Data deduplication method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of vehicle infrastructure cooperation, and discloses a data deduplication method and device, computer equipment and a storage medium, and the method comprises the steps: receiving target data sensed by a plurality of pieces of sensing equipment; the target data comprises pose parameters of the target; according to the target data sensed by the plurality of sensing devices, judging whether the targets sensed by different sensing devices coincide or not; under the condition that the targets sensed by different sensing devices coincide, effective target data are extracted from the target data with the coincident targets; the effective target data comprises the target data with the highest priority in the target data with the overlapped targets; and determining the actual pose of the coincident target according to the effective target data. According to the invention, the effective target data with high precision is extracted according to the priority of the target data, the target data with poor precision provided by other sensing devices can be excluded, the influence of the target data with poor precision can be avoided, and the actual pose of the coincident target can be determined more accurately.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-road collaboration technology, and in particular to a method, device, computer equipment and storage medium for data deduplication. Background Art

[0002] With the rapid development of intelligent connected vehicles, perception sharing between vehicles and roads, and between vehicles is very important. Vehicle-road collaboration can improve road safety and efficiency, reduce traffic jams and accidents, and also improve the vehicle's automated control capabilities and driving comfort. In vehicle-road collaboration, multiple perception devices may recognize duplicate perception data of the same target, and duplicate perception data needs to be deduplicated.

[0003] The current data deduplication scheme generally assumes that these perception data are reliable, and uses duplicate perception data as a supplement to the same data. However, with the increasing number of intelligent connected vehicles and roadside perception devices, the perception accuracy of each perception device varies, resulting in poor reliability of perception data after deduplication. Summary of the invention

[0004] The purpose of the present invention is to provide a method, apparatus, computer device and storage medium for data deduplication, so as to solve the problem of poor reliability of perceived data after deduplication in the prior art.

[0005] In a first aspect, the present invention provides a method for deduplication of data, comprising:

[0006] Receiving target data respectively sensed by a plurality of sensing devices; the target data includes position and posture parameters of the target;

[0007] According to the target data sensed by the plurality of sensing devices, determining whether the targets sensed by different sensing devices overlap;

[0008] In the case where the targets sensed by different sensing devices overlap, valid target data is extracted from the target data of the overlapping targets; the valid target data includes the target data with the highest priority among the target data of the overlapping targets;

[0009] The actual position and posture of the overlapping target is determined according to the valid target data.

[0010] In some optional implementations, the target data further includes a timestamp;

[0011] The receiving target data respectively sensed by a plurality of sensing devices comprises:

[0012] Determining a delay threshold according to the speed of the receiving end receiving the target data; the delay threshold is negatively correlated with the speed of the receiving end;

[0013] Receive target data respectively sensed by a plurality of sensing devices, the target data having a timestamp that does not exceed the delay threshold.

[0014] In some optional implementations, the target data further includes a timestamp;

[0015] The step of extracting valid target data from target data with overlapping targets includes:

[0016] From the target data with overlapping targets, the target data with the highest priority and the latest timestamp is extracted as the valid target data.

[0017] In some optional embodiments, the pose parameters include the perceived position, velocity and heading angle of the target, and the target data also includes a timestamp;

[0018] The step of judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the plurality of sensing devices includes:

[0019] Determine, according to the speed of the target, the moving distance that the target can move from the timestamp to the current time;

[0020] Based on the perceived position of the target, increasing the moving distance along the heading angle to determine a predicted position of the target;

[0021] Determine whether the predicted positions of the targets perceived by different sensing devices overlap.

[0022] In some optional embodiments, the sensed location includes a sensed latitude and a sensed longitude;

[0023] The step of increasing the moving distance along the heading angle based on the perceived position of the target to determine the predicted position of the target includes:

[0024] Based on the perceived latitude of the target, adding the component of the moving distance in the latitude direction to determine the predicted latitude of the target;

[0025] Based on the perceived longitude of the target, the component of the moving distance in the longitude direction is added to determine the predicted longitude of the target.

[0026] In some optional implementations, the target data further includes the size of the target;

[0027] The step of judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the plurality of sensing devices includes:

[0028] Determine bounding boxes that can cover the target, respectively, based on the target data sensed by the plurality of sensing devices;

[0029] Determine whether the bounding boxes of the targets perceived by different sensing devices overlap.

[0030] In some optional embodiments, the bounding box is a directional bounding box;

[0031] The determining whether the bounding boxes of the targets sensed by different sensing devices overlap includes:

[0032] Taking the normal of one edge of the first oriented bounding box and the second oriented bounding box as the separation axis, cyclically performing the projection operation on the first oriented bounding box and the second oriented bounding box until it is determined that the first oriented bounding box does not overlap with the first oriented bounding box, or traversing all edges of the first oriented bounding box and the second oriented bounding box; the first oriented bounding box is the oriented bounding box of the target perceived by one perception device, and the second oriented bounding box is the oriented bounding box of the target perceived by another perception device;

[0033] The projection operation includes:

[0034] Projecting the first oriented bounding box and the second oriented bounding box onto the separation axis respectively;

[0035] Determine whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis;

[0036] In the case where the two projections do not overlap, determining that the first oriented bounding box does not overlap with the second oriented bounding box;

[0037] When the two projections overlap, the normal of one side of the first oriented bounding box and the second oriented bounding box that is not used as the separation axis is used as a new separation axis to perform the next projection operation.

[0038] In some optional implementations, determining whether a projection of the first oriented bounding box on the separation axis coincides with a projection of the second oriented bounding box on the separation axis includes:

[0039] Determine a center projection distance of a line segment between a center of the first oriented bounding box and a center of the second oriented bounding box on the separation axis;

[0040] Determine a first projection distance of the first oriented bounding box on the separation axis, and determine a second projection distance of the second oriented bounding box on the separation axis;

[0041] When the center projection distance is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections coincide with each other.

[0042] In some optional implementations, when the center projection distance is less than half of the sum of the first projection distance and the second projection distance, determining that the two projections coincide with each other includes:

[0043] When the sum of the central projection distance and a preset overlap error is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections overlap.

[0044] In some optional embodiments, the method further comprises:

[0045] Determining an error threshold according to the speed of the receiving end receiving the target data; the error threshold is positively correlated with the speed of the receiving end;

[0046] An error in the target position perceived by the perception device is determined, and a minimum value between the error in the target position and the error threshold is used as the overlap error.

[0047] In some optional implementations, determining the bounding boxes that can cover the target respectively according to the target data sensed by the plurality of sensing devices includes:

[0048] Determine axis-aligned bounding boxes of the targets respectively according to the target data sensed by the plurality of sensing devices;

[0049] In case the axis-aligned bounding boxes of the target perceived by at least some of the sensing devices coincide, an oriented bounding box of the target perceived by at least some of the sensing devices is determined.

[0050] In some optional implementations, judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the multiple sensing devices includes:

[0051] According to the lane lines, the sensing detection range is divided into multiple sub-areas;

[0052] According to the target data sensed by the plurality of sensing devices, it is determined whether the targets in the same sub-area sensed by different sensing devices overlap.

[0053] In some optional implementations, the receiving target data respectively sensed by a plurality of sensing devices includes:

[0054] Collect target data perceived by local sensing devices;

[0055] Receive at least one of target data in a perception sharing message format sent by other perception devices, target data in a basic safety message format sent by a perceived device, and target data in a roadside unit message format sent by a roadside device.

[0056] In a second aspect, the present invention provides a device for deduplication of data, comprising:

[0057] A receiving module, used to receive target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target;

[0058] A judgment module, used for judging whether the targets sensed by different sensing devices overlap according to the target data sensed by the plurality of sensing devices;

[0059] An extraction module is used to extract valid target data from the target data of the target overlap when the targets sensed by different sensing devices overlap; the valid target data includes the target data with the highest priority among the target data of the target overlap;

[0060] A processing module is used to determine the actual position and posture of the overlapping target according to the valid target data.

[0061] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the data deduplication method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0062] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the data deduplication method of the first aspect or any corresponding embodiment thereof.

[0063] Beneficial effects of the present invention:

[0064] (1) When targets perceived by different sensing devices overlap, the present invention extracts effective target data with higher accuracy according to the priority of the target data, filters the uneven target data, and excludes target data with lower accuracy provided by other sensing devices, thereby avoiding the influence of target data with lower accuracy, achieving data deduplication, and more accurately determining the actual position and posture of the overlapping targets based on the effective target data;

[0065] (2) The target data sensed by the sensing device is unified from its timestamp to the current time. Based on the predicted position of the target at the current time, it is determined whether the target positions sensed by different sensing devices overlap. This can eliminate the interference caused by different timestamps and ensure the accuracy of overlap detection.

[0066] (3) Using bounding boxes instead of targets can easily determine whether targets overlap. First, a preliminary judgment is made based on the axis-aligned bounding boxes to exclude non-overlapping targets. The calculation is simple and can speed up the data deduplication process. Then, the directional bounding boxes of the targets that are preliminarily determined to overlap are determined to accurately determine whether the targets with these axis-aligned bounding boxes overlap, making the final deduplication result more accurate. Based on the target data, the center projection distance of the centers of the two directional bounding boxes on the separation axis and the projection distance of each directional bounding box on the separation axis can be quickly determined. Based on the size relationship between the three projection distances, it is possible to simply and quickly determine whether the projections of the directional bounding boxes on the separation axis overlap. This does not require a complex calculation process and has high processing efficiency. It is suitable for scenarios with high timeliness requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 It is a schematic diagram of a vehicle-road collaboration scenario;

[0069] Figure 2 is a flowchart of a method for deduplication of data according to an embodiment of the present invention;

[0070] Figure 3 is a schematic diagram of deduplication of a target sensed by multiple sensing devices according to an embodiment of the present invention;

[0071] Figure 4 is a flow chart of another method for data deduplication according to an embodiment of the present invention;

[0072] Figure 5 is a flow chart of another method for deduplication of data according to an embodiment of the present invention;

[0073] Figure 6 is a schematic diagram of determining whether targets overlap based on oriented bounding boxes according to an embodiment of the present invention;

[0074] Figure 7 is a schematic diagram of determining whether objects overlap based on axis-aligned bounding boxes according to an embodiment of the present invention;

[0075] Figure 8 is a schematic diagram of dividing sub-areas according to lane lines according to an embodiment of the present invention;

[0076] Fig. 9is a structural block diagram of a data deduplication device according to an embodiment of the present invention;

[0077] Fig.10 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0078] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0079] A scenario of vehicle-road collaboration can be seen in Figure 1 As shown. Figure 1 In the scenario shown, multiple sensing devices are involved: local vehicle 110, sensed vehicle 120, other vehicles 130, and roadside equipment 140. Figure 1 As shown, the sensed vehicle 120 is taken as the sensed target, and the sensed vehicle 120 is located in front of the local vehicle 110 and in the sensed detection range 111 of the onboard radar and camera of the local vehicle 110. The local vehicle 110 can sense the target data of the sensed vehicle 120, and the target data may include the position and speed of the sensed vehicle 120. For the convenience of distinction, the target data sensed by the local vehicle 110 is referred to as target data A. The sensed vehicle 120 itself can collect its own position and speed, so the sensed vehicle 120 can also generate target data B related to itself.

[0080] In addition, if Figure 1 As shown in FIG. 1 , the sensed vehicle 120 is located in the sensing detection range 131 of other vehicles 130, so other vehicles 130 can also sense the target data C of the sensed vehicle 120. In addition, the roadside device 140 can sense the data of vehicles and pedestrians in a larger range around it, such as Figure 1 As shown, the sensed vehicle 120 is located in the sensing detection range 141 of the roadside device 140, so the roadside device 140 can also sense the target data D of the sensed vehicle 120. It can be understood that the sensing device can sense multiple targets, and for each sensed target, corresponding target data can be generated; for example, Figure 1As shown, the roadside device 140 can sense the local vehicle 110, the sensed vehicle 120, and the other vehicles 130. All three are targets that the roadside device 140 can sense. The roadside device 140 can generate target data for the local vehicle 110, target data for the sensed vehicle 120 (i.e., the target data D mentioned above), and target data for the other vehicles 130.

[0081] Among them, the sensed vehicle 120, other vehicles 130, and roadside equipment 140 can all send their sensed target data to the local vehicle 110; for example, the three can broadcast the corresponding target data by broadcasting or other means, so that the local vehicle 110 can obtain the target data B sensed by the sensed vehicle 120, the target data C sensed by other vehicles 130, and the target data D sensed by the roadside equipment 140, and determine the position, speed, etc. of the sensed vehicle 120 by combining the target data A, target data B, target data C, and target data D.

[0082] In traditional data deduplication schemes, it is generally believed that the above target data A, target data B, target data C, and target data D are all reliable, but in fact the perception accuracy of different sensing devices is different; for example, with the development of technologies such as laser radar, the laser radar on the local vehicle 110 can more accurately determine the position and speed of the sensed vehicle 120, but the processing capacity of the roadside device 140 is generally prioritized, and the accuracy of the position and speed of the target it senses is relatively low, that is, the accuracy of the above target data A is generally greater than the accuracy of the target data D. The data deduplication method provided in the embodiment of the present invention selects appropriate data based on priority to achieve data deduplication.

[0083] According to an embodiment of the present invention, a method embodiment of data deduplication is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0084] In this embodiment, a method for deduplication of data is provided, which can be used for sensing devices, such as the local vehicle 110, roadside device 140, etc., and can also be used for servers. Figure 1 is a flow chart of a method for data deduplication according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps.

[0085] Step S201, receiving target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target.

[0086] In this embodiment, in the vehicle-road collaboration scenario, there are multiple sensing devices, and each sensing device can sense the target within its sensing detection range to determine the sensed target data. Figure 1 As shown, the local vehicle 110, the sensed vehicle 120, the other vehicles 130, and the roadside equipment 140 can all be sensed; if the local vehicle 110 is used to execute the method provided in this embodiment, for example, the controller of the local vehicle 110 executes the method provided in this embodiment, if the sensed vehicle 120 is the sensed target, then the local vehicle 110 can receive the target data B, target data C, and target data D sensed by the sensed vehicle 120, the other vehicles 130, and the roadside equipment 140, respectively.

[0087] The target data includes the position and posture parameters of the target, which can represent the position of the target. Furthermore, the position and posture parameters can also represent the orientation angle, speed, etc. of the target.

[0088] Step S202: judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the multiple sensing devices.

[0089] In this embodiment, for the same target, when multiple sensing devices sense the target data of the target, due to different precisions of the sensing devices, time delays, etc., the target data of the same target sensed by different sensing devices may be different. Before deduplication of target data, this embodiment needs to first determine which target data is overlapped, that is, it needs to determine whether the targets sensed by different sensing devices are overlapped.

[0090] Specifically, it is possible to determine whether the targets perceived by different sensing devices overlap based on the positions of the targets represented by the target data. For example, a coincidence threshold (e.g., 2 m) may be pre-set. If the distance difference between the positions of the targets determined by the two sensing devices is less than the coincidence threshold, it means that the two targets are relatively close, that is, the two sensing devices perceive the same target and the targets perceived by the two sensing devices overlap.

[0091] Step S203, when the targets sensed by different sensing devices overlap, valid target data is extracted from the target data of the overlapped targets; the valid target data includes the target data with the highest priority among the target data of the overlapped targets.

[0092] In this embodiment, if the targets sensed by multiple different sensing devices overlap, the priority of each target data of the target overlap can be determined, and appropriate valid target data can be selected from them. The valid target data at least includes the target data with the highest priority. For example, the target data with the highest priority among the target data of the target overlap can be used as the valid target data.

[0093] For example, Figure 3 The figure shows a schematic diagram of the location of a target sensed by multiple sensing devices. Figure 3 As shown, based on the target data A sensed by the local vehicle 110, the first target 301 is determined; based on the target data B sensed by the sensed vehicle 120, the second target 302 is determined; based on the target data C sensed by the other vehicle 130, the third target 303 is determined; based on the target data D sensed by the roadside device 140, the fourth target 304 is determined; the relative positions of the first target 301, the second target 302, the third target 303, and the fourth target 304 can be seen in Figure 3 As shown. Figure 3 It can be seen that if the four targets sensed by the four sensing devices overlap with each other, valid target data can be selected from target data A, target data B, target data C, and target data D; for example, if target data B has the highest priority, target data B can be used as valid target data.

[0094] It can be understood that if a target perceived by a certain sensing device does not overlap with other targets, the target can be regarded as an independent target. For example, if the fourth target 304 does not overlap with the first target 301, the second target 302, and the third target 303, the fourth target 304 is regarded as an independent target perceived; accordingly, at this time, the valid target data is selected from the target data A, the target data B, and the target data C.

[0095] Among them, the priority of the target data is positively correlated with the accuracy of the target data, that is, the higher the accuracy of the target data, the greater the priority of the target data. For example, the target data fed back by the sensing device may also include the sensing accuracy, and the priority of the target data is determined based on the sensing accuracy; or, the priority of the target data may be determined based on the sensing source of the target data, that is, based on which sensing device senses the target data, and this embodiment does not limit this.

[0096] Step S204, determining the actual position and posture of the overlapping target according to the valid target data.

[0097] In this embodiment, the target data with lower priority among the target data of target overlap is not used in the process of determining the position and posture of the target, that is, the target data with lower priority among the target data of target overlap can be discarded; wherein, for the convenience of description, if the targets perceived by different sensing devices overlap, it can be explained that the different sensing devices perceive the same target, and the same target is called an overlapped target. Based on the extracted valid target data, the actual position and posture of the overlapped target, that is, the actual position, etc. can be determined.

[0098] Generally, there can be only one valid target data, which is the target data with the highest priority. For example, if target data B has the highest priority, then target data B can be used as the valid target data; see Figure 3 As shown, the position and posture of the second target 302 determined based on the target data B can be used as the actual position and posture finally determined for the overlapping target.

[0099] The data deduplication method provided in this embodiment, when the targets sensed by different sensing devices overlap, extracts the valid target data with higher accuracy according to the priority of the target data, filters the uneven target data, and can exclude the target data with poor accuracy provided by other sensing devices, thereby avoiding the influence of the target data with poor accuracy and achieving data deduplication. Moreover, based on the valid target data, the actual posture of the overlapping targets can be determined more accurately.

[0100] In this embodiment, a method for deduplication of data is provided, which can be used in local vehicles 110, roadside equipment 140, etc., and can also be used in servers. Figure 4 is a flow chart of a method for data deduplication according to an embodiment of the present invention. Figure 4 As shown, the process includes the following steps.

[0101] Step S401, receiving target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target.

[0102] For more details, please refer to Figure 2 Step S201 of the illustrated embodiment will not be described in detail here.

[0103] In some optional implementations, the target data may include a timestamp in addition to the position and posture parameters. The timestamp indicates the time when the sensing device senses the target data; for example, the timestamp may specifically be the time when the sensing device sends the target data.

[0104] Furthermore, the process of the above step S401 "receiving target data respectively sensed by a plurality of sensing devices" may specifically include the following steps A1 to A2.

[0105] Step A1, determining a delay threshold according to the speed of the receiving end receiving the target data; there is a negative correlation between the delay threshold and the speed of the receiving end.

[0106] Step A2, receiving target data respectively sensed by multiple sensing devices, the timestamp of which does not exceed the delay threshold.

[0107] In this embodiment, a delay threshold for indicating the allowed delay size is predetermined. If the timestamp of the target data sensed by the sensing device does not exceed the delay threshold, it means that the delay of the target data is within a reasonable range and meets the requirements, and the subsequent process of determining the actual position of the target can be performed based on the target data; on the contrary, if the timestamp of the target data sensed by the sensing device exceeds the delay threshold, it means that the delay of the target data is large and the target data is directly discarded.

[0108] The delay threshold may be a time difference. When determining whether the timestamp exceeds the delay threshold, the time difference between the current time and the timestamp may be determined, and the time difference may represent the delay of the target data; the time difference is compared with the delay threshold, and if the time difference is greater than the delay threshold, it indicates that the timestamp exceeds the delay threshold.

[0109] In this embodiment, the receiving end that receives the target data is also the executing end of the data deduplication method of this embodiment. When the receiving end itself has a certain speed, the delay threshold is determined based on the speed of the receiving end. In addition, there is a negative correlation between the delay threshold and the speed of the receiving end, that is, the greater the speed of the receiving end, the smaller the delay threshold; specifically, the greater the speed of the receiving end, the higher the requirement of the receiving end for the target data delay, and the target data with a high delay is prone to misjudgment. Therefore, setting the delay threshold to a smaller value can ensure that the receiving end can determine stable and reliable target data when moving at high speed.

[0110] For example, if Figure 1 The local vehicle 110 shown executes the method. During the driving process of the local vehicle 110, it has a certain speed. The delay threshold is determined based on its own speed. The faster the speed of the local vehicle 110, the smaller the allowed delay. For example, if the speed of the local vehicle 110 is less than 60 km / h, the delay threshold can be set to 300 ms; if the speed of the local vehicle 110 is greater than 60 km / h, the delay threshold can be set to 200 ms.

[0111] Step S402: judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the multiple sensing devices.

[0112] In this embodiment, the target data may include a timestamp in addition to the posture parameters; and the posture parameters specifically include the perceived position, speed and heading angle of the target. It can be understood that the posture parameters are the parameters perceived by the perception device at the timestamp; for example, the perceived position is the location of the target at the timestamp.

[0113] The above-mentioned step S402 "determining whether the targets sensed by different sensing devices overlap based on the target data sensed by multiple sensing devices" may specifically include the following steps S4021 to S4024.

[0114] Step S4021, determine whether it is necessary to perform time synchronization on the target data sensed by different sensing devices; if time synchronization is required, execute step S4022.

[0115] In this embodiment, since different sensing devices may sense at different times, the target data sent by different sensing devices may have different timestamps; in order to accurately determine whether the targets corresponding to different target data overlap, it is necessary to synchronize these target data, that is, it is necessary to unify the targets corresponding to these target data to the same time. It is understandable that if the timestamps of different target data are the same, time synchronization is not required.

[0116] Step S4022, determining the moving distance that the target can move from the timestamp to the current time according to the speed of the target.

[0117] In this embodiment, the target data sensed by all sensing devices are unified to the current time. In addition, the moving distance that the target can move from the timestamp to the current time is determined, and the predicted position of the target at the current time is predicted based on the moving distance.

[0118] Among them, since the time period from the timestamp to the current time is short, for example, the time period between the two can be less than the above-mentioned delay threshold; in a short time period, it can be considered that the target maintains a uniform speed and the moving direction remains unchanged, so the moving distance that the target can move from the timestamp to the current time is the distance that the target moves at a uniform speed. For example, the speed of the target is speed, if the time period between the timestamp and the current time is time_offset, then the moving distance Distance is: Distance = speed*time_offset.

[0119] Step S4023, based on the perceived position of the target, the moving distance is increased along the heading angle to determine the predicted position of the target.

[0120] In this embodiment, from the timestamp of the target data to the current time, it can be considered that the moving direction of the target remains unchanged, that is, the heading angle remains unchanged, so based on the perceived position of the target at the timestamp plus the corresponding moving distance, the position of the target after uniformly moving along the heading angle to the current time can be determined, that is, the predicted position. For target data perceived by different sensing devices, the predicted positions of the targets perceived by different sensing devices at the current time can be determined in accordance with the above steps S4021 to S4022.

[0121] Optionally, when the sensing device is a vehicle, the sensed position collected by it is generally expressed in the form of longitude and latitude, that is, the sensed position includes sensed latitude and sensed longitude. In this case, the above step S4022 "based on the sensed position of the target, adding the moving distance along the heading angle to determine the predicted position of the target" can specifically include the following steps B1 to B2.

[0122] Step B1, based on the perceived latitude of the target, the component of the moving distance in the latitude direction is added to determine the predicted latitude of the target.

[0123] Step B2, based on the perceived longitude of the target, the component of the moving distance in the longitude direction is added to determine the predicted longitude of the target.

[0124] In this embodiment, after determining the moving distance, the component of the moving distance in the latitude direction and the component of the longitude direction can be determined based on the heading angle, and then the perceived latitude at the timestamp can be updated to the predicted latitude at the current time, and the perceived longitude at the timestamp can be updated to the predicted longitude at the current time. For example, if the perceived latitude is Lat_before, the perceived longitude is Long_before, and the heading angle of the target is head, the predicted latitude Lat_after and predicted longitude Long_after of the target are respectively:

[0125] Lat_after=Lat_before+Distance*cos(head)*Parameter_lat;

[0126] Long_after=Long_before+Distance*sin(head)*Parameter_long.

[0127] Parameter_lat represents a conversion factor for converting distance into latitude, for example, a conversion factor for converting 1 meter into 1 degree of latitude; Parameter_long represents a conversion factor for converting distance into longitude, for example, a conversion factor for converting 1 meter into 1 degree of longitude.

[0128] Step S4024, determining whether the predicted positions of the targets sensed by different sensing devices overlap.

[0129] In this embodiment, when judging whether the targets perceived by different sensing devices overlap, the judgment is made based on the predicted position of the target, that is, based on the unified position of the target at the current time, so as to avoid the problem of incorrect judgment due to different timestamps.

[0130] Step S403, when the targets sensed by different sensing devices overlap, valid target data is extracted from the target data of the overlapped targets; the valid target data includes the target data with the highest priority among the target data of the overlapped targets.

[0131] For details, please refer to Figure 2 Step S203 of the illustrated embodiment will not be described in detail here.

[0132] In some optional embodiments, when the target data also includes a timestamp, the step S403 "extracting valid target data from target data with overlapping targets" may include: extracting target data with the highest priority and the latest timestamp as valid target data from target data with overlapping targets.

[0133] In this embodiment, for multiple target data with overlapping targets, the higher the priority of the target data, the higher the accuracy of the target data; and the newer the timestamp of the target data, the smaller the delay of the target data and the smaller the error. Therefore, the target data with the highest priority and the latest timestamp can be used as the valid target data.

[0134] Specifically, for multiple target data with overlapping targets, the target data with the highest priority can be used as the valid target data; if the priorities are the same, that is, there are multiple target data with the highest priority, the target data with the latest timestamp (that is, the smallest delay) can be retained and used as the valid target data. The remaining target data with lower priority and larger delay can be discarded.

[0135] Otherwise, the priority of the target perception source is compared. The priority can be set by obtaining the confidence of the perception source, and the target with higher priority is retained. If the priority is the same, the delay is compared, and the target with lower delay is retained. The target with lower priority and lower delay is discarded.

[0136] Step S404, determining the actual position and posture of the overlapping target according to the valid target data.

[0137] For details, please refer to Figure 2 Step S204 of the illustrated embodiment will not be described in detail here.

[0138] This embodiment provides a method for deduplication of data, which unifies the target data sensed by the sensing device from its timestamp to the current time, and judges whether the target positions sensed by different sensing devices overlap based on the predicted position of the target at the current time. This can eliminate the interference caused by different timestamps and ensure the accuracy of overlap detection.

[0139] In this embodiment, a method for deduplication of data is provided, which can be used in local vehicles 110, roadside equipment 140, etc., and can also be used in servers. Figure 4is a flow chart of a method for data deduplication according to an embodiment of the present invention. Figure 5 As shown, the process includes the following steps.

[0140] Step S501, receiving target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target.

[0141] For more details, please refer to Figure 2 Step S201 of the illustrated embodiment will not be described in detail here.

[0142] In some optional implementations, the above step S501 "receiving target data respectively sensed by multiple sensing devices" may specifically include the following steps C1 to C2.

[0143] Step C1, collecting target data sensed by the local sensing device.

[0144] Step C2, receiving at least one of target data in a perception sharing message format sent by other perception devices, target data in a basic safety message format sent by a perceived device, and target data in a roadside unit message format sent by a roadside device.

[0145] In this embodiment, the method may be executed by a certain sensing device. For the convenience of distinction, the sensing device executing the method is referred to as a local sensing device. For example, the local sensing device may be Figure 1 The local vehicle 110 shown. Among them, the local sensing device can use its own vehicle-mounted radar, camera, etc. to collect target data of targets in the surrounding environment, and receive target data sensed by other sensing devices. At present, sensing devices generally share data according to their own message formats. In particular, when the sensing device is an intelligent networked vehicle, different car manufacturers may transmit target data according to their own developed communication protocols, which makes it difficult to share target data between sensing devices and is costly. In this embodiment, V2X standard messages are used to transmit target data between sensing devices, which facilitates data sharing between sensing devices and is applicable to the industry standard of intelligent networked vehicles.

[0146] Specifically, after other sensing devices sense the target data of a certain target, they can share the sensed data by sending a sensor sharing message (SSM). By sending the target data in the format of a sensor sharing message to the local sensing device, the sensed target data can be shared with the local sensing device. As a target that can be sensed by other sensing devices, the sensed device itself can collect its own target data and sense the target data by sending a basic safety message (BSM). By sending the target data in the format of a basic safety message to the local sensing device, the sensed target data can also be shared with the local sensing device. Similarly, the roadside unit can sense the target data by sending a road test unit message (RSM). The roadside device sends the target data in the format of a road side unit message and can share the sensed target data with the local sensing device.

[0147] For example, the vehicle has a vehicle-side perception module, a vehicle-side communication module (OBU, On Board Unit) and a vehicle-side computing module. The vehicle-side perception module may include cameras, lidars, millimeter-wave radars, etc., which can collect data from other vehicles. The vehicle's CAN (Controller Area Network) and GNSS (Global Navigation Satellite System) positioning are used to collect data from the vehicle itself; the vehicle-side communication module can share perception through BSM messages and SSM messages. Figure 1 Taking the shown scenario as an example, the sensed vehicle 120 can send a BSM message to the local vehicle 110 and send the target data B collected by itself to the local vehicle 110; other vehicles 130 can send an SSM message to the local vehicle 110 and send the target data C collected by itself to the local vehicle 110.

[0148] The roadside equipment includes a roadside perception module, a roadside communication module (RSU) and a roadside computing module. The roadside perception module can include cameras and laser radars, etc., which can collect data from vehicles within the range, and perform perception fusion through the roadside computing module to generate target data; the roadside communication module can share perception through RSM messages. For example, Figure 1 Taking the illustrated scenario as an example, the roadside device 140 sends an RSM message to the local vehicle 110 , and sends the target data D collected by itself to the local vehicle 110 .

[0149] Step S502: judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the multiple sensing devices.

[0150] When judging whether the targets perceived by different sensing devices overlap, it is possible to judge whether the distance between the two targets is less than the overlap threshold. If the distance between the two targets is less than the overlap threshold, it means that the two targets are relatively close and overlap. However, this method is equivalent to representing the target with a circle. When judging whether targets of other shapes (such as square vehicles) overlap, it is easy to cause the problem of insufficient judgment accuracy. In this embodiment, the target's bounding box is used to determine whether the target overlaps.

[0151] Specifically, the target data sent by the sensing device also includes the size of the target; the information contained in the size may be based on the type or shape of the target. For example, if the target is a square, the size of the target may include the length and width; or, if the target is a triangle, the size of the target may include the base and height, etc. For example, if the target is a vehicle, the vehicle may be represented by a rectangle, and the target data may include the length and width of the vehicle. Furthermore, the above step S502 "determining whether the targets sensed by different sensing devices overlap based on the target data sensed by multiple sensing devices" may include the following steps S5021 to S5022.

[0152] Step S5021, determining bounding boxes that can cover the target based on the target data sensed by multiple sensing devices.

[0153] In this embodiment, for each target perceived by the sensing device, a bounding box that can cover the target can be determined based on the position and size of the target, and the position and size of the target are represented based on the bounding box. For example, the bounding box can be an oriented bounding box (OBB).

[0154] Step S5022, determine whether the bounding boxes of the targets sensed by different sensing devices overlap.

[0155] In this embodiment, the target is represented by a bounding box. If the bounding boxes of the targets sensed by different sensing devices overlap, it can be indicated that the targets sensed by the sensing devices themselves also overlap.

[0156] In some optional implementations, the bounding box is a directional bounding box, based on which the position and shape of the target are accurately represented. In addition, this embodiment uses a separation axis to determine whether the bounding boxes overlap. Specifically, the above step S5022 "determine whether the bounding boxes of the targets perceived by different sensing devices overlap" may include the following step D1.

[0157] Step D1, using the normal of one edge in the first oriented bounding box and the second oriented bounding box as the separation axis, cyclically performing the projection operation on the first oriented bounding box and the second oriented bounding box until it is determined that the first oriented bounding box does not overlap with the first oriented bounding box, or traversing all edges of the first oriented bounding box and the second oriented bounding box; the first oriented bounding box is the oriented bounding box of the target perceived by one perception device, and the second oriented bounding box is the oriented bounding box of the target perceived by another perception device.

[0158] In this embodiment, multiple bounding boxes can be compared in pairs to determine the overlap between these bounding boxes. For the convenience of description, the oriented bounding box of the target perceived by one perception device is called the first oriented bounding box, and the oriented bounding box of the target perceived by another perception device is called the second oriented bounding box, and it is determined whether the first oriented bounding box overlaps with the first oriented bounding box. Wherein, the first oriented bounding box and the second oriented bounding box both have multiple edges, and the normal of each edge is used as a separation axis, and a projection operation is performed to determine whether the projections of the first oriented bounding box and the second oriented bounding box on the separation axis overlap. If the projections of the first oriented bounding box and the second oriented bounding box on a certain separation axis do not overlap, it can be said that the first oriented bounding box and the second oriented bounding box do not overlap; if after traversing all the edges, it is determined that the projections of the first oriented bounding box and the second oriented bounding box on any separation axis overlap, it can be said that the first oriented bounding box and the second oriented bounding box overlap.

[0159] The above-mentioned projection operation specifically includes the following steps D11 to D14.

[0160] Step D11: project the first oriented bounding box and the second oriented bounding box onto the separation axis respectively.

[0161] See also Figure 6 As shown, the shape of the target is represented by a rectangle, 601 represents the target perceived by a certain sensing device, 602 represents the target perceived by another sensing device, and the directional bounding box of the target 601 is the first directional bounding box 603, and the directional bounding box of the target 602 is the second directional bounding box 604. The directional bounding box is consistent with the outer contour of the target, so that the target and the directional bounding box can be distinguished. Figure 6 There is a certain gap between the oriented bounding box in and the target.

[0162] Since the first oriented bounding box 603 and the second oriented bounding box 604 both have two parallel sides, it is only necessary to use the normals of the long side and the short side of the first oriented bounding box 603 and the second oriented bounding box 604 as separation axes. That is, the normals of the long side and the short side of the first oriented bounding box 603 can both be used as separation axes, and the normals of the long side and the short side of the second oriented bounding box 606 can also be used as separation axes, for a total of four separation axes. Figure 6, the separation axis 1 corresponding to the short side normal of the first oriented bounding box 603 and the separation axis 2 corresponding to the long side normal of the first oriented bounding box 603 are shown.

[0163] During each projection operation, the first oriented bounding box and the second oriented bounding box are projected onto a separation axis. For example, if the separation axis selected for this projection operation is separation axis 1, the first oriented bounding box 603 and the second oriented bounding box 604 can be projected onto separation axis 1.

[0164] Step D12, determining whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis.

[0165] In this embodiment, the projection of the oriented bounding box on the separation axis is a line segment. Based on the projection position of the oriented bounding box on the separation axis, it can be determined whether the projections of two oriented bounding boxes (i.e., the first oriented bounding box and the second oriented bounding box) on the separation axis overlap.

[0166] For example, the projection coordinates of the oriented bounding box on the separation axis can be determined, and whether the projections overlap can be determined based on the coordinate size relationship of the two projections; however, this method has a large amount of calculation. Optionally, the above step D12 "determining whether the projection of the first oriented bounding box on the separation axis overlaps with the projection of the second oriented bounding box on the separation axis" can include the following steps D121 to D123.

[0167] Step D121, determining the center projection distance of the line segment between the center of the first oriented bounding box and the center of the second oriented bounding box on the separation axis.

[0168] In this embodiment, based on the target data provided by the sensing device, the position of the target can be determined, and the position can be used as the center of the oriented bounding box; for example, the position is a predicted position obtained based on the sensed position prediction, and the predicted position is used as the center of the corresponding oriented bounding box. In addition, the target data includes the orientation angle of the target, and combined with the orientation angle, the projection distance between the centers of the two oriented bounding boxes on the separation axis can be determined, that is, the center projection distance. In this embodiment, L0 represents the center projection distance.

[0169] For example, see Figure 6 As shown, the center coordinates of the first oriented bounding box 603 are (x1, y1), and its orientation angle is head1; the center coordinates of the second oriented bounding box 604 are (x2, y2), and its orientation angle is head2; wherein the center coordinates and orientation angle are coordinates and relative orientation angles in an XY coordinate system, and the XY coordinate system may be a coordinate system determined by a device that executes the method (such as a local sensing device, etc.).

[0170] For example, if the current separation axis is separation axis 1, the center projection distance L0 of the centers of the two oriented bounding boxes on the separation axis can be found in Figure 6 And, the center projection distance L0 may specifically be: L0 =|(x1-x2)*cos(head1)+(y1-y2)*sin(head1)|.

[0171] Step D122, determining a first projection distance of the first oriented bounding box on the separation axis, and determining a second projection distance of the second oriented bounding box on the separation axis.

[0172] In this embodiment, the projection distances of the two oriented bounding boxes on the separation axis are determined respectively, and the projection distance of the first oriented bounding box on the separation axis is called the first projection distance, and the projection distance of the second oriented bounding box on the separation axis is called the second projection distance.

[0173] The projection distance of the oriented bounding box on the separation axis can be determined based on the length and width of the oriented bounding box and the relative angle between the oriented bounding box and the separation axis. When the normal of one side of a certain oriented bounding box is used as the separation axis, the projection distance of the oriented bounding box on the separation axis is only related to the length or width of the oriented bounding box.

[0174] For example, Figure 6 As shown, the current separation axis is separation axis 1 , which is the short side normal of the first oriented bounding box 603 , so the first projection distance of the first oriented bounding box 603 on the separation axis 1 is the length of the first oriented bounding box 603 . Figure 6 L1 represents half of the first projection distance. In other words, the first projection distance is 2*L1.

[0175] Furthermore, the second projection distance of the second oriented bounding box 604 on the separation axis 1 includes the projection distance of the long side of the second oriented bounding box 604 on the separation axis 1 and the projection distance of the short side on the separation axis 1. Specifically, if the length of the second oriented bounding box 604 is length2 and the width is width2, the projection distance L_length2 of the long side and the projection distance L_width2 of the short side on the separation axis 1 are respectively:

[0176] L_length2=|length2*cos(head2)*cos(head1)+length2*sin(head2)*sin(head1)|;

[0177] L_width2=|width2*cos(head2+π / 2)*cos(head1)+width2*sin(head2+π / 2)*sin(head1)|.

[0178] The second projected distance of the second oriented bounding box 604 on the separation axis 1 is the sum of the projected distance L_length2 and the projected distance L_width2, that is, L_width2 + L_length2. Figure 6 Let L2 represent half of this second projected distance, that is, L2 = (L_width2 + L_length2) / 2.

[0179] Step D123: When the central projected distance is less than half of the sum of the first projected distance and the second projected distance, it is determined that the two projections coincide.

[0180] In this embodiment, if the central projected distance is less than half of the sum of the first projected distance and the second projected distance, it can be shown that the projections of the two oriented bounding boxes on the separation axis coincide; conversely, if the central projected distance is greater than half of the sum of the first projected distance and the second projected distance, it can be shown that the projections of the two oriented bounding boxes on the separation axis do not coincide. As Figure 6 shown, if L0 < L1 + L2, it can be determined that the two projections coincide; Figure 6 Taking the non - coincidence of the two projections as an example. Among them, if L0 = L1 + L2, it can be determined that the projections coincide, or it can be determined that the projections do not coincide, which can be specifically determined based on the actual situation.

[0181] In this embodiment, based on the target data, the central coordinates and the orientation angle of the oriented bounding box can be conveniently determined. Furthermore, the central projected distance of the centers of the two oriented bounding boxes on the separation axis, and the projected distance of each oriented bounding box on the separation axis can be simply and quickly determined. Based on the magnitude relationship among the three projected distances (central projected distance, first projected distance, second projected distance), it is possible to simply and quickly determine whether the projections on the separation axis coincide, without a complex calculation process, with high processing efficiency, and is suitable for scenarios with high requirements for timeliness.

[0182] In some alternative embodiments, the above step D123 "When the central projected distance is less than half of the sum of the first projected distance and the second projected distance, it is determined that the two projections coincide" may specifically include: When the sum of the central projected distance and a preset overlap error is less than half of the sum of the first projected distance and the second projected distance, it is determined that the two projections coincide.

[0183] In this embodiment, when determining whether two projections on the separation axis coincide, an overlap error is introduced to allow for errors in the vehicle's positioning accuracy. For example, still using L0 to represent the central projection distance, L1 to represent half of the first projection distance, and L2 to represent half of the second projection distance, if the overlap error is posion_error, when the sum of the central projection distance and the preset overlap error is less than half of the sum of the first projection distance and the second projection distance, that is, when L0 + posion_error < L1 + L2, it can be considered that the two projections on the separation axis coincide. If L0 + posion_error > L1 + L2, it can be considered that the two projections on the separation axis do not coincide, and at this time, it can be further determined that the two oriented bounding boxes do not coincide.

[0184] Optionally, the method further includes a process of determining the overlap error, and this process may specifically include the following steps E1 to E2.

[0185] Step E1, determine an error threshold according to the speed of the receiving end that receives the target data; there is a positive correlation between the error threshold and the speed of the receiving end.

[0186] Step E2, determine the error of the sensing device in sensing the target position, and use the minimum value of the error of the target position and the error threshold as the overlap error.

[0187] In this embodiment, the receiving end that receives the target data is also the executing end that executes the method for deduplicating data in this embodiment. When the receiving end itself has a certain speed, the error threshold is determined based on the speed of the receiving end. For example, based on the speed of the receiving end, the current driving scenario can be determined, and an appropriate error threshold Emax can be set; for example, when the speed of the receiving end > 60 km / h, it is determined that the current is a fast section, and Emax can be set to 0.5 m; conversely, when it is determined that the current is a slow section, Emax can be set to 0.2 m.

[0188] Moreover, the error of the sensing device in sensing the target position can also be determined. For example, the target data sent by the sensing device includes the position accuracy, and the corresponding error can be determined based on this position accuracy. For example, if the position accuracy includes the longitude error E1 and the latitude error E2, then the minimum value among the error threshold Emax, the longitude error E1, and the latitude error E2 can be used as the overlap error posion_error.

[0189] Step D13, when the two projections do not coincide, determine that the first oriented bounding box and the second oriented bounding box do not coincide.

[0190] Step D14, when the two projections coincide, use the normal line of one side that has not been used as the separation axis in the first oriented bounding box and the second oriented bounding box as the new separation axis to perform the next projection operation.

[0191] In this embodiment, in this projection operation, if it is determined that the two projections do not overlap, it can be directly determined that the first oriented bounding box and the second oriented bounding box do not overlap, and there is no need to determine whether the projections on other separation axes overlap, that is, there is no need to perform the projection operation again. If the two projections overlap in this projection operation, a new separation axis needs to be determined until all separation axes are traversed, or the projections on a certain separation axis do not overlap.

[0192] In some optional implementations, the above-mentioned step S5021 "determining bounding boxes that can cover the target respectively according to the target data sensed by multiple sensing devices" may include the following steps F1 to F2.

[0193] Step F1, determining axis-aligned bounding boxes of the targets respectively according to target data sensed by multiple sensing devices;

[0194] Step F2, when the axis-aligned bounding boxes of the target perceived by at least some of the sensing devices overlap, determining the oriented bounding box of the target perceived by at least some of the sensing devices.

[0195] In this embodiment, before making an overlap judgment based on the oriented bounding box, an overlap judgment can be made based on the axis-aligned bounding box (AABB). Since the range of the axis-aligned bounding box is larger than the range of the oriented bounding box, if the axis-aligned bounding boxes of the two targets do not overlap, it can be directly determined that the two targets do not overlap; if the axis-aligned bounding boxes of the two targets overlap, an overlap judgment can be further made based on the oriented bounding boxes of the two targets, that is, the oriented bounding boxes of the corresponding targets are determined at this time, that is, the above steps S5021 to S5022 are executed again. Although the axis-aligned bounding box will enlarge the actual area occupied by the target, the judgment logic is simple and can be detected quickly; the oriented bounding box can obtain an envelope that fits the target more closely and can be detected accurately.

[0196] For example, see Figure 7 As shown, there are currently four perceived targets, namely 701, 702, 703, and 704. An XY coordinate system can be determined based on one of the targets, and the axis-aligned bounding box of each target can be determined based on the XY coordinate system. Figure 7 As shown, the axis-aligned bounding box 705 of target 701, the axis-aligned bounding box 706 of target 702, the axis-aligned bounding box 707 of target 703, and the axis-aligned bounding box 708 of target 704 can be determined; and the maximum value Xmax and the minimum value Xmin of each axis-aligned bounding box on the X-axis, as well as the maximum value Ymax and the minimum value Ymin on the Y-axis can be determined. Based on the maximum and minimum values ​​of the two axis-aligned bounding boxes, it can be quickly determined whether the two axis-aligned bounding boxes overlap.

[0197] For example, if the maximum and minimum values ​​of one axis-aligned bounding box are X1max, X1min, Y1max, and Y1min, respectively, and the maximum and minimum values ​​of another axis-aligned bounding box are X2max, X2min, Y2max, and Y2min, respectively, if any of the following four judgment conditions are met, it can be determined that the two axis-aligned bounding boxes do not overlap. The four judgment conditions are: (1) Y2max <Y1min;(2)Y1max<Y2min;(3)X2max<X1min;(4)X1max<X2min。

[0198] like Figure 7 As shown, the axis-aligned bounding box 706 of target 702 coincides with the axis-aligned bounding box 707 of target 703, so it is only necessary to determine the oriented bounding boxes of targets 702 and 703, that is, it is not necessary to determine the oriented bounding boxes of targets 701 and 704. In addition, it can be understood that in order to clearly indicate the projection of the oriented bounding box onto the separation axis, Figure 6 Two targets that are far apart are used as an example.

[0199] This embodiment first performs a preliminary judgment based on the axis-aligned bounding box, which can exclude non-overlapping targets. The calculation is simple and can speed up the data deduplication speed. Then, the directional bounding box of the overlapping targets is preliminarily determined to accurately judge whether the targets whose axis-aligned bounding boxes overlap do not overlap, so that the final deduplication result is more accurate.

[0200] In some optional implementations, the above step S502 "determining whether the targets sensed by different sensing devices overlap based on the target data sensed by multiple sensing devices" may include the following steps G1 to G2.

[0201] Step G1, dividing the sensing detection range into multiple sub-areas according to lane lines.

[0202] Step G2, based on the target data sensed by the multiple sensing devices, it is determined whether the targets in the same sub-area sensed by different sensing devices overlap.

[0203] Since there may be many targets in the vehicle-road collaboration scenario, each perception device can perceive the target data of multiple targets. If all targets are judged whether they overlap with other targets, it will result in a large amount of data processing. The present embodiment divides the perception detection range into multiple sub-areas. By separately judging whether the targets in each sub-area overlap, overlap detection of all targets can be achieved. There is no need to judge the targets in different sub-areas, which can reduce the processing amount and improve processing efficiency.

[0204] Furthermore, in a vehicle-road collaboration scenario, a vehicle generally moves between two lane lines, so this embodiment divides the sensing detection range into multiple sub-areas according to the lane lines. Figure 8 A seven-lane scenario is shown; Figure 8 As shown in FIG. 1 , the sensing detection range of the local vehicle can be divided into multiple sub-areas according to multiple lane lines 801, wherein the area between two lane lines 801 can be used as a sub-area. In addition, the area before and after the local vehicle can be divided into 14 sub-areas. For details of these 14 sub-areas, see Figure 8 shown.

[0205] Step S503, when the targets sensed by different sensing devices overlap, valid target data is extracted from the target data of the overlapped targets; the valid target data includes the target data with the highest priority among the target data of the overlapped targets.

[0206] For more details, please refer to Figure 2 Step S203 of the illustrated embodiment will not be described in detail here.

[0207] Step S504, determining the actual position and posture of the overlapping target according to the valid target data.

[0208] For more details, please refer to Figure 2 Step S204 of the illustrated embodiment will not be described in detail here.

[0209] The data deduplication method provided in this embodiment uses a bounding box instead of a target, which can easily determine whether the targets overlap. First, a preliminary judgment is made based on the axis-aligned bounding box, which can exclude non-overlapping targets. The calculation is simple and can speed up the data deduplication speed; then the directional bounding boxes of the targets that are preliminarily determined to overlap are determined to accurately determine whether the targets that overlap with these axis-aligned bounding boxes overlap, so that the final deduplication result is more accurate. Based on the target data, the center projection distance of the centers of the two directional bounding boxes on the separation axis and the projection distance of each directional bounding box on the separation axis can be quickly determined. Based on the size relationship between the three projection distances, it can be simply and quickly determined whether the projections of the directional bounding boxes on the separation axis overlap. It does not require a complicated calculation process and has high processing efficiency. It is suitable for scenes with high timeliness requirements.

[0210] In this embodiment, a device for data deduplication is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0211] This embodiment provides a data deduplication device, such as Fig. 9 As shown, including:

[0212] The receiving module 901 is used to receive target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target;

[0213] A judgment module 902 is used to judge whether the targets sensed by different sensing devices overlap according to the target data sensed by the multiple sensing devices;

[0214] An extraction module 903 is used to extract valid target data from the target data of the target overlap when the targets sensed by different sensing devices overlap; the valid target data includes the target data with the highest priority among the target data of the target overlap;

[0215] The processing module 904 is used to determine the actual position and posture of the overlapping target according to the valid target data.

[0216] In some optional implementations, the target data further includes a timestamp;

[0217] The receiving module 901 receives target data respectively sensed by multiple sensing devices, including: determining a delay threshold according to the speed of a receiving end that receives the target data; the delay threshold is negatively correlated with the speed of the receiving end; and receiving target data respectively sensed by multiple sensing devices, whose timestamps do not exceed the delay threshold.

[0218] In some optional implementations, the target data further includes a timestamp; the extraction module 903 extracts valid target data from the target data that overlaps with the target, including:

[0219] From the target data with overlapping targets, the target data with the highest priority and the latest timestamp is extracted as the valid target data.

[0220] In some optional embodiments, the pose parameters include the perceived position, velocity and heading angle of the target, and the target data also includes a timestamp;

[0221] The judging module 902 judges whether the targets sensed by different sensing devices overlap according to the target data sensed by the plurality of sensing devices, including:

[0222] Determine, according to the speed of the target, the moving distance that the target can move from the timestamp to the current time;

[0223] Based on the perceived position of the target, increasing the moving distance along the heading angle to determine a predicted position of the target;

[0224] Determine whether the predicted positions of the targets perceived by different sensing devices overlap.

[0225] In some optional embodiments, the sensed location includes a sensed latitude and a sensed longitude;

[0226] The determination module 902 determines the predicted position of the target by increasing the moving distance along the heading angle based on the perceived position of the target, including:

[0227] Based on the perceived latitude of the target, adding the component of the moving distance in the latitude direction to determine the predicted latitude of the target;

[0228] Based on the perceived longitude of the target, the component of the moving distance in the longitude direction is added to determine the predicted longitude of the target.

[0229] In some optional embodiments, the target data further includes the size of the target;

[0230] The judging module 902 judges whether the targets sensed by different sensing devices overlap according to the target data sensed by the plurality of sensing devices, including:

[0231] Determine bounding boxes that can cover the target, respectively, based on the target data sensed by the plurality of sensing devices;

[0232] Determine whether the bounding boxes of the targets perceived by different sensing devices overlap.

[0233] In some optional embodiments, the bounding box is a directional bounding box;

[0234] The determining module 902 determines whether the bounding boxes of the targets sensed by different sensing devices overlap, including:

[0235] Taking the normal of one edge of the first oriented bounding box and the second oriented bounding box as the separation axis, cyclically performing the projection operation on the first oriented bounding box and the second oriented bounding box until it is determined that the first oriented bounding box does not overlap with the first oriented bounding box, or traversing all edges of the first oriented bounding box and the second oriented bounding box; the first oriented bounding box is the oriented bounding box of the target perceived by one perception device, and the second oriented bounding box is the oriented bounding box of the target perceived by another perception device;

[0236] The projection operation includes:

[0237] Projecting the first oriented bounding box and the second oriented bounding box onto the separation axis respectively;

[0238] Determine whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis;

[0239] In the case where the two projections do not overlap, determining that the first oriented bounding box does not overlap with the second oriented bounding box;

[0240] When the two projections overlap, the normal of one side of the first oriented bounding box and the second oriented bounding box that is not used as the separation axis is used as a new separation axis to perform the next projection operation.

[0241] In some optional implementations, the determining module 902 determines whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis, including:

[0242] Determine a center projection distance of a line segment between a center of the first oriented bounding box and a center of the second oriented bounding box on the separation axis;

[0243] Determine a first projection distance of the first oriented bounding box on the separation axis, and determine a second projection distance of the second oriented bounding box on the separation axis;

[0244] When the center projection distance is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections coincide with each other.

[0245] In some optional implementations, the judging module 902 determines that the two projections coincide when the center projection distance is less than half of the sum of the first projection distance and the second projection distance, including:

[0246] When the sum of the central projection distance and a preset overlap error is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections overlap.

[0247] In some optional embodiments, the device also includes: an overlapping error determination module, which is used to: determine an error threshold based on the speed of a receiving end that receives the target data; the error threshold is positively correlated with the speed of the receiving end; determine the error of the target position perceived by the sensing device, and take the minimum value between the error of the target position and the error threshold as the overlapping error.

[0248] In some optional implementations, the judgment module 902 determines the bounding boxes that can cover the target respectively according to the target data sensed by the plurality of sensing devices, including:

[0249] Determine axis-aligned bounding boxes of the targets respectively according to the target data sensed by the plurality of sensing devices;

[0250] In case the axis-aligned bounding boxes of the target perceived by at least some of the sensing devices coincide, an oriented bounding box of the target perceived by at least some of the sensing devices is determined.

[0251] In some optional implementations, the judging module 902 judges whether the targets perceived by different sensing devices overlap based on the target data perceived by the multiple sensing devices, including:

[0252] According to the lane lines, the sensing detection range is divided into multiple sub-areas;

[0253] According to the target data sensed by the plurality of sensing devices, it is determined whether the targets in the same sub-area sensed by different sensing devices overlap.

[0254] In some optional implementations, the receiving module 901 receives target data respectively sensed by a plurality of sensing devices, including:

[0255] Collect target data perceived by local sensing devices;

[0256] Receive at least one of target data in a perception sharing message format sent by other perception devices, target data in a basic safety message format sent by a perceived device, and target data in a roadside unit message format sent by a roadside device.

[0257] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0258] The data deduplication device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0259] The embodiment of the present invention also provides a computer device having the above Fig. 9 The device for deduplication of data is shown.

[0260] See also Fig.10 , Fig.10 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig.10As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.10 A processor 10 is taken as an example.

[0261] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0262] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0263] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0264] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0265] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Fig.10 The example of connecting through bus is taken in the following.

[0266] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0267] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0268] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A method for deduplication of data, characterized in that: The method comprises: Receiving target data respectively sensed by a plurality of sensing devices; the target data includes position and posture parameters of the target; According to the target data sensed by the plurality of sensing devices, determining whether the targets sensed by different sensing devices overlap; In the case where the targets sensed by different sensing devices overlap, valid target data is extracted from the target data of the overlapping targets; the valid target data includes the target data with the highest priority among the target data of the overlapping targets; The actual position and posture of the overlapping target is determined according to the valid target data.

2. The method according to claim 1, characterized in that: The target data also includes a timestamp; The receiving target data respectively sensed by a plurality of sensing devices comprises: Determining a delay threshold according to the speed of the receiving end receiving the target data; the delay threshold is negatively correlated with the speed of the receiving end; Receive target data respectively sensed by a plurality of sensing devices, the target data having a timestamp that does not exceed the delay threshold.

3. The method according to claim 1, characterized in that The target data also includes a timestamp; The step of extracting valid target data from target data with overlapping targets includes: From the target data with overlapping targets, the target data with the highest priority and the latest timestamp is extracted as the valid target data.

4. The method according to claim 1, characterized in that: The pose parameters include the perceived position, speed and heading angle of the target, and the target data also includes a timestamp; The step of judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the plurality of sensing devices includes: Determine, according to the speed of the target, the moving distance that the target can move from the timestamp to the current time; Based on the perceived position of the target, increasing the moving distance along the heading angle to determine a predicted position of the target; Determine whether the predicted positions of the targets perceived by different sensing devices overlap.

5. The method according to claim 4, characterized in that The sensed position includes a sensed latitude and a sensed longitude; The step of increasing the moving distance along the heading angle based on the perceived position of the target to determine the predicted position of the target includes: Based on the perceived latitude of the target, adding the component of the moving distance in the latitude direction to determine the predicted latitude of the target; Based on the perceived longitude of the target, the component of the moving distance in the longitude direction is added to determine the predicted longitude of the target.

6. The method according to claim 1, characterized in that The target data also includes the size of the target; The step of judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the plurality of sensing devices includes: Determine bounding boxes that can cover the target, respectively, based on the target data sensed by the plurality of sensing devices; Determine whether the bounding boxes of the targets perceived by different sensing devices overlap.

7. The method according to claim 6, characterized in that The bounding box is a directional bounding box; The determining whether the bounding boxes of the targets sensed by different sensing devices overlap includes: Taking the normal of one edge of the first oriented bounding box and the second oriented bounding box as the separation axis, cyclically performing the projection operation on the first oriented bounding box and the second oriented bounding box until it is determined that the first oriented bounding box does not overlap with the first oriented bounding box, or traversing all edges of the first oriented bounding box and the second oriented bounding box; the first oriented bounding box is the oriented bounding box of the target perceived by one perception device, and the second oriented bounding box is the oriented bounding box of the target perceived by another perception device; The projection operation includes: Projecting the first oriented bounding box and the second oriented bounding box onto the separation axis respectively; Determine whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis; In the case where the two projections do not overlap, determining that the first oriented bounding box does not overlap with the second oriented bounding box; When the two projections overlap, the normal of one side of the first oriented bounding box and the second oriented bounding box that is not used as the separation axis is used as a new separation axis to perform the next projection operation.

8. The method according to claim 7, characterized in that The determining whether the projection of the first oriented bounding box on the separation axis coincides with the projection of the second oriented bounding box on the separation axis includes: Determine a center projection distance of a line segment between a center of the first oriented bounding box and a center of the second oriented bounding box on the separation axis; Determine a first projection distance of the first oriented bounding box on the separation axis, and determine a second projection distance of the second oriented bounding box on the separation axis; When the center projection distance is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections coincide with each other.

9. The method according to claim 8, characterized in that The step of determining that the two projections coincide when the center projection distance is less than half of the sum of the first projection distance and the second projection distance includes: When the sum of the central projection distance and a preset overlap error is less than half of the sum of the first projection distance and the second projection distance, it is determined that the two projections overlap.

10. The method according to claim 9, characterized in that Also includes: determining an error threshold according to a speed of a receiving end receiving the target data; There is a positive correlation between the error threshold and the speed of the receiving end; An error in the target position perceived by the perception device is determined, and a minimum value between the error in the target position and the error threshold is used as the overlap error.

11. The method according to claim 7, characterized in that The step of respectively determining bounding boxes that can cover the target based on the target data sensed by the plurality of sensing devices includes: Determine axis-aligned bounding boxes of the targets respectively according to the target data sensed by the plurality of sensing devices; In case the axis-aligned bounding boxes of the target perceived by at least some of the sensing devices coincide, an oriented bounding box of the target perceived by at least some of the sensing devices is determined.

12. The method according to claim 1, characterized in that The step of judging whether the targets sensed by different sensing devices overlap based on the target data sensed by the plurality of sensing devices includes: According to the lane lines, the sensing detection range is divided into multiple sub-areas; According to the target data sensed by the plurality of sensing devices, it is determined whether the targets in the same sub-area sensed by different sensing devices overlap.

13. The method according to claim 1, characterized in that The receiving target data respectively sensed by a plurality of sensing devices comprises: Collect target data perceived by local sensing devices; Receive at least one of target data in a perception sharing message format sent by other perception devices, target data in a basic safety message format sent by a perceived device, and target data in a roadside unit message format sent by a roadside device.

14. A device for deduplication of data, characterized in that: The device comprises: A receiving module, used to receive target data sensed by multiple sensing devices respectively; the target data includes position and posture parameters of the target; A judgment module, used for judging whether the targets sensed by different sensing devices overlap according to the target data sensed by the plurality of sensing devices; An extraction module is used to extract valid target data from the target data of the target overlap when the targets sensed by different sensing devices overlap; the valid target data includes the target data with the highest priority among the target data of the target overlap; A processing module is used to determine the actual position and posture of the overlapping target according to the valid target data.

15. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the data deduplication method described in any one of claims 1 to 13 by executing the computer instructions.

16. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the data deduplication method according to any one of claims 1 to 13.

Images