] From the ninth identifier, select the identifier that does not exist in the (X-1)th frame of the i-th period radar data, and use it as the eleventh identifier;

[0296] Determine the acceleration of the object represented by the eleventh identifier in the (X-1)th frame;

[0297] Based on the acceleration and velocity of the object represented by the eleventh identifier in the (X-1)th frame, determine the target displacement of the object represented by the eleventh identifier, wherein the target displacement is the displacement of the object represented by the eleventh identifier from its position in the (X-1)th frame within one frame time.

[0298] Based on the target displacement, determine the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame, and use them as the second target coordinates;

[0299] If the distance between the second target coordinates and the first reference coordinates is greater than or equal to a first preset value, and the distance between the second target coordinates and the second reference coordinates is greater than or equal to the first preset value, the eleventh identifier is added to the Xth frame, and the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame are determined to be the second target coordinates.

[0300] Wherein, the first reference coordinate and the second reference coordinate are the latitude and longitude coordinates of two objects located before and after the second target coordinate in the Xth frame, respectively.

[0301] Optionally, when determining the acceleration of the object represented by the eleventh identifier in the (X-1)th frame, the third constraint module is specifically used for:

[0302] If the object represented by the eleventh identifier is obtained as the following object in the X-1 frame, the acceleration of the object represented by the eleventh identifier in the X-1 frame is determined based on the speed and latitude / longitude coordinates of the object represented by the eleventh identifier in the X-1 frame, the speed and latitude / longitude coordinates of the following object in the X-1 frame, and a predetermined following model.

[0303] If the following object is not obtained, the acceleration of the object represented by the eleventh identifier in the (X-1)th frame is determined as the preset acceleration.

[0304] Optionally, the device further includes: a fourth constraint module;

[0305] The fourth constraint module is used for:

[0306] Obtain the sorted position of the object represented by the ninth identifier in the Xth frame;

[0307] When j takes the value 1 K, the following process is executed:

[0308] Calculate the distance between the j-th object and the (j-1)-th object in the position sorting, and if the distance is less than a first preset value, determine the latitude and longitude coordinates of the j-th object in the X-th frame as the third target coordinates, wherein the third target coordinates are located on the center line, the distance between the third target coordinates and the latitude and longitude coordinates of the (j-1)-th object in the X-th frame is equal to the first preset value, and K is the number of the ninth identifiers.

[0309] This invention also provides an electronic device, such as... Figure 17 As shown, it includes a processor 171, a communication interface 172, a memory 173, and a communication bus 174, wherein the processor 171, the communication interface 172, and the memory 173 communicate with each other through the communication bus 174.

[0310] Memory 173 is used to store computer programs;

[0311] The processor 171 is used to execute the program stored in the memory 173 to implement the radar data processing method described above.

[0312] The aforementioned memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0313] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0314] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform any of the radar data processing methods described in the above embodiments.

[0315] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the radar data processing methods described in the above embodiments.

[0316] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0317] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0318] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0319] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A radar data processing method, characterized in that, The method includes: Acquire radar data to be processed, which includes radar data detected by multiple radars on at least one object within a preset time period; Based on the radar data to be processed, determine the ghosting objects within the preset time period, and delete the radar data belonging to the ghosting objects from the radar data to be processed; The step of determining the ghosting objects within the preset time period based on the radar data to be processed, and deleting the radar data belonging to the ghosting objects from the radar data to be processed, includes: Based on the radar data to be processed, identify the first ghost object whose trajectory duration is greater than the first preset value within the preset time period, and delete the radar data belonging to the first ghost object from the radar data to be processed. Based on the radar data to be processed, determine the second ghost object whose trajectory duration is less than or equal to the first preset value within the preset time period, and delete the radar data belonging to the second ghost object from the radar data to be processed. The step of determining, based on the radar data to be processed, a first ghosting object whose trajectory duration is greater than a first preset value within the preset time period, and deleting radar data belonging to the first ghosting object from the radar data to be processed, includes: When i takes the integer value from 1 to M, the following process is executed: The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1. Obtain the fourth identifier that meets the third preset condition in the radar data of the i-th period, and store the fourth identifier in the first set, wherein the third preset condition includes that it exists in the (X-1)-th frame, the X-th frame, and the Y+1-th frame of the radar data of the i-th period; Obtain the fifth identifier that meets the fourth preset condition in the radar data of the i-th period, and store the fifth identifier in the second set, wherein the fourth preset condition includes that it exists in the (X-1)-th frame, and does not exist in the X-th frame and the Y+1-th frame; Obtain a first target identifier from the first set and a second target identifier from the second set, wherein the distance between the objects represented by the first target identifier and the second target identifier in the (X-1)th frame is less than a first preset value; Delete the radar data corresponding to the second target identifier in the (X-1)th to (Y+1)th frames of the i-th period radar data.

2. The radar data processing method according to claim 1, characterized in that, Radar data includes object identifiers; The step of determining, based on the radar data to be processed, a second ghost object whose trajectory duration within the preset time period is less than or equal to the first preset value, and deleting radar data belonging to the second ghost object from the radar data to be processed, includes: When i takes the integer value from 1 to M, the following process is executed: The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1. Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1; Obtain a second identifier from the first identifier that meets the first preset condition and the second preset condition, wherein the first preset condition includes that it exists in the (X+1)th frame of the i-th period radar data and does not exist in the Y-th and Y+1th frames of the i-th period radar data, and the second preset condition includes that it does not exist in the (X-1)th frame of the i-th period radar data, where Y=N-1; Delete the radar data corresponding to the second identifier in the radar data of the i-th cycle.

3. The radar data processing method according to claim 2, characterized in that, Radar data includes object identifiers; The step of determining a second ghosting object whose trajectory duration within the preset time period is less than or equal to the first preset value based on the radar data to be processed, and deleting radar data belonging to the second ghosting object from the radar data to be processed, further includes: Obtain the sixth identifier that meets the fifth preset condition in the radar data of the i-th cycle, and store the sixth identifier in the third set. The fifth preset condition includes that it exists in both the (X-1)th frame and the Xth frame in the radar data of the i-th cycle, and does not exist in both the Yth frame and the Y+1th frame in the radar data of the i-th cycle, where X is an integer greater than 1 and Y=N-1. Obtain the seventh identifier that meets the sixth preset condition in the radar data of the i-th period, and store the seventh identifier in the fourth set, wherein the sixth preset condition includes not existing in the (X-1)-th frame and existing in the X-th frame; A third target identifier is obtained from the third set, and a fourth target identifier is obtained from the fourth set, wherein the distance between the objects represented by the third target identifier and the fourth target identifier in the (X-1)th frame is less than a first preset value, and the distance between the objects represented by the third target identifier and the fourth target identifier in the target frame is less than the first preset value, and the target frame is the last frame in which the third target identifier appears in the radar data of the i-th cycle; Delete the radar data corresponding to the third target identifier in the target frame from the (X-1)th frame to the target frame in the i-th period radar data.

4. The radar data processing method according to claim 1, characterized in that, Radar data includes object identifiers; Before determining the first ghosting object with a trajectory duration greater than a first preset value within the preset time period based on the radar data to be processed, and deleting the radar data belonging to the first ghosting object from the radar data to be processed, the method further includes: When i takes the integer value from 1 to M, the following process is executed: The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1. Obtain the eighth identifier in the radar data of the i-th cycle that meets the seventh preset condition, wherein the seventh preset condition includes that it exists in both the Y-1 frame and the Y+1 frame in the radar data of the i-th cycle, and does not exist in the Y frame in the radar data of the i-th cycle, Y=N-1; Based on the radar data of the object represented by the eighth identifier in the (Y-1)th frame and the radar data in the (Y+1)th frame, determine the radar data of the object represented by the eighth identifier in the Yth frame.

5. The radar data processing method according to claim 4, characterized in that, Radar data includes object type, object size, object identification number, velocity, and latitude and longitude coordinates; The step of determining the radar data of the object represented by the eighth identifier in the Y-1 frame based on the radar data of the object represented by the eighth identifier in the Y-1 frame and the radar data in the Y+1 frame includes: The object type, object size, and number information of the object represented by the eighth identifier in the (Y-1)th frame are respectively determined as the object type, object size, and number information of the object represented by the eighth identifier in the Yth frame; The speed of the object represented by the eighth identifier in the (Y-1)th frame and the speed of the object represented by the eighth identifier in the (Y+1)th frame are taken as the speed of the object represented by the eighth identifier in the Yth frame. The latitude and longitude coordinates of the object represented by the eighth identifier in the (Y-1)th frame and the average latitude and longitude coordinates of the object represented by the eighth identifier in the (Y+1)th frame are determined as the latitude and longitude coordinates of the object represented by the eighth identifier in the Yth frame.

6. The radar data processing method according to claim 1, characterized in that, Radar data includes the latitude and longitude coordinates of the object; After determining the ghosting objects within the preset time period based on the radar data to be processed, and deleting the radar data belonging to the ghosting objects from the radar data to be processed, the method further includes: Based on the positional relationship between the radar's detected object and the predetermined boundary line of the target area, the latitude and longitude coordinates of the object in the radar data to be processed are adjusted.

7. The radar data processing method according to claim 6, characterized in that, Radar data also includes object identification; The step of adjusting the latitude and longitude coordinates of the object in the radar data to be processed based on the positional relationship between the detected object of the radar and the boundary line of a pre-determined target area includes: When i takes the integer value from 1 to M, the following process is executed: The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1. Obtain the first identifier in the Xth frame of the i-th period radar data, where X is an integer greater than 1; Obtain the identifiers whose latitude and longitude coordinates are within the boundary line of the target area from the first identifier, and use them as the ninth identifier; The latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are adjusted to the first target coordinates, wherein the straight line containing the first point and the second point is perpendicular to the center line of the boundary line of the target area, and the first point is located on the center line. The first point is the point represented by the first target coordinates, and the second point is the point represented by the original latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame.

8. The radar data processing method according to claim 7, characterized in that, After adjusting the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame to the first target coordinates, the method further includes: Based on the slope of the centerline, determine the first driving direction angle of the object represented by the ninth identifier in the Xth frame and the X+1th frame of the i-th period radar data; Based on the latitude and longitude coordinates of the object represented by the ninth identifier in the X+2 and X+3 frames of the i-th cycle radar data, determine the second driving direction angle of the object represented by the ninth identifier in the X+2 and X+3 frames; When the signs of the first driving direction angle and the second driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the ninth identifier in the X+1th frame.

9. The radar data processing method according to claim 8, characterized in that, The method further includes: Obtain the identifier in the first identifier whose latitude and longitude coordinates are outside the boundary line of the target area, and use it as the tenth identifier; Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the first and Xth frames of the i-th cycle radar data, determine the third driving direction angle of the object represented by the tenth identifier in the (X-1)th and Xth frames; Based on the latitude and longitude coordinates of the object represented by the tenth identifier in the X+2 and X+3 frames of the i-th cycle radar data, determine the fourth driving direction angle of the object represented by the tenth identifier in the X+2 and X+3 frames; When the signs of the third driving direction angle and the fourth driving direction angle are opposite, the latitude and longitude coordinates of the object represented by the tenth identifier in the Xth frame are determined as the latitude and longitude coordinates of the object represented by the tenth identifier in the X+1th frame.

10. The radar data processing method according to claim 8, characterized in that, After determining the latitude and longitude coordinates of the object represented by the ninth identifier in the Xth frame as the latitude and longitude coordinates of the object represented by the ninth identifier in the (X+1)th frame, the method further includes: From the ninth identifier, select the identifier that does not exist in the (X-1)th frame of the i-th period radar data, and use it as the eleventh identifier; Determine the acceleration of the object represented by the eleventh identifier in the (X-1)th frame; Based on the acceleration and velocity of the object represented by the eleventh identifier in the (X-1)th frame, determine the target displacement of the object represented by the eleventh identifier, wherein the target displacement is the displacement of the object represented by the eleventh identifier from its position in the (X-1)th frame within one frame time. Based on the target displacement, determine the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame, and use them as the second target coordinates; If the distance between the second target coordinates and the first reference coordinates is greater than or equal to a first preset value, and the distance between the second target coordinates and the second reference coordinates is greater than or equal to the first preset value, the eleventh identifier is added to the Xth frame, and the latitude and longitude coordinates of the object represented by the eleventh identifier in the Xth frame are determined to be the second target coordinates. Wherein, the first reference coordinate and the second reference coordinate are the latitude and longitude coordinates of two objects located before and after the second target coordinate in the Xth frame, respectively.

11. The radar data processing method according to claim 10, characterized in that, Determining the acceleration of the object represented by the eleventh identifier in the (X-1)th frame includes: If the object represented by the eleventh identifier is obtained as the following object in the X-1 frame, the acceleration of the object represented by the eleventh identifier in the X-1 frame is determined based on the speed and latitude / longitude coordinates of the object represented by the eleventh identifier in the X-1 frame, the speed and latitude / longitude coordinates of the following object in the X-1 frame, and a predetermined following model. If the following object is not obtained, the acceleration of the object represented by the eleventh identifier in the (X-1)th frame is determined as the preset acceleration.

12. The radar data processing method according to claim 10, characterized in that, After adding the eleventh identifier to the Xth frame, representing the latitude and longitude coordinates of the object in the Xth frame as indicated by the ninth identifier, and determining that the latitude and longitude coordinates of the object in the Xth frame represented by the eleventh identifier are the second target coordinates, the method further includes: Obtain the sorted position of the object represented by the ninth identifier in the Xth frame; When j is 1 and the value K is , the following process is executed: Calculate the distance between the j-th object and the (j-1)-th object in the position sorting, and if the distance is less than a first preset value, determine the latitude and longitude coordinates of the j-th object in the X-th frame as the third target coordinates, wherein the third target coordinates are located on the center line, the distance between the third target coordinates and the latitude and longitude coordinates of the (j-1)-th object in the X-th frame is equal to the first preset value, and K is the number of the ninth identifiers.

13. A radar data processing device, characterized in that, The device includes: The data acquisition module is used to acquire radar data to be processed, which includes radar data of multiple radars detecting at least one object within a preset time period. The ghosting processing module is used to determine the ghosting objects within the preset time period based on the radar data to be processed, and to delete the radar data belonging to the ghosting objects from the radar data to be processed. The ghosting processing module includes: The long-duration ghosting deletion submodule is used to determine, based on the radar data to be processed, the first ghosting object whose trajectory duration is greater than a first preset value within the preset time period, and delete the radar data belonging to the first ghosting object from the radar data to be processed; The short-term ghosting deletion submodule is used to determine, based on the radar data to be processed, a second ghosting object whose trajectory duration is less than or equal to the first preset value within the preset time period, and to delete the radar data belonging to the second ghosting object from the radar data to be processed; The long-term ghosting deletion submodule is specifically used for: When i takes the integer value from 1 to M, the following process is executed: The radar data from the i-th frame to the (i+N)-th frame of the radar data to be processed is determined as the radar data of the i-th period, where M is the difference between the total number of frames included in the radar data to be processed and N, and N is an integer greater than 1. Obtain the fourth identifier that meets the third preset condition in the radar data of the i-th period, and store the fourth identifier in the first set, wherein the third preset condition includes that it exists in the (X-1)-th frame, the X-th frame, and the Y+1-th frame of the radar data of the i-th period; Obtain the fifth identifier that meets the fourth preset condition in the radar data of the i-th period, and store the fifth identifier in the second set, wherein the fourth preset condition includes that it exists in the (X-1)-th frame, and does not exist in the X-th frame and the Y+1-th frame; Obtain a first target identifier from the first set and a second target identifier from the second set, wherein the distance between the objects represented by the first target identifier and the second target identifier in the (X-1)th frame is less than a first preset value; Delete the radar data corresponding to the second target identifier in the (X-1)th to (Y+1)th frames of the i-th period radar data.

14. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in a memory, implements the radar data processing method according to any one of claims 1-12.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the radar data processing method as described in any one of claims 1-12.

Images