Regional target tracking method, system and computer equipment based on spatial point cloud

Abstract

The present invention relates to a method, system and computer equipment for regional target tracking based on spatial point cloud, and belongs to the field of smart home. The method includes: obtaining point cloud data of a detection area; if a strong target in the detection area tracked by a target tracker disappears in the current frame point cloud data, the information of the strong target tracked before the corresponding moment of the current frame is placed in a preset recovery area; if a frame of point cloud data within a preset recovery time after the corresponding moment of the current frame has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then the information of the strong target is put back into the target tracker to continue tracking the strong target. The present invention can reduce the occurrence of target loss, thereby reducing the occurrence of cascade operation failures.

Description

Technical Field

[0001] The present invention relates to a method, system and computer equipment for tracking a regional target based on a space point cloud, and belongs to the field of smart home. Background Art

[0002] With the rapid development of Internet of Things technology, smart home devices are playing an increasingly important role in users' lives, and different home devices can also be linked. For example, when a target (i.e., user) is detected in the detection area, home devices are linked and cascade events occur, such as turning on lights and air conditioners. When no target is detected in the detection area, the cascade events are ended, such as turning off lights and air conditioners.

[0003] However, the existing methods can easily cause the tracking target (i.e., user) to disappear when the user is stationary or has very small movements after entering the area to be detected, resulting in misjudgment, that is, giving no-person information when there is someone, causing the cascade operation to fail. Summary of the invention

[0004] The purpose of the present invention is to provide a method, system and computer device for regional target tracking based on spatial point cloud, so as to solve the problem in the prior art that after a person enters an area and becomes stationary, target detection is easily lost, resulting in failure of cascade operation.

[0005] To achieve the above object, the solution of the present invention includes:

[0006] A method for tracking a regional target based on a spatial point cloud of the present invention comprises the following steps:

[0007] Obtain point cloud data of the detection area;

[0008] If a strong target in the detection area tracked by the target tracker disappears in the point cloud data of the current frame, the information of the strong target tracked before the corresponding time of the current frame is placed in the preset recovery area;

[0009] If a frame of point cloud data has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds within a preset recovery time after the corresponding moment of the current frame, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then the information of the strong target is put back into the target tracker to continue tracking the strong target;

[0010] A strong target is a target that needs special attention; a micro-dynamic point cloud represents a non-static point cloud with a relatively small movement amplitude; a dynamic point cloud represents a non-static point cloud other than a micro-dynamic point cloud; the second preset number is less than the first preset number; a missing time period represents the time period between the moment corresponding to the current frame and the moment corresponding to a frame of a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds; a target tracker is used to track a target.

[0011] Furthermore, methods for confirming whether it is a strong target include:

[0012] Get the trajectory information of each target tracked in the target tracker;

[0013] When the trajectory information of a certain target indicates that the trajectory of the target enters the detection area from outside the detection area, and the trajectory position data in the detection area is greater than a preset number, and the point cloud data density of each frame of the certain target is greater than the preset density, a strong target label is added to the certain target to characterize the certain target as a strong target.

[0014] Further, after a strong target label is added to the certain target, the strong target label is updated;

[0015] The process of updating the strong target label includes: when a certain target moves out of the detection area for at least one frame, deleting the strong target label.

[0016] Furthermore, the process of repairing the information of the strong target in the missing time period by using the information of the strong target in the restoration area includes: updating the information of the strong target in the missing time period to the information of the strong target in the previous frame of the current frame.

[0017] Furthermore, the target tracking process of the target tracker includes: obtaining each target to be matched in the point cloud data of the frame to be matched, matching each target to be matched with each target in the target tracker respectively, and if a target to be matched and a target in the target tracker meet all matching conditions, it is considered that the two are matched successfully;

[0018] The matching condition includes: the straight-line distance between the centroid position of a target to be matched and the latest position of a target in the target tracker on the horizontal plane is less than a preset straight-line distance.

[0019] Furthermore, the matching condition also includes: the distances between the centroid position of a target to be matched and the latest position of a target in the target tracker on two mutually perpendicular components of the horizontal plane respectively satisfy corresponding preset supplementary distances.

[0020] Further, the target to be matched is divided into a primary target and a secondary target. When matching the primary target and the secondary target with each target in the target tracker, the primary target is first matched with each target in the target tracker; after the primary target is matched, the secondary target is matched with the targets in the target tracker except those that are successfully matched with the primary target;

[0021] The preset straight-line distance corresponding to the first-level target is greater than the preset straight-line distance corresponding to the second-level target; the first-level target refers to the target to be matched whose number of dynamic point clouds is greater than the third preset number or the target to be matched whose number of non-static point clouds is greater than the fourth preset number; the second-level target refers to the target to be matched whose number of dynamic point clouds is less than the third preset number or the target to be matched whose number of non-static point clouds is less than the fourth preset number.

[0022] Furthermore, the target information includes: index number, whether it is a strong target label and whether it is a strong motion label;

[0023] The index number is the label of each target;

[0024] When the position difference of two consecutive frames of a certain target in a certain direction of the horizontal plane is greater than a preset position difference, a strong motion label is marked for the certain target to indicate that the certain target is a strong motion target.

[0025] Furthermore, the detection area is an area excluding the interference area; the interference area is an area that does not need to be detected.

[0026] Furthermore, the target tracking process of the target tracker further includes: filtering out symmetrical mirror targets;

[0027] The symmetrical mirror target is a target with equal velocity components in two mutually perpendicular directions on the horizontal plane in two consecutive frames of point cloud data.

[0028] Furthermore, the regional target tracking method based on spatial point cloud also includes: outputting whether there is a target in the detection area, the number of targets and the position of the targets according to the tracking situation of the target tracker.

[0029] A computer device of the present invention includes a processor, characterized in that the processor executes a computer program to implement the steps of the regional target tracking method based on spatial point cloud as described above.

[0030] The present invention provides a regional target tracking system based on spatial point cloud, including a radar and a processor. The radar is used to collect point cloud data in the detection area and transmit the point cloud data to the processor. The processor is used to execute a computer program to implement the steps of the regional target tracking method based on spatial point cloud.

[0031] The beneficial effects of the present invention are as follows: as a pioneering invention, the regional target tracking method, system and computer device based on spatial point cloud provided by the present invention can know the number of targets in the detection area by acquiring the point cloud data of the detection area; by placing the information of the strong target tracked before the corresponding moment of the current frame in the preset recovery area when the strong target in the detection area tracked by the target tracker disappears in the current frame point cloud data, the time for the strong target to disappear can be extended, and the occurrence of misjudgment due to misidentification can be reduced; if a frame of point cloud data has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds within the preset recovery time after the corresponding moment of the current frame, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then put the information of the strong target back into the target tracker to continue tracking the strong target, and the occurrence of the situation that the track is lost and cannot be tracked can be reduced by repairing the track of the strong target. The present invention can reduce the occurrence of target loss, thereby reducing the occurrence of cascade operation failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flowchart of a method for tracking a regional target based on a spatial point cloud provided by an embodiment of the present invention;

[0033] Figure 2 It is a schematic diagram of a trajectory update process provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0035] The idea of ​​the present invention is to set a short-time recovery area where the target disappears to improve the risk of preventing the target detection from being lost when the person is still, resulting in the failure of the cascade operation.

[0036] Specifically, by acquiring the point cloud data of the detection area, the number of targets in the detection area can be known; by placing the information of the strong target tracked before the corresponding moment of the current frame in the preset recovery area when the strong target in the detection area tracked by the target tracker disappears in the current frame point cloud data, the time for the strong target to disappear can be extended, and the occurrence of misjudgment due to misidentification can be reduced; if a frame of point cloud data within the preset recovery time after the corresponding moment of the current frame has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then put the information of the strong target back into the target tracker to continue tracking the strong target, and the occurrence of the situation where the track is lost and the tracking cannot be achieved can be reduced by repairing the track of the strong target. The present invention can reduce the occurrence of target loss, thereby reducing the occurrence of cascade operation failure.

[0037] An embodiment of a regional target tracking method based on spatial point cloud:

[0038] Figure 1 FIG. 1 is a flow chart of a method for tracking a regional target based on a spatial point cloud provided by an embodiment of the present invention. Figure 1 As shown, the method includes the following steps: S101, obtaining point cloud data of the detection area.

[0039] Specifically, the point cloud data of the detection area is acquired according to a preset sampling frequency and a preset sampling frame number.

[0040] The preset sampling frequency may be 1 second, 2 seconds, etc., and may be set according to actual conditions, and the present invention does not specifically limit this. The following takes the preset sampling frequency of 1 second as an example for exemplary description.

[0041] The preset sampling frame number may be 5 frames, 10 frames, etc., and may be set according to actual conditions, and the present invention does not specifically limit this. The following takes the preset sampling frame number of 5 frames as an example for exemplary description.

[0042] When the preset sampling frequency is 1 second and the preset sampling frame number is 5 frames: the point cloud data of the detection area is acquired according to the acquisition method of 5 frames per second.

[0043] The detection area is the area excluding the interference area; the interference area is the area that does not need to be detected.

[0044] S102: If a strong target in the detection area tracked by the target tracker disappears in the point cloud data of the current frame, the information of the strong target tracked before the corresponding time of the current frame is placed in a preset recovery area.

[0045] The method of determining whether it is a strong target includes the following steps S102A and S102B.

[0046] S102A: Obtaining trajectory information of each target tracked by the target tracker.

[0047] The trajectory information of 100 frames may be obtained, or other numbers may be obtained, which is not particularly limited in the present invention. The following description will be given by taking obtaining 100 frames as an example for illustrative explanation.

[0048] Specifically, if the target's trajectory information exceeds 100 frames, only the latest 100 frames of trajectory information are taken; if the target's trajectory data is less than 100 frames, all frames of trajectory information are read.

[0049] S102B. When the trajectory information of a certain target indicates that the trajectory of the target enters the detection area from outside the detection area, and the trajectory position data in the detection area is greater than a preset number, and the point cloud data density of each frame of the certain target is greater than a preset density, a strong target label is added to the certain target to indicate that the certain target is a strong target.

[0050] In order to improve tracking efficiency, after the strong target label is added to the target, the strong target label needs to be updated. The process of updating the strong target label includes: when the target moves out of the detection area for at least one frame, the strong target label is deleted.

[0051] The strong target is a target that needs to be focused on; the disappearance of the strong target may be the disappearance of the label of the strong target, or the disappearance of the strong target itself, which is not particularly limited in the present invention.

[0052] It is understandable that no matter whether the label of a strong target disappears or the strong target itself disappears, a large number of point clouds will disappear or decrease. After this phenomenon occurs, the information of the strong target tracked before the corresponding moment of the current frame is placed in the preset recovery area, which can extend the time when the strong target disappears and reduce the occurrence of misjudgment due to misidentification.

[0053] S103. If a frame of point cloud data contains a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds within a preset recovery time after the corresponding moment of the current frame, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then the information of the strong target is put back into the target tracker to continue tracking the strong target.

[0054] Among them, the micro-dynamic point cloud represents a non-static point cloud with a relatively small motion amplitude; the dynamic point cloud represents a non-static point cloud other than the micro-dynamic point cloud; the second preset number is less than the first preset number; the missing time period represents the time period between the moment when the strong target disappears and the moment when a frame corresponding to the first preset number of micro-dynamic point clouds or the second preset number of dynamic point clouds appears; the target tracker is used to track the target; the preset recovery time can be 2 minutes, or other time, etc., which can be selected according to actual conditions, and the present invention does not make any special limitation on this.

[0055] Among them, the first preset number of micro-dynamic point clouds or the second preset number of dynamic point clouds may be those of disappeared strong targets or new targets, etc., and the present invention does not make any special limitation on this.

[0056] The process of using the information of the strong target in the recovery area to repair the information of the strong target in the missing time period includes: updating the information of the strong target in the missing time period to the information of the strong target in the previous frame of the current frame. Repairing the trajectory of the strong target can reduce the occurrence of the situation where the trajectory is lost and the tracking cannot be carried out.

[0057] In order to improve the tracking efficiency, as an optional implementation method, the process of target tracking in the target tracker may also include: step S104, obtaining each to-be-matched target in the to-be-matched frame point cloud data, and matching each to-be-matched target with each target in the target tracker respectively; if a to-be-matched target and a target in the target tracker meet all matching conditions, it is considered that the two are matched successfully.

[0058] The matching condition includes: a straight-line distance between a centroid position of a target to be matched and a latest position of a target in the target tracker on a horizontal plane is less than a preset straight-line distance.

[0059] In order to reduce the occurrence of mismatching, as an optional implementation, the matching condition also includes: the distances between the center of mass position of a target to be matched and the latest position of a target in the target tracker on two mutually perpendicular components of the horizontal plane respectively satisfy corresponding preset supplementary distances.

[0060] Among them, the targets to be matched can be divided into primary targets and secondary targets.

[0061] The process of dividing the primary target and the secondary target may include the following sub-steps: step S104A and step S104B.

[0062] S104A, initializing the acquired point cloud data.

[0063] Specifically, S104A1, convert all point clouds in the millimeter wave radar coordinate system into point clouds in the real world coordinate system.

[0064] Specifically, a rotation matrix is ​​established according to the angle of the millimeter-wave radar chip: after testing, it was found that when the angle is 45 degrees, the point cloud effect in the plane is best, and the rotation matrix at this time is T; all point cloud coordinates are traversed, and each point cloud is multiplied by the rotation matrix T to obtain the coordinate position in the real world.

[0065] S104A2. Perform point cloud filtering on the point cloud data in the world coordinate system.

[0066] Specifically, the height limit threshold and the low degree threshold are obtained through scene statistical analysis, and all points in each point cloud are traversed, and all high interference points and low degree interference points are filtered out through the height and low degree thresholds. The purpose of filtering out high interference points and low degree interference points is to make the obtained point cloud belong to the target point cloud as much as possible.

[0067] S104A3. After filtering, the non-static point cloud data is divided into dynamic point cloud and micro-dynamic point cloud according to the frequency Doppler effect.

[0068] S104A4. Cluster the dynamic point cloud, and based on the clustering result, evaluate whether the target is the target that really needs to be tracked.

[0069] The point cloud data of the detection area directly obtained may include point cloud data other than the user. These point cloud data do not need to be tracked and are false targets. Therefore, it is necessary to cluster the dynamic point cloud to obtain the real target, that is, the target that needs to be tracked.

[0070] The clustering may adopt a density-based clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), or other clustering algorithms, which are not particularly limited in the present invention. The clustering algorithm is DBSCAN algorithm as an example for exemplary description.

[0071] For the DBSCAN algorithm, specifically, 1) Calculate the ε neighborhood of all points: For each point P in the data set, calculate how many neighbors there are in its ε neighborhood. The threshold of the number of neighbors is usually defined by a parameter MinPts.

[0072] 2) Marking core points: If the number of points in the ε neighborhood of a point is greater than or equal to MinPts, then the point is marked as a core point.

[0073] 3) Find density-connected points: For each core point, find all points that are density-connected to it. If point P is in the ε neighborhood of point O, and O is a core point, then P is a point that is density-connected to O.

[0074] 4) Marking noise points and boundary points: Points that are not marked as core points are marked as noise points. Points that are densely connected to a core point but are not core points themselves are marked as boundary points.

[0075] 5) Assign an independent cluster label to each core point or point connected to it by density: Assign an independent cluster label to each core point or point connected to it by density. If a point is densely connected to multiple core points, it will be assigned the cluster label of the first core point found.

[0076] S104A5. After clustering, the scattered points are filtered out.

[0077] After clustering, there may be some scattered points, which will reduce the effect of subsequent tracking. Therefore, these scattered points need to be filtered out.

[0078] Specifically, noise point cloud clusters in the point cloud data are eliminated, and point cloud clusters whose number is less than a first preset number of points in the point cloud data clusters are also eliminated.

[0079] Among them, the noise point cloud cluster is a point cloud cluster that does not belong to the target to be tracked, such as tables, chairs, air conditioners, etc.; the first preset number of points can be 10 points, or 5 points, etc., the present invention does not make any special limitation on this, and the following illustrative explanation is given by taking the first preset number of points being 10 points as an example.

[0080] As an optional implementation, the noise point cloud clusters may be removed first, and then the point cloud clusters with less than the first preset number of points may be removed; as another optional implementation, the point cloud clusters with less than the first preset number of points may be removed first, and then the noise point cloud clusters may be removed. The present invention does not specifically limit this, and the following takes the example of first removing the noise point cloud clusters and then removing the point cloud clusters with less than the first preset number of points as an example for exemplary description.

[0081] S104A6. After filtering out scattered points, evaluating whether the target is the target that really needs to be tracked based on the preset distance between each point cloud cluster and the preset tracking target.

[0082] Exemplarily, it is assumed that there are 5 point cloud clusters obtained by clustering, and the number of points in one of the point cloud clusters is less than 10; and it is assumed that the preset tracking targets include a first preset tracking target and a second preset tracking target.

[0083] The point cloud clusters with less than 10 points are removed, and 4 point cloud clusters are left. The distances between the remaining 4 point cloud clusters and the first preset tracking target and the distances between the remaining 4 point cloud clusters and the second preset tracking target are calculated respectively. The point cloud clusters that meet the first preset distance are classified as the point cloud clusters of the first preset tracking target, and the point cloud clusters that meet the second preset distance are classified as the point cloud clusters of the second preset tracking target. After the point cloud clusters are classified, the target is tracked according to the classification results.

[0084] S104A7: Analyze the dynamic point cloud after filtering out scattered points according to its density and voxel density to obtain the number of dynamic point clouds of the target, wherein the voxel density refers to the number of point clouds in a unit space.

[0085] S104B: Determine the target type of the target to be matched according to the number of dynamic point clouds or the number of non-static point clouds.

[0086] Specifically, a target to be matched whose number of dynamic point clouds is greater than the third preset number or whose number of non-static point clouds is greater than the fourth preset number is a first-level target; a target to be matched whose number of dynamic point clouds is less than the third preset number or whose number of non-static point clouds is less than the fourth preset number is a second-level target.

[0087] Among them, the third preset number can be 5 or other values, which can be set according to actual conditions. The present invention does not make any special limitation on this. The third preset number of 5 is taken as an example for exemplary explanation. The fourth preset number can be 10 or other values, which can be set according to actual conditions. The present invention does not make any special limitation on this. The fourth preset number of 10 is taken as an example for exemplary explanation.

[0088] When the targets to be matched are divided into primary targets and secondary targets, the specific matching process is as follows:

[0089] If the straight-line distance between the centroid position of a certain first-level target to be matched and the latest position of a certain target in the target tracker on the horizontal plane is less than the first preset straight-line distance, the two are matched successfully.

[0090] When the straight-line distance on the horizontal plane between the centroid position of a secondary target to be matched and the latest position of a target other than the target successfully matched with the primary target in the target tracker is less than the second preset straight-line distance, the two are successfully matched. Among them, the second preset straight-line distance is less than the first preset straight-line distance; the first preset straight-line distance can be 0.8 meters or other numbers, and the present invention does not make special restrictions on this. The first preset straight-line distance is 0.8 meters as an example for exemplary description; the second preset straight-line distance can be 0.5 meters or other numbers, and the present invention does not make special restrictions on this. The second preset straight-line distance is 0.5 meters as an example for exemplary description.

[0091] In order to reduce the occurrence of mismatching, as another optional implementation, when the straight-line distance on the horizontal plane between the center of mass position of a primary target to be matched and the latest position of a target in the target tracker is less than a first preset straight-line distance, and the distances between the two vertical components on the horizontal plane respectively meet the corresponding preset supplementary distances, the two are matched successfully.

[0092] When the straight-line distance on the horizontal plane between the center of mass position of a secondary target to be matched and the latest position of a target in the target tracker other than the target that successfully matches the primary target is less than the second preset straight-line distance, and the distances between the two on the two vertical components of the horizontal plane respectively meet the corresponding preset supplementary distances, the two are matched successfully.

[0093] In order to improve the matching efficiency, as an optional implementation, before matching the secondary target, the track that is successfully associated with the primary target can be erased.

[0094] Among them, the corresponding preset supplementary distance is a horizontal distance or a vertical distance; the value of the preset supplementary distance can be 0.45 meters, or other values, etc., which can be set according to actual needs. The present invention does not make any special limitation on this. This embodiment takes 0.45 meters as an example for illustrative explanation.

[0095] S105 , updating the information of the primary target or the secondary target that has been successfully associated and matched with the target in the target tracker to the corresponding successfully associated and matched target.

[0096] The target information may include the target's speed, acceleration, trajectory information, whether it is a strong target tag, whether it is a strong motion tag, and the like.

[0097] Among them, the index number is the label of each target; the strong target is the target that needs to be paid special attention to; when the position difference of two consecutive frames of a certain target in a certain direction on the horizontal plane is greater than the preset position difference, the strong motion label is marked on the certain target to indicate that the certain target is a strong motion target.

[0098] In order to improve the tracking accuracy, as an optional implementation, the target tracking process in the target tracker may further include: step S105, filtering out symmetrical mirror targets.

[0099] The symmetrical mirror target is a target whose velocity components in two vertical directions corresponding to the horizontal plane are equal in two consecutive frames of point cloud data.

[0100] Specifically, the data of the previous and next frames are selected to see whether the velocity component of the X-axis and the velocity component of the Y-axis of the same target are the same in the previous frame and the next frame. If they are the same, then see whether they are symmetrical about a certain point or a certain plane. If they are symmetrical about a certain point or a certain plane, the target is a symmetrical mirror target.

[0101] S106, outputting whether there is a target in the detection area, the number of targets and the position of the targets according to the tracking status of the target tracker, and updating the target tracker according to the current frame status.

[0102] The number of targets may be one or more, and the present invention does not impose any special limitation on this.

[0103] Figure 2 is a schematic diagram of a trajectory update process provided by an embodiment of the present invention, such as Figure 2 As shown, there are multiple target detections, multiple old tracks in the area, and multiple new tracks in the area. When the old tracks and the targets to be detected are matched for the first time according to the spatial distance, some targets are matched to the old tracks, and some targets are not matched to the old tracks. The targets to be detected that are not matched to the old tracks, the old tracks that are not matched to the targets to be detected (that is, the remaining old tracks and the remaining targets to be detected), all new tracks, and new targets to be detected are matched again according to the spatial distance. During the secondary matching, some targets to be detected (including the old remaining targets to be detected and the new targets to be detected) are matched to the new tracks, and some targets to be detected are matched to the old tracks.

[0104] Delete the target to be tested that has not been matched to the trajectory for two consecutive times (i.e., unmatched detection), delete the old trajectory that has not been matched to the target to be tested within the preset time; delete the remaining old target to be tested that has not been matched to the trajectory for two consecutive times; and temporarily define the new trajectory that has not matched any target as a new trajectory.

[0105] Before the target tracker performs target tracking, step S107 can be executed to initialize the multi-target tracker. Specifically, the tracker is first initialized according to relevant parameters such as the detection area, the folding area, the rotation angle, the number of detections when the target is set as a fixed target, etc. The initialization at least includes: initialization of target data classification, initialization of single target tracking, and initialization of multi-target set update.

[0106] Among them, the detection area refers to the area in the room that needs to be detected; the folding area refers to the area in the room that is not detected, that is, the interference area; the rotation angle is the rotation angle required to convert all the point clouds in the millimeter wave radar coordinate system into the point clouds in the real coordinate system.

[0107] Among them, the initialization of target data classification refers to: initializing the collected point cloud data in the multi-target tracker, and dividing these point clouds into dynamic point clouds and micro-dynamic point clouds according to the number of points and distance in the point clouds.

[0108] The initialization of single target tracking means: setting how many frames of continuous detection are needed for the target to be tracked to be the real target, as well as some information about the target itself. For example, if 5 frames of data are collected in 1 second, if the target to be tracked appears in 5 frames in 1 second, it is considered to be the real target to be tracked and is put into the multi-target tracker as the preset target; if the target to be tracked does not appear in 5 frames in 1 second, it is considered not to be the real target to be tracked and is not put into the multi-target tracker.

[0109] Initialization of multi-target collection update means: the multi-target aggregator also needs to set initialization parameters; the multi-target aggregator is used to store some information related to multiple targets. For example, if a strong target disappears, the strong target disappears while walking, store this target first, and call it out when it appears next time, so as to initialize faster.

[0110] The method for tracking regional targets based on spatial point clouds provided by the embodiment of the present invention can know the number of targets in the detection area by acquiring the point cloud data of the detection area; by placing the information of the strong targets tracked before the corresponding moment of the current frame in the preset recovery area when the strong targets in the detection area tracked by the target tracker disappear in the current frame point cloud data, the time for the strong targets to disappear can be extended, and the occurrence of misjudgment due to misidentification can be reduced; if a frame of point cloud data has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds within the preset recovery time after the corresponding moment of the current frame, the information of the strong targets in the recovery area is used to repair the information of the strong targets in the missing time period and then put the information of the strong targets back into the target tracker to continue tracking the strong targets, and the occurrence of the situation where the track is lost and the tracking cannot be achieved can be reduced by repairing the track of the strong targets. The present invention can reduce the occurrence of target loss, thereby reducing the occurrence of cascade operation failure.

[0111] An embodiment of a computer device:

[0112] A computer device of the present invention includes a processor, and the processor executes a computer program to implement the steps of the above-mentioned regional target tracking method based on spatial point cloud.

[0113] An embodiment of the present invention provides a computer device that can achieve the same beneficial effects as the aforementioned method for tracking regional targets based on spatial point clouds, which will not be described in detail here.

[0114] An embodiment of a regional target tracking system based on spatial point cloud:

[0115] An embodiment of the present invention provides a regional target tracking system based on spatial point cloud, including a radar and a processor, wherein the radar is used to collect point cloud data in a detection area and transmit the point cloud data to the processor; the processor is used to execute a computer program to implement the steps of the regional target tracking method based on spatial point cloud as described above.

[0116] An embodiment of the present invention provides a regional target tracking system based on spatial point cloud, which can achieve the same beneficial effects as the aforementioned regional target tracking method based on spatial point cloud, and will not be described in detail here.

Claims

1. A method for tracking regional targets based on spatial point clouds, characterized in that: The steps include: Obtain point cloud data of the detection area; If a strong target in the detection area tracked by the target tracker disappears in the point cloud data of the current frame, the information of the strong target tracked before the corresponding time of the current frame is placed in a preset recovery area; If a frame of point cloud data has a first preset number of micro-dynamic point clouds or a second preset number of dynamic point clouds within a preset recovery time after the corresponding moment of the current frame, the information of the strong target in the recovery area is used to repair the information of the strong target in the missing time period and then the information of the strong target is put back into the target tracker to continue tracking the strong target; The strong target is a target that needs to be paid special attention to; the micro-dynamic point cloud represents a non-static point cloud with a relatively small motion amplitude; the dynamic point cloud represents a non-static point cloud other than the micro-dynamic point cloud; the second preset number is less than the first preset number; the missing time period represents the time period between the moment corresponding to the current frame and the moment corresponding to a frame of the first preset number of micro-dynamic point clouds or the second preset number of dynamic point clouds; the target tracker is used to track the target.

2. The method for tracking regional targets based on spatial point cloud according to claim 1, characterized in that: Ways to confirm whether it is a strong target include: Get the trajectory information of each target tracked in the target tracker; When the trajectory information of a certain target indicates that the trajectory of the target enters the detection area from outside the detection area, and the trajectory position data in the detection area is greater than a preset number, and the point cloud data density of each frame of the certain target is greater than the preset density, a strong target label is added to the certain target to characterize the certain target as a strong target.

3. The method for tracking regional targets based on spatial point clouds according to claim 2, characterized in that: After the strong target label is added to the target, the strong target label is updated; The process of updating the strong target label includes: when the target moves out of the detection area for at least one frame, deleting the strong target label.

4. The method for tracking regional targets based on spatial point cloud according to claim 1, characterized in that: The process of repairing the information of the strong target in the missing time period by using the information of the strong target in the recovery area includes: updating the information of the strong target in the missing time period to the information of the strong target in the previous frame of the current frame.

5. The method for tracking regional targets based on spatial point cloud according to claim 1, characterized in that: The target tracking process of the target tracker includes: obtaining each target to be matched in the point cloud data of the frame to be matched, matching each target to be matched with each target in the target tracker respectively, and if a target to be matched and a target in the target tracker meet all the matching conditions, then the two are considered to be matched successfully; The matching condition includes: a straight-line distance on a horizontal plane between a centroid position of a target to be matched and a latest position of a target in the target tracker is less than a preset straight-line distance.

6. The method for tracking regional targets based on spatial point cloud according to claim 5, characterized in that: The matching condition also includes: the distances between the centroid position of a target to be matched and the latest position of a target in the target tracker on two mutually perpendicular components of the horizontal plane respectively satisfy corresponding preset supplementary distances.

7. The method for tracking regional targets based on spatial point cloud according to claim 5 or 6, characterized in that: The targets to be matched are divided into primary targets and secondary targets. When matching the primary targets and secondary targets with the targets in the target tracker, the primary targets are first matched with the targets in the target tracker. After the primary targets are matched, the secondary targets are matched with the targets in the target tracker except those that are successfully matched with the primary targets. The preset straight-line distance corresponding to the first-level target is greater than the preset straight-line distance corresponding to the second-level target; the first-level target refers to the target to be matched whose number of dynamic point clouds is greater than the third preset number or the target to be matched whose number of non-static point clouds is greater than the fourth preset number; the second-level target refers to the target to be matched whose number of dynamic point clouds is less than the third preset number or the target to be matched whose number of non-static point clouds is less than the fourth preset number.

8. The method for tracking regional targets based on spatial point cloud according to claim 3, characterized in that: The target information includes: index number, whether it is a strong target label and whether it is a strong motion label; The index number is the label of each target; When the position difference of two consecutive frames of a certain target in a certain direction of the horizontal plane is greater than a preset position difference, the certain target is marked with the strong motion label to indicate that the certain target is a strong motion target.

9. The method for tracking regional targets based on spatial point cloud according to claim 1, characterized in that: The detection area is an area excluding the interference area; the interference area is an area that does not need to be detected.

10. The method for tracking regional targets based on spatial point cloud according to claim 5, characterized in that: The target tracking process of the target tracker also includes: filtering out symmetrical mirror targets; The symmetrical mirror target is a target having equal velocity components in two mutually perpendicular directions on the horizontal plane in two consecutive frames of point cloud data.

11. The method for tracking regional targets based on spatial point cloud according to claim 1, characterized in that: The regional target tracking method based on spatial point cloud also includes: outputting whether there is a target in the detection area, the number of targets and the position of the targets according to the tracking situation of the target tracker.

12. A computer device comprising a processor, characterized in that: The processor executes a computer program to implement the steps of the method for regional target tracking based on spatial point cloud as described in any one of claims 1-11.

13. A regional target tracking system based on spatial point cloud, comprising a radar and a processor, wherein the radar is used to collect point cloud data in a detection area and transmit the point cloud data to the processor, characterized in that: The processor is used to execute a computer program to implement the steps of the regional target tracking method based on spatial point cloud as described in any one of claims 1-11.

Images