Target state detection method and device, storage medium and product

Abstract

The invention discloses a target state detection method and device, a storage medium and a product, and relates to the field of radars. The detection method comprises the following steps: acquiring target trajectory data; wherein the target trajectory data comprises a target current motion state and target detection data obtained according to a current frame radar echo signal; the motion state comprises a normal driving state, a parking state and a starting state, and the target detection data comprises a target speed; and judging the motion state of the next frame of the target according to the target track data currently in different motion states, thereby realizing detection and updating of the target motion state. According to the invention, the detection of the motion state of each target is independent and does not need to depend on a stop line of an intersection and the parking state of a front vehicle, so that the problems that traffic radar parameters are difficult to configure and large-scale continuous misjudgment is easily caused are avoided; the detection precision of the radar on the target speed is high, and the correctness of target motion state judgment is ensured.

Description

Technical Field

[0001] The present invention belongs to the field of radar technology, and in particular relates to a target state detection method, device, storage medium and product. Background Art

[0002] Traffic radar is used to detect traffic participants on the road, which usually include motor vehicles, non-motor vehicles, pedestrians, etc. Traffic radar is usually installed on the crossbar with speed cameras and can obtain information such as the distance, angle, speed, etc. of the target. However, due to factors such as environmental interference, traffic radar usually filters out static clutter in the signal processing algorithm, resulting in the inability to detect stationary targets.

[0003] Vehicles often need to stop moving due to factors such as waiting for red lights or congestion. At this time, traffic radar cannot detect stationary vehicles and can only determine whether the vehicle has stopped based on the detection data of the vehicle's past movement in the signal processing algorithm, and then retain the target status to maintain continuous tracking of the target.

[0004] At present, the specific steps of the target vehicle status identification solution include:

[0005] S1: Traverse all trajectories. When the speed of the target vehicle is lower than the threshold and multiple frames of detection data are lost continuously, the target vehicle is taken as the target to be identified.

[0006] S2: Filter out the front vehicle track in the same lane as the target to be identified from all tracks;

[0007] S3: If the vehicle ahead stops and the distance to the target to be identified is less than a threshold, or the distance to the stop line at the traffic intersection is less than a threshold, the target to be identified is considered to have entered a stopped state; if not, the target to be identified is considered not to have entered a stopped state.

[0008] This method determines whether to stop the maneuver based on the change in target speed and the situation of the target relative to the stop line. It relies on the demarcation of the stop line. The stop line distance is not easy to obtain when installing the traffic radar, which makes it difficult to configure the traffic radar installation parameters. At the same time, the applicable scene is single, and it is only applicable to scenes with stop lines such as intersections, and cannot be applied on ordinary roads. In addition, this method also relies on the status information of the preceding vehicle. The parking status of the preceding vehicle is also determined by the previous parking strategy, resulting in the correctness of the parking status judgment of the preceding vehicle affecting the parking status judgment of the following vehicle. If the preceding vehicle is misjudged, it will lead to a chain effect such as continuous misjudgment of the following vehicle, affecting the final parking queue detection result.

[0009] Traffic radar can also be combined with visual detection equipment to identify the movement status of targets, which increases complexity while introducing more data and leads to higher costs. Summary of the invention

[0010] The purpose of the present invention is to provide a target state detection method, device, storage medium and product to solve the problem that traditional parking queue detection relies on the stop line and the parking state of the preceding vehicle, resulting in difficulty in configuring traffic radar parameters, single applicable scenario and easy to cause continuous misjudgment.

[0011] The present invention solves the above technical problems through the following technical solutions: a target state detection method, comprising:

[0012] Acquire target trajectory data; wherein the target trajectory data includes the current motion state of the target and target detection data obtained according to the radar echo signal of the current frame; the motion state includes a normal driving state, a parking state and a starting state, and the target detection data includes a target speed;

[0013] Traverse all target trajectory data that are currently in a normal driving state. If the target speed in the target trajectory data is lower than a set speed threshold and is not associated with new target detection data in consecutive frames, determine that the target corresponding to the target trajectory data has entered a parking state, and update the current motion state of the target in the target trajectory data; otherwise, determine that the target corresponding to the target trajectory data has not entered a parking state, and update the target detection data in the target trajectory data;

[0014] Traversing all target trajectory data currently in the parking state, if the target trajectory data is associated with new target detection data, determining that the target corresponding to the target trajectory data has entered the starting state, and updating the target current motion state and target detection data in the target trajectory data; otherwise, determining that the target corresponding to the target trajectory data is still in the parking state;

[0015] All target trajectory data currently in the starting state are traversed. If the number of times that the target trajectory data is associated with the target detection data since the parking state reaches a set first number threshold, it is determined that the target corresponding to the target trajectory data enters the normal driving state, and the target current motion state and target detection data in the target trajectory data are updated; if the number of times that the target trajectory data is not associated with the target detection data since the parking state reaches a set second number threshold, it is determined that the target corresponding to the target trajectory data enters the parking state, and the target current motion state and target detection data in the target trajectory data are updated; if the number of times that the target trajectory data is not associated with the target detection data since the parking state does not reach the set second number threshold, it is determined that the target corresponding to the target trajectory data is still in the starting state, and the target detection data in the target trajectory data is updated.

[0016] Further, if the target speed in the target trajectory data is lower than a set speed threshold and is not associated with new target detection data in consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated, specifically including:

[0017] If the target speed in the target trajectory data is lower than a set first speed threshold and is not associated with new target detection data within at least N1 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated;

[0018] If the target speed in the target trajectory data is lower than a set second speed threshold and no new target detection data is associated within at least N2 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated;

[0019] If the target speed in the target trajectory data is lower than a set third speed threshold and is not associated with new target detection data within at least N3 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated;

[0020] If the target speed in the target trajectory data is lower than a set fourth speed threshold and is not associated with new target detection data within at least N4 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated;

[0021] Among them, the first speed threshold < the second speed threshold < the third speed threshold < the fourth speed threshold, N1 < N2 < N3 < N4.

[0022] Furthermore, the first speed threshold is 1 m / s, the second speed threshold is 3 m / s, the third speed threshold is 5 m / s, the fourth speed threshold is 8 m / s, N1 is 5, N2 is 10, N3 is 15, and N4 is 20.

[0023] Furthermore, the first number threshold is set to 3, and the second number threshold is set to 4.

[0024] Furthermore, the detection method further includes determining whether two targets overlap in space, and adjusting the positions of the two targets that overlap in space, specifically including:

[0025] Acquire road information; wherein the road information includes the number of lanes, the width of each lane, and the horizontal offset of the radar installation position, wherein the horizontal offset of the radar installation position refers to the vertical distance from the radar installation position to the road sideline;

[0026] Calculating the range of each lane in the radar coordinate system according to the number of lanes, the width of each lane, and the horizontal offset of the radar installation position;

[0027] For targets in normal driving or starting state, the target position is determined based on the radar echo signal of the current frame; for targets in parking state, the target position is determined based on the position of the target when it enters the parking state;

[0028] Determine the lane information to which the target belongs according to the range of each lane and the target position;

[0029] Determine whether two targets are in the same lane according to the lane information to which the targets belong;

[0030] Determine the target type and then the target length according to the radar echo signal of the current frame;

[0031] For two targets in the same lane, whether there is spatial overlap is determined based on the target positions and target lengths of the two targets. If so, the positions of the two targets with spatial overlap are adjusted; if not, the positions of the two targets with spatial overlap are not adjusted.

[0032] Furthermore, the range of the i-th lane is:

[0033] [-x0+(w1+…+w i-1 ),-x0+(w1+…+w i )];

[0034] Among them, x0 represents the horizontal offset of the radar installation position, w1, ..., w i-1 、w i They represent the width of the 1st, ..., i-1, and i lanes respectively. When i=1, w0=0.

[0035] Furthermore, the positions of two targets with spatial overlap are adjusted, including:

[0036] Determine the front-to-back positional relationship of the two targets according to the target positions of the two targets; assume that target A is located in front of target B, and the front refers to the direction close to the radar;

[0037] When both target A and target B are in a non-parking state and target detection data is not lost, the target positions of target A and target B remain unchanged;

[0038] When target B is in a non-parking state and has not lost target detection data, and target A is in a non-parking state and has lost target detection data, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap;

[0039] When target B is in a non-parking state and has not lost target detection data, and target A is in a parking state, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap;

[0040] When target B is in a non-parking state and loses target detection data, and target A is in a non-parking state and does not lose target detection data, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap;

[0041] When both target A and target B are in a non-parking state and the target detection data is lost, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap;

[0042] When target B is in a non-parking state and loses target detection data, and target A is in a parking state, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap;

[0043] When target B is in a parking state, and target A is in a non-parking state and has not lost target detection data, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap;

[0044] When target B is in a parking state, and target A is in a non-parking state and loses target detection data, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap;

[0045] When both target A and target B are in the parking state, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap.

[0046] Based on the same concept, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the target state detection method as described above.

[0047] Based on the same concept, the present invention also provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the target state detection method as described above is implemented.

[0048] Based on the same concept, the present invention also provides a computer program product, including a computer program / instruction, which implements the target state detection method as described above when executed by a processor.

[0049] Beneficial Effects

[0050] Compared with the prior art, the advantages of the present invention are:

[0051] The present invention detects the next motion state of a target according to the current motion state of the target, the target speed and whether new target detection data is associated in continuous frames. The motion state detection of each target is independent and does not need to rely on the stop line of the intersection and the parking state of the preceding vehicle, thus avoiding the difficulty in configuring traffic radar parameters and the problem of large-scale continuous misjudgment. The radar has a high detection accuracy for the target speed, thus ensuring the correctness of the target motion state determination. The present invention can be applied to sections with and without stop lines, to a variety of scenarios, and can detect illegal parking and other behaviors.

[0052] The present invention can be used to guide the tracking algorithm to predict the maneuvering behavior that may occur to the target, especially to propose the conversion from the starting state to the parking state, and can quickly correct the target back to the parking state after it is erroneously converted to the starting state. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 is a flow chart of a target state detection method in an embodiment of the present invention;

[0055] Figure 2 is a target motion state conversion relationship diagram in an embodiment of the present invention;

[0056] Figure 3 is a schematic diagram of the target motion state in the 400th frame of target trajectory data in an embodiment of the present invention;

[0057] Figure 4 is a diagram illustrating the current motion state of a target in an embodiment of the present invention;

[0058] Figure 5 is a schematic diagram of the target motion state in the 745th frame of target trajectory data in an embodiment of the present invention;

[0059] Figure 6 It is a schematic diagram of the target motion state in the 858th frame of target trajectory data in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The following is a clear and complete description of the technical solutions in the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0062] Example 1

[0063] Figure 1 FIG. 2 shows a flow chart of the target state detection method provided by the present invention. Figure 1 As shown, the target state detection method includes the following steps:

[0064] Step 1: Get target trajectory data.

[0065] The target trajectory data includes the current motion state of the target and the target detection data obtained according to the radar echo signal of the current frame, that is, the target trajectory data includes the latest motion state of the target and the latest target detection data. The targets in this embodiment include motor vehicles and non-motor vehicles, and the motion state of the target includes normal driving state, parking state and starting state. The normal driving state can be converted to the parking state, the parking state can be converted to the starting state, and the starting state can be converted to the normal driving state or the parking state. Figure 2 As shown in the figure, since the traffic radar can only detect moving targets, the initial motion state of the targets is set to the normal driving state.

[0066] Target detection data includes target speed, target distance and target azimuth. Target distance refers to the straight-line distance from the target to the traffic radar antenna surface. The traffic radar radiates electromagnetic wave signals. The electromagnetic wave signals are reflected after passing through the target. The traffic radar receives the reflected echo signals and processes the echo signals to obtain the target distance, target azimuth and target speed. The target position can be determined based on the target distance and target azimuth. The target position is used to judge whether there is spatial overlap in the future.

[0067] According to the current running state of the target in the target trajectory data, it can be judged whether each target trajectory data is currently in a normal driving state, a parking state, or a starting state.

[0068] Step 2: Determine the motion state of the target trajectory data currently in different motion states, and update the current motion state of the target in the target trajectory data according to the determination result.

[0069] For the target trajectory data that is currently in a normal driving state, if the target speed in the target trajectory data is lower than the set speed threshold and is not associated with new target detection data in a continuous frame, it is determined that the target corresponding to the target trajectory data has entered a parking state, and the current motion state of the target in the target trajectory data is updated, that is, the normal driving state in the target trajectory data is converted to a parking state, and the position corresponding to the moment of entering the parking state is the parking position of the target, and the target speed in the parking state is always 0. If the target speed in the target trajectory data is higher than or equal to the set speed threshold, and / or is associated with new target detection data in a continuous frame, it is determined that the target corresponding to the target trajectory data has not entered a parking state, the current running state in the target trajectory data is still a normal driving state, and the target detection data in the target trajectory data is updated according to the latest target detection data.

[0070] Since existing traffic radars can only detect moving targets in order to filter out environmental clutter, when no new target detection data is associated in consecutive frames (i.e., the radar does not detect the target in consecutive frames or the detection data is lost in consecutive frames), it indicates that the target is stationary and is determined to be in a parked state.

[0071] In this embodiment, the speed threshold and the number of consecutive frames are set according to actual needs. In order to more accurately detect the parking position of the target, the speed threshold and the number of consecutive frames are divided into multiple gears, and the parking state is determined according to different gears, and then the parking position is determined, that is, the position of the target at the moment it enters the parking state is the parking position of the target. Specifically, if the target speed in the target trajectory data is lower than the set first speed threshold and is not associated with new target detection data within at least N1 consecutive frames, it is determined that the target corresponding to the target trajectory data enters the parking state, and the current motion state of the target in the target trajectory data is updated, that is, the normal driving state in the target trajectory data is converted to the parking state;

[0072] If the target speed in the target trajectory data is lower than the set second speed threshold and is not associated with new target detection data within at least N2 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated, that is, the normal driving state in the target trajectory data is converted into a parking state;

[0073] If the target speed in the target trajectory data is lower than the set third speed threshold and is not associated with new target detection data within at least N3 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated, that is, the normal driving state in the target trajectory data is converted to a parking state;

[0074] If the target speed in the target trajectory data is lower than the set fourth speed threshold and is not associated with new target detection data within at least N4 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated, that is, the normal driving state in the target trajectory data is converted to a parking state;

[0075] Among them, the first speed threshold < the second speed threshold < the third speed threshold < the fourth speed threshold, N1 < N2 < N3 < N4.

[0076] For a traffic radar with a detection period of 80ms, the first speed threshold is set to 1m / s, the second speed threshold is set to 3m / s, the third speed threshold is set to 5m / s, the fourth speed threshold is set to 8m / s, N1 is set to 5, N2 is set to 10, N3 is set to 15, and N4 is set to 20. Exemplarily, if the target speed in the target trajectory data is lower than 1m / s and is not associated with new target detection data within at least 5 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state.

[0077] For target trajectory data that is currently in a parking state, if the target trajectory data is re-associated with new target detection data, it is determined that the target corresponding to the target trajectory data enters a starting state, and the target current motion state and target detection data in the target trajectory data are updated, that is, the parking state in the target trajectory data is converted to a starting state, and the new target detection data is used to replace the original target detection data; otherwise, it is determined that the target corresponding to the target trajectory data is still in a parking state.

[0078] Since existing traffic radars can only detect moving targets in order to filter out environmental clutter, when new target detection data is associated (i.e., the radar detects a target), it indicates that the target is in motion and is determined to have entered the starting state.

[0079] For the target trajectory data currently in the starting state, if the number of times the target trajectory data is associated with the target detection data since the parking state reaches the set first number threshold, it is determined that the target corresponding to the target trajectory data has entered the normal driving state, and the target current motion state and target detection data in the target trajectory data are updated, that is, the starting state in the target trajectory data is converted to the normal driving state, and the target detection data in the target trajectory data is updated according to the latest target detection data. If the number of times the target trajectory data is not associated with the target detection data since the parking state reaches the set second number threshold, it is determined that the target corresponding to the target trajectory data has entered the parking state, and the target current motion state and target detection data in the target trajectory data are updated, that is, the starting state in the target trajectory data is converted to the parking state, and the target detection data in the target trajectory data is updated according to the latest target detection data. If the number of times the target trajectory data is not associated with the target detection data since the parking state does not reach the set second number threshold, it is determined that the target corresponding to the target trajectory data is still in the starting state, and the target detection data in the target trajectory data is updated, that is, the target detection data in the target trajectory data is updated according to the latest target detection data.

[0080] A target in a parking state may be associated with the correct target detection data because the target is actually moving, or it may be erroneously associated with the detection data of a pedestrian crossing a zebra crossing. In order to solve the problem of erroneous association, it is necessary to obtain the first count threshold and the second count threshold through engineering debugging. During the engineering debugging process, the first count threshold and the second count threshold are repeatedly debugged according to the actual collected data until the target that is actually parked does not mistakenly enter the normal driving state due to erroneous association in most cases, but returns to the parking state, and the target that is actually moving again can quickly enter the normal driving state. In this embodiment, the first count threshold is set to 3, and the second count threshold is set to 4.

[0081] Step 3: Determine whether there is spatial overlap between two targets, and adjust the positions of the two targets that have spatial overlap.

[0082] In a specific embodiment of the present invention, determining whether two targets have spatial overlap specifically includes:

[0083] Step 3.1: Obtain road information; the road information includes the number of lanes, the width of each lane, and the horizontal offset of the radar installation position, where the horizontal offset of the radar installation position refers to the vertical distance from the radar installation position to the road sideline;

[0084] Step 3.2: Calculate the range of each lane in the radar coordinate system according to the number of lanes, the width of each lane, and the horizontal offset of the radar installation position;

[0085] Step 3.3: For a target in a normal driving state or a starting state, the target position is determined according to the radar echo signal of the current frame; for a target in a parking state, the target position is determined according to the position of the target when it enters the parking state (i.e., the position determined according to the radar echo signal when the target enters the parking state); specifically, the target distance and the target azimuth can be obtained according to the radar echo signal of the current frame, and the target position is determined according to the target distance and the target azimuth;

[0086] Step 3.4: Determine the lane information to which the target belongs according to the range of each lane and the target position;

[0087] Step 3.5: Determine whether the two targets are in the same lane based on the lane information of the targets;

[0088] Step 3.6: Determine the target type according to the current frame radar echo signal, and then determine the target length; specifically, the target length characteristic value can be determined according to the current frame radar echo signal, and the target type is determined according to the target length characteristic value, and then the target length is determined;

[0089] Step 3.7: For two targets in the same lane, determine whether there is spatial overlap based on the target positions and target lengths of the two targets; if so, adjust the positions of the two targets with spatial overlap; if not, do not adjust the positions of the two targets with spatial overlap.

[0090] In a specific implementation of the present invention, the range of the i-th lane is:

[0091] [-x0+(w1+…+w i-1 ),-x0+(w1+…+w i )](1)

[0092] Among them, x0 represents the horizontal offset of the radar installation position, w1, ..., w i-1 、w i They represent the widths of the 1st, ..., i-1th, and ith lanes respectively. When i = 1, w0 = 0. When the target horizontal coordinate is within the range of the i-th lane, it is determined that the target is in the i-th lane.

[0093] A rectangular coordinate system is constructed with the radar installation position as the origin (in this embodiment, the radar installation position is located at the midpoint of the road), the direction parallel to the road sideline and away from the radar as the positive direction of the Y axis, and the road cross-section as the X axis, as shown in FIG. Figure 3 , Figure 5 and Figure 6 As shown, the road has 4 lanes, each lane is 3.5m wide, and the horizontal offset of the radar installation position is 7m. From this, the range of each lane can be calculated, and then it can be determined which lane the target belongs to.

[0094] In a specific embodiment of the present invention, the positions of two targets with spatial overlap are adjusted according to the credibility of the target positions. The credibility relationship of the target positions of the two targets can be determined according to the current motion state of the targets and whether the target detection data is lost, as shown in Table 1. For targets A and B with spatial overlap, the front and rear position relationship of targets A and B can be determined according to the target positions of targets A and B. In this embodiment, it is assumed that target A is located in front of target B, the front refers to the direction close to the radar, and the rear refers to the direction away from the radar. In Table 1, R A represents the target position credibility of target A, R B represents the target position credibility of target B. The target in non-parking state means that the target is in normal driving state or starting state. Lost target detection data means that multiple consecutive frames are not associated with new target detection data. When R A =R B When R A >R B When , the target position of target A is more reliable, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap; R A <R B , the target position of target B is more reliable, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap.

[0095] Table 1 Determine the credibility of the target position according to the current motion state of the target and whether the target detection data is lost

[0096]

[0097] The present invention utilizes the characteristics of real objects that their length is constant and they cannot overlap, and under the premise of accurate target detection, uses a target position with higher credibility to infer the target position that spatially overlaps with it (i.e., there is a position conflict), so that the target position is more real and accurate.

[0098] Taking a traffic radar installed at an intersection in a certain city as an example, the parking queue and queue passage effects achieved by the present invention are as follows:

[0099] The radar obtains road information and target trajectory data of the current cycle. Figure 3As shown, the radar installation position is taken as the origin (in this embodiment, the radar installation position is located at the midpoint of the road), the direction parallel to the road sideline and away from the radar is taken as the positive direction of the Y-axis, and the road cross-section (such as the stop line) is taken as the X-axis. The horizontal axis represents the offset of each lane relative to the radar installation position, and the vertical axis represents the distance between each position and the radar installation position. The horizontal line at 65m on the vertical axis is the stop line, and the vertical line is the lane line. Four lanes are delineated in the interval from the stop line to 120m on the vertical axis, and each lane is 3.5m wide. In the 400th frame of target trajectory data, there are 5 tracking targets numbered 3, 1, 2, 7, and 4 from left to right. Among them, target No. 3 is in a parking state, and targets No. 1, 2, 7, and 4 are in a normal driving state. The specific legend is as follows Figure 4 shown.

[0100] In the 745th frame target trajectory data, Figure 5 As shown in the figure, targets 1, 2, and 7 have entered the parking state because their target speed is lower than the set speed threshold and no new target detection data is associated in the continuous frames; targets 14 and 19 have entered the parking state, but the distance between the parking position when entering the parking state and the stop line is greater than the actual stopping position (the position when targets 14 and 19 actually enter the stopping state). At this time, target 21 is in the same lane as targets 14 and 19, and is still in a normal driving state without losing target detection data.

[0101] In the 858th frame data, Figure 6 As shown, since targets 14, 19, and 21 overlapped in space during this period, the parking positions of targets 14 and 19 were adjusted forward to the correct positions; targets 5 and 12 were switched from the parking state to the starting state due to the re-association with the detection data; targets 1, 2, 3, etc. were switched from the starting state to the normal driving state.

[0102] from Figures 3 to 6 It can be seen that the method of the present invention can effectively cope with the intersection queuing scene and continuously track the entire process of the target from normal driving to stopping and then to normal driving.

[0103] Example 2

[0104] An embodiment of the present invention further provides an electronic device, which includes: a memory, a processor, and a computer program / instructions stored in the memory, and the processor executes the computer program / instructions to implement the target state detection method in the embodiment of the present application.

[0105] Although not shown, the electronic device includes a processor, which can perform various appropriate operations and processes according to the program and / or data stored in the read-only memory (ROM) or the program and / or data loaded from the storage portion into the random access memory (RAM). The processor can be a multi-core processor, or it can include multiple processors. In some embodiments, the processor can include a general main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. In RAM, various programs and data required for device operation are also stored. The processor, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.

[0106] The processor and memory are used together to execute the program / instructions stored in the memory. When the program / instructions are executed by the computer, the methods, steps or functions described in the above embodiments can be implemented.

[0107] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the target state detection method in the embodiment of the present application is implemented.

[0108] Readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0109] Although not shown, an embodiment of the present invention further provides a computer program product, including: a computer program / instruction, which implements the target state detection method in the embodiment of the present application when the computer program / instruction is executed by a processor.

[0110] What is disclosed above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, which should be covered within the protection scope of the present invention.

Claims

1. A target state detection method, characterized in that: The detection method comprises: Acquire target trajectory data; wherein the target trajectory data includes the current motion state of the target and target detection data obtained according to the radar echo signal of the current frame; the motion state includes a normal driving state, a parking state and a starting state, and the target detection data includes a target speed; Traverse all target trajectory data that are currently in a normal driving state. If the target speed in the target trajectory data is lower than a set speed threshold and is not associated with new target detection data in consecutive frames, determine that the target corresponding to the target trajectory data has entered a parking state, and update the current motion state of the target in the target trajectory data; otherwise, determine that the target corresponding to the target trajectory data has not entered a parking state, and update the target detection data in the target trajectory data; Traversing all target trajectory data currently in the parking state, if the target trajectory data is associated with new target detection data, determining that the target corresponding to the target trajectory data has entered the starting state, and updating the target current motion state and target detection data in the target trajectory data; otherwise, determining that the target corresponding to the target trajectory data is still in the parking state; All target trajectory data currently in the starting state are traversed. If the number of times that the target trajectory data is associated with the target detection data since the parking state reaches a set first number threshold, it is determined that the target corresponding to the target trajectory data enters the normal driving state, and the target current motion state and target detection data in the target trajectory data are updated; if the number of times that the target trajectory data is not associated with the target detection data since the parking state reaches a set second number threshold, it is determined that the target corresponding to the target trajectory data enters the parking state, and the target current motion state and target detection data in the target trajectory data are updated; if the number of times that the target trajectory data is not associated with the target detection data since the parking state does not reach the set second number threshold, it is determined that the target corresponding to the target trajectory data is still in the starting state, and the target detection data in the target trajectory data is updated.

2. The target state detection method according to claim 1, characterized in that: If the target speed in the target trajectory data is lower than the set speed threshold and no new target detection data is associated in consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated, specifically including: If the target speed in the target trajectory data is lower than a set first speed threshold and is not associated with new target detection data within at least N1 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated; If the target speed in the target trajectory data is lower than a set second speed threshold and no new target detection data is associated within at least N2 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated; If the target speed in the target trajectory data is lower than a set third speed threshold and is not associated with new target detection data within at least N3 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated; If the target speed in the target trajectory data is lower than a set fourth speed threshold and is not associated with new target detection data within at least N4 consecutive frames, it is determined that the target corresponding to the target trajectory data enters a parking state, and the current motion state of the target in the target trajectory data is updated; Among them, the first speed threshold < the second speed threshold < the third speed threshold < the fourth speed threshold, N1 < N2 < N3 < N4.

3. The target state detection method according to claim 2, characterized in that: The first speed threshold is 1 m / s, the second speed threshold is 3 m / s, the third speed threshold is 5 m / s, the fourth speed threshold is 8 m / s, N1 is 5, N2 is 10, N3 is 15, and N4 is 20.

4. The target state detection method according to claim 1, characterized in that: The first number threshold is set to 3, and the second number threshold is set to 4.

5. The target state detection method according to any one of claims 1 to 4, characterized in that: The detection method further includes determining whether two targets overlap in space, and adjusting the positions of the two targets that overlap in space, specifically including: Acquire road information; wherein the road information includes the number of lanes, the width of each lane, and the horizontal offset of the radar installation position, wherein the horizontal offset of the radar installation position refers to the vertical distance from the radar installation position to the road sideline; Calculating the range of each lane in the radar coordinate system according to the number of lanes, the width of each lane, and the horizontal offset of the radar installation position; For targets in normal driving or starting state, the target position is determined based on the radar echo signal of the current frame; for targets in parking state, the target position is determined based on the position of the target when it enters the parking state; Determine the lane information to which the target belongs according to the range of each lane and the target position; Determine whether two targets are in the same lane according to the lane information to which the targets belong; Determine the target type and then the target length according to the radar echo signal of the current frame; For two targets in the same lane, whether there is spatial overlap is determined based on the target positions and target lengths of the two targets. If so, the positions of the two targets with spatial overlap are adjusted; if not, the positions of the two targets with spatial overlap are not adjusted.

6. The target state detection method according to claim 5, characterized in that: The range of the i-th lane is: [-x0+(w1+…+w i-1 ),-x0+(w1+…+w i )]; Among them, x0 represents the horizontal offset of the radar installation position, w1, ..., w i-1 、w i They represent the width of the 1st, ..., i-1, and i lanes respectively. When i=1, w0=0.

7. The target state detection method according to claim 5, characterized in that: Adjust the positions of two targets that overlap in space, including: Determine the front-to-back positional relationship of the two targets according to the target positions of the two targets; assume that target A is located in front of target B, and the front refers to the direction close to the radar; When both target A and target B are in a non-parking state and target detection data is not lost, the target positions of target A and target B remain unchanged; When target B is in a non-parking state and has not lost target detection data, and target A is in a non-parking state and has lost target detection data, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap; When target B is in a non-parking state and has not lost target detection data, and target A is in a parking state, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap; When target B is in a non-parking state and loses target detection data, and target A is in a non-parking state and does not lose target detection data, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap; When both target A and target B are in a non-parking state and the target detection data is lost, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap; When target B is in a non-parking state and loses target detection data, and target A is in a parking state, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap; When target B is in a parking state, and target A is in a non-parking state and has not lost target detection data, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap; When target B is in a parking state, and target A is in a non-parking state and loses target detection data, the target position of target B remains unchanged, and the target position of target A is adjusted to the front of target B without spatial overlap; When both target A and target B are in the parking state, the target position of target A remains unchanged, and the target position of target B is adjusted to the rear of target A without spatial overlap.

8. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the target state detection method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the target state detection method according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the target state detection method according to any one of claims 1 to 7 is implemented.

Images