A target tracking method, a computer readable storage medium and a driving device

Abstract

The present application relates to the technical field of intelligent driving, and particularly provides a target tracking method, a computer readable storage medium and a driving device, aiming at solving the problem of existing target tracking method output result jitter and easy to cause automatic parking failure. For this purpose, the target tracking method of the present application comprises: acquiring a current frame image of a parking scene, and determining a target to be tracked in the current frame image; acquiring a binding relationship between the target to be tracked and a parking space; acquiring position information of the bound parking space; based on the position information of the bound parking space and the relative position information between the target to be tracked and the bound parking space in the binding relationship, determining the tracking position of the target to be tracked, which avoids directly detecting the image of the target to be tracked based on the detection model to obtain the tracking position of the target to be tracked, and the relative position relationship between the target to be tracked and the bound parking space is more stable, which can effectively avoid the jitter of the tracking position, and thus is conducive to improving the accuracy of automatic parking.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent driving, and particularly provides a target tracking method, a computer readable storage medium and a driving device. BACKGROUND

[0002] With the development of technology, many vehicles can realize automatic parking function. In the prior art, when automatic parking is performed, an image acquisition device is usually used to acquire images of a parking space and obstacles in a parking scene, the parking space and the obstacles in the images are recognized, and target tracking is performed on the obstacles based on the images, so as to further determine position information of the parking space and guide the vehicle to complete automatic parking.

[0003] However, in the existing automatic parking scene, target tracking on the obstacles is generally directly performed by using a detection model to detect images containing the obstacles and determine positions, sizes and the like of the obstacles. Since the input of the detection model is an IPM (Inverse Perspective Mapping) image, and the model is limited by computing power and is relatively small, the output result of the detection model is often relatively jittery, which causes the tracked positions of the obstacles to be jittery and easily causes automatic parking to fail. SUMMARY

[0004] The present application aims to solve the above technical problem, that is, to solve the problem that the output result of the existing target tracking method is jittery and easily causes automatic parking to fail.

[0005] In a first aspect, the present application provides a target tracking method, comprising:

[0006] acquiring a current frame image of a parking scene and determining a target to be tracked in the current frame image;

[0007] acquiring a binding relationship between the target to be tracked and a parking space, the binding relationship comprising target to be tracked identification information, binding parking space identification information and relative position information between the target to be tracked and the binding parking space;

[0008] acquiring position information of the binding parking space;

[0009] determining a tracking position of the target to be tracked according to the position information of the binding parking space and the relative position information between the target to be tracked and the binding parking space.

[0010] In some embodiments, the acquiring of the binding relationship between the target to be tracked and the parking space comprises:

[0011] acquiring historical frame tracking data of the target to be tracked;

[0012] determining whether the to-be-tracked target and the parking space have a historical binding relationship according to the historical frame tracking data;

[0013] If yes, taking the historical binding relationship as the binding relationship of the to-be-tracked target and the parking space in the current frame image; if no, determining the binding relationship of the to-be-tracked target and the parking space based on a stage corresponding to the parking scene at a current moment.

[0014] In some embodiments, the determining the binding relationship of the to-be-tracked target and the parking space based on the stage corresponding to the parking scene at the current moment comprises:

[0015] When the parking scene corresponds to a parking stage, obtaining a first distance between the to-be-tracked target and a to-be-parked parking space;

[0016] determining whether the first distance is less than a first distance threshold;

[0017] If yes, taking the to-be-parked parking space as the binding parking space and constructing the binding relationship of the to-be-tracked target and the parking space based on the to-be-parked parking space;

[0018] Or,

[0019] When the parking scene corresponds to a searching parking space stage, determining a target parking space closest to the to-be-tracked target and taking the target parking space as the binding parking space;

[0020] constructing the binding relationship of the to-be-tracked target and the parking space based on the target parking space.

[0021] In some embodiments, the method further comprises updating the binding relationship of the to-be-tracked target and the parking space in the current frame image based on the stage corresponding to the parking scene.

[0022] In some embodiments, the updating the binding relationship of the to-be-tracked target and the parking space in the current frame image based on the stage corresponding to the parking scene comprises:

[0023] When the parking scene corresponds to a parking stage, determining whether a to-be-parked parking space and the binding parking space are a same parking space, and determining whether a second distance between the to-be-parked parking space and the to-be-tracked target is less than a second distance threshold;

[0024] When the to-be-parked parking space and the binding parking space are not the same parking space, and the second distance between the to-be-parked parking space and the to-be-tracked target is less than the second distance threshold, replacing the binding parking space with the to-be-parked parking space and updating the binding relationship of the to-be-tracked target and the parking space based on the to-be-parked parking space;

[0025] Or,

[0026] when the parking scene corresponds to a searching parking space stage, obtaining a target parking space closest to the to-be-tracked target;

[0027] determining whether the target parking space and the bound parking space are the same parking space, and determining whether a third distance between the target parking space and the to-be-tracked target satisfies a preset condition;

[0028] when the target parking space and the bound parking space are not the same parking space, and the third distance between the target parking space and the to-be-tracked target satisfies the preset condition, replacing the bound parking space with the target parking space and updating the binding relationship between the to-be-tracked target and the parking space based on the target parking space.

[0029] In some embodiments, the determining the to-be-tracked target in the current frame image comprises:

[0030] obtaining a historical frame image of the parking scene;

[0031] matching the target in the current frame image with a tracking target in the historical frame image;

[0032] when the matching is consistent, taking the target in the current frame image as the to-be-tracked target.

[0033] In some embodiments, the method further comprises:

[0034] when the matching is consistent, updating parameter information of the to-be-tracked target in the current frame image according to the parameter information of the tracking target in the historical frame image and the parameter information of the to-be-tracked target in the current frame image.

[0035] In some embodiments, the method further comprises:

[0036] when the matching is inconsistent, taking the target in the current frame image as the to-be-tracked target and constructing a binding relationship between the to-be-tracked target and a parking space based on a stage corresponding to the parking scene at a current time.

[0037] In a second aspect, the present application provides a computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and the computer program is executed by a processor to implement the target tracking method according to any one of the preceding aspects.

[0038] In a third aspect, the present application provides a driving device, which comprises a driving device body, a memory and a processor, wherein a computer program is stored in the memory, and the computer program is executed by the processor to implement the target tracking method according to any one of the preceding aspects.

[0039] By employing the above technical solution, this invention can determine the tracking position of the target by acquiring the binding relationship between the target and the parking space established based on the relative position information between the target and the bound parking space, as well as the position information of the bound parking space that has a binding relationship with the target. This method avoids directly detecting the image of the target based on the detection model to obtain the tracking position of the target, and the relative positional relationship between the target and the bound parking space is more stable, which can effectively avoid the jitter of the tracking position, thereby improving the accuracy of automatic parking. Attached Figure Description

[0040] The preferred embodiments of the present invention are described below with reference to the accompanying drawings, in which:

[0041] Figure 1 This is a flowchart illustrating a target tracking method provided in an embodiment of the present invention;

[0042] Figure 2 This is a schematic diagram of the method for determining the binding relationship between the target to be tracked and the parking space based on the stage corresponding to the parking scenario at the current moment, provided by an embodiment of the present invention;

[0043] Figure 3 This is a flowchart illustrating a method for determining the binding relationship between a target to be tracked and a parking space based on the stage corresponding to the parking scenario at the current moment, according to another embodiment of the present invention.

[0044] Figure 4 This is a flowchart illustrating a target tracking method according to another embodiment of the present invention;

[0045] Figure 5 This is a schematic flowchart of the method for implementing step S43 provided in an embodiment of the present invention;

[0046] Figure 6 This is a schematic diagram of the method flow for implementing step S43 provided in another embodiment of the present invention. Detailed Implementation

[0047] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0048] In the existing automatic parking scene, target tracking of an obstacle is generally directly performed by using a detection model to detect an image containing the obstacle and determine a position, size, etc. of the obstacle. Because the input of the detection model is an IPM (Inverse Perspective Mapping) image, and the output of the detection model is relatively jittery due to limited computing power and a small model, the tracking position of the obstacle is jittery and automatic parking is prone to failure.

[0049] Therefore, the present application provides a target tracking method. The tracking position of a target to be tracked is determined by obtaining a binding relationship between the target to be tracked and a parking space, which is established based on relative position information between the target to be tracked and the parking space, and position information of a binding parking space that has the binding relationship with the target to be tracked. The method avoids directly detecting an image of the target to be tracked based on a detection model to obtain the tracking position of the target to be tracked, and the relative position relationship between the target to be tracked and the binding parking space is more stable, which can effectively avoid jitter of the tracking position, thereby facilitating improvement of the accuracy of automatic parking.

[0050] Referring to Figure 1 Figure 1 is a flowchart of a target tracking method provided by an embodiment of the present application, which can include the following steps:

[0051] Step S11: Obtain a current frame image of a parking scene, and determine a target to be tracked in the current frame image.

[0052] Step S12: Obtain a binding relationship between the target to be tracked and a parking space, the binding relationship including target to be tracked identification information, binding parking space identification information, and relative position information between the target to be tracked and the binding parking space.

[0053] Step S13: Obtain position information of the binding parking space.

[0054] Step S14: Determine the tracking position of the target to be tracked according to the position information of the binding parking space and the relative position information between the target to be tracked and the binding parking space.

[0055] In some embodiments, the target to be tracked in the current frame image can be determined in step S11 by the following steps:

[0056] Obtain a historical frame image of the parking scene.

[0057] Match a target in the current frame image with a tracking target in the historical frame image.

[0058] When the matching is consistent, the target in the current frame image is taken as the target to be tracked.

[0059] ​As an example, two frames of images corresponding to a parking scene at time t-1 and time t can be acquired, and a target in a current frame of image corresponding to time t is matched with a tracking target in a historical frame of image corresponding to time t-1, when the matching is consistent, the target in the current frame of image is taken as a target to be tracked. The target to be tracked can include a parking pole or a cone barrel.

[0060] In some other embodiments, the method can further include: when the matching is consistent, updating the parameter information of the target to be tracked in the current frame of image according to the parameter information of the tracking target in the historical frame of image and the parameter information of the target to be tracked in the current frame of image.

[0061] When the matching is consistent, the same target as the tracking target in the historical frame of image is taken as the target to be tracked in the current frame of image, wherein the parameter information of the target to be tracked in the current frame of image can be updated based on new parameter information obtained by weighted summation of the parameter information of the tracking target in the historical frame of image and the parameter information of the target to be tracked in the current frame of image.

[0062] As an example, the parameter information can include the center point coordinates, length, etc. of the target to be tracked.

[0063] In some other embodiments, the method can further include: when the matching is inconsistent, taking a target in the current frame of image as a target to be tracked and constructing a binding relationship between the target to be tracked and a parking space based on a stage corresponding to the parking scene at the current time.

[0064] When the matching is inconsistent, i.e. the target in the current frame of image is not detected in the tracking target in the historical frame of image, the target in the current frame can be taken as a new target to be tracked.

[0065] The new target to be tracked has no historical binding relationship, which can be seen in the following Figure 2 Or Figure 3 The method shown is based on a stage corresponding to the parking scene at the current time to construct a binding relationship between the new target to be tracked and a parking space, which can be used to determine the tracking position of the target to be tracked in the next frame of image.

[0066] In some embodiments, a plurality of frames of images of the parking scene can be acquired by a vehicle-mounted image acquisition device.

[0067] In some embodiments, step S12 can be specifically:

[0068] Acquiring historical frame tracking data of the target to be tracked;

[0069] Determining whether the target to be tracked and the parking space have a historical binding relationship according to the historical frame tracking data;

[0070] If yes, the historical binding relationship is taken as the binding relationship between the target to be tracked and the parking space in the current frame image; if no, the binding relationship between the target to be tracked and the parking space is determined based on the stage corresponding to the current parking scene.

[0071] The historical binding relationship at least includes target-to-be-tracked identification information, binding parking space identification information, and relative position information between the target to be tracked and the binding parking space.

[0072] In some embodiments, as shown in FIG. 13, the binding relationship between the target to be tracked and the parking space can be determined based on the stage corresponding to the current parking scene by the following steps: Figure 2

[0073] Step S21: When the parking scene corresponds to a parking stage, a first distance between the target to be tracked and the parking space to be parked into is acquired.

[0074] Step S22: It is judged whether the first distance is less than a first distance threshold.

[0075] Step S23: If yes, the parking space to be parked into is taken as the binding parking space, and the binding relationship between the target to be tracked and the parking space is constructed based on the parking space to be parked into.

[0076] The first distance threshold can be set according to actual application needs.

[0077] In some other embodiments, as shown in FIG. 14, the binding relationship between the target to be tracked and the parking space can also be determined based on the stage corresponding to the current parking scene by the following steps: Figure 3

[0078] Step S31: When the parking scene corresponds to a search parking space stage, a target parking space closest to the target to be tracked is determined, and the target parking space is taken as the binding parking space.

[0079] Step S32: The binding relationship between the target to be tracked and the parking space is constructed based on the target parking space.

[0080] As an example, the relative position information between the target to be tracked and the binding parking space can include relative distance, relative angle and the like. In some embodiments, the relative position information between the target to be tracked and the binding parking space can be determined by a center point of the target to be tracked relative to a midpoint of an arbitrary side parking line of the binding parking space.

[0081] In some embodiments, step S13 can be specifically that the position information of the binding parking space is determined by acquiring an image of the binding parking space at a current time. The current time is a time corresponding to acquisition of the current frame image of the parking scene.

[0082] ​​In some embodiments, step S14 can be specifically determining the tracking position of the target to be tracked according to the position information of the binding parking space and the relative position information between the target to be tracked and the binding parking space. The position information of the binding parking space can include the position coordinates of the binding parking space in the vehicle coordinate system, and the position coordinates of the target to be tracked in the vehicle coordinate system can be determined according to the position coordinates of the binding parking space and the relative position information between the target to be tracked and the binding parking space.

[0083] The above is a target tracking method provided by an embodiment of the present application. The tracking position of the target to be tracked is determined by obtaining the binding relationship between the target to be tracked and the parking space established based on the relative position information between the target to be tracked and the binding parking space, and the position information of the binding parking space having the binding relationship with the target to be tracked. This method avoids directly detecting the image of the target to be tracked based on the detection model to obtain the tracking position of the target to be tracked, and the relative position relationship between the target to be tracked and the binding parking space is more stable, which can effectively avoid the jitter of the tracking position, thereby facilitating the improvement of the accuracy and success rate of automatic parking.

[0084] Referring to Figure 4 illustrated, Figure 4 is a flow diagram of a target tracking method provided by another embodiment of the present application, which can include:

[0085] Step S41: obtaining a current frame image of a parking scene and determining a target to be tracked in the current frame image;

[0086] Step S42: obtaining a binding relationship between the target to be tracked and the parking space, the binding relationship including target to be tracked identification information, binding parking space identification information, and relative position information between the target to be tracked and the binding parking space;

[0087] Step S43: updating the binding relationship between the target to be tracked and the parking space in the current frame image based on the corresponding stage of the parking scene;

[0088] Step S44: obtaining position information of the binding parking space;

[0089] Step S45: determining the tracking position of the target to be tracked according to the position information of the binding parking space and the relative position information between the target to be tracked and the binding parking space.

[0090] Steps S41 and S42 can be performed in the same manner as steps S11 and S12, respectively, and will not be described again here for brevity. For details, please refer to the description in the foregoing.

[0091] In some embodiments, referring to Figure 5 illustrated, step S43 can be specifically implemented by the following steps:

[0092] Step S51: When the parking scene corresponds to the parking stage, it is determined whether the to-be-parked-in parking space and the bound parking space are the same parking space, and whether the second distance between the to-be-parked-in parking space and the to-be-tracked target is less than a second distance threshold.

[0093] Step S52: When the to-be-parked-in parking space and the bound parking space are not the same parking space, and the second distance between the to-be-parked-in parking space and the to-be-tracked target is less than the second distance threshold, the bound parking space is replaced by the to-be-parked-in parking space, and the binding relationship between the to-be-tracked target and the parking space is updated based on the to-be-parked-in parking space.

[0094] The second distance threshold can be set according to actual needs.

[0095] In some embodiments, step S52 can be specifically that when the to-be-parked-in parking space and the bound parking space are not the same parking space, and the second distance between the to-be-parked-in parking space and the to-be-tracked target is less than the second distance threshold, the to-be-parked-in parking space identification information and the relative position information between the to-be-parked-in parking space and the to-be-tracked target are obtained, and a new binding relationship between the to-be-tracked target and the parking space is constructed based on the to-be-parked-in parking space identification information, the relative position information between the to-be-parked-in parking space and the to-be-tracked target, and the to-be-tracked target identification information.

[0096] In other embodiments, referring to FIG. 4, step S43 can be specifically implemented by the following steps: Figure 6

[0097] Step S61: When the parking scene corresponds to the searching parking space stage, a target parking space closest to the to-be-tracked target is obtained.

[0098] Step S62: It is determined whether the target parking space and the bound parking space are the same parking space, and whether a third distance between the target parking space and the to-be-tracked target satisfies a preset condition.

[0099] Step S63: When the target parking space and the bound parking space are not the same parking space, and the third distance between the target parking space and the to-be-tracked target satisfies the preset condition, the bound parking space is replaced by the target parking space, and the binding relationship between the to-be-tracked target and the parking space is updated based on the target parking space.

[0100] In some embodiments, the third distance between the target parking space and the to-be-tracked target can be compared with a relative distance between the bound parking space and the to-be-tracked target, and when the third distance is less than or equal to one-half of the relative distance, it is determined that the third distance satisfies the preset condition; otherwise, the third distance does not satisfy the preset condition.

[0101] ​In some embodiments, replacing the binding parking space with the target parking space and updating the binding relationship between the target to be tracked and the parking space based on the target parking space can specifically be: obtaining target parking space identification information and relative position information between the target parking space and the target to be tracked, and constructing a new binding relationship between the target to be tracked and the parking space based on the target parking space identification information, the relative position information between the target parking space and the target to be tracked, and the target to be tracked identification information.

[0102] In some embodiments, when the parking scene corresponds to a parking phase, the parking space to be parked in and the binding parking space are not the same parking space, and the second distance between the parking space to be parked in and the target to be tracked is less than the second distance threshold, step S44 can specifically be: obtaining the position information of the parking space to be parked in.

[0103] Step S45 can determine the tracking position of the target to be tracked according to the updated binding relationship between the target to be tracked and the parking space, which can specifically be: determining the tracking position of the target to be tracked according to the position information of the parking space to be parked in and the relative position information between the target to be tracked and the parking space to be parked in.

[0104] In other embodiments, when the parking scene corresponds to a search parking space phase, the target parking space and the binding parking space are not the same parking space, and the third distance between the target parking space and the target to be tracked meets the preset condition, step S44 can specifically be: obtaining the position information of the target parking space.

[0105] In some embodiments, step S45 can specifically be: determining the tracking position of the target to be tracked according to the position information of the target parking space and the relative position information between the target to be tracked and the target parking space.

[0106] The above is a target tracking method provided by another embodiment of the present application. The method can also update the binding relationship between the target to be tracked and the parking space in the current frame image based on the phase corresponding to the parking scene, to determine the tracking position of the target to be tracked based on the updated binding relationship between the target to be tracked and the parking space. In addition to the same beneficial effects as the corresponding embodiments, the binding relationship can be flexibly adjusted according to the phase corresponding to the parking scene, further improving the accuracy and success rate of automatic parking. Figure 1 In addition to the same beneficial effects as the corresponding embodiments, the binding relationship can be flexibly adjusted according to the phase corresponding to the parking scene, further improving the accuracy and success rate of automatic parking.

[0107] Another aspect of the present application also provides a computer readable storage medium having a computer program stored therein, which can implement the target tracking method in any of the above embodiments when executed by a processor. The computer readable storage medium can be a storage device formed by various electronic devices, and optionally, the computer readable storage medium in the embodiments of the present application is a non-transitory computer readable storage medium.

[0108] Another aspect of the present application also provides a driving device, which can comprise a driving device body, a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to implement the target tracking method in any of the above embodiments.

[0109] The driving device provided by the present application can be used to execute the target tracking method, and achieve the same beneficial effects as the target tracking method in the above embodiments.

[0110] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will all fall within the protection scope of the present application.

Claims

1. A target tracking method characterized by, The method comprises: acquiring a current frame image of a parking scene, and determining a target to be tracked in the current frame image; acquiring a binding relationship between the target to be tracked and a parking space, the binding relationship comprising target to be tracked identification information, binding parking space identification information, and relative position information between the target to be tracked and the binding parking space; acquiring position information of the binding parking space; determining a tracking position of the target to be tracked according to the position information of the binding parking space and the relative position information between the target to be tracked and the binding parking space; the acquiring of the binding relationship between the target to be tracked and the parking space comprises determining the binding relationship between the target to be tracked and the parking space based on a stage corresponding to the parking scene at a current time, wherein the determining of the binding relationship between the target to be tracked and the parking space based on the stage corresponding to the parking scene at the current time comprises: when the parking scene corresponds to a parking stage, acquiring a first distance between the target to be tracked and a target parking space; determining whether the first distance is less than a first distance threshold; if yes, taking the target parking space as the binding parking space and constructing the binding relationship between the target to be tracked and the parking space based on the target parking space; or, when the parking scene corresponds to a searching parking space stage, determining a target parking space closest to the target to be tracked, and taking the target parking space as the binding parking space; constructing the binding relationship between the target to be tracked and the parking space based on the target parking space.

2. The method of claim 1, wherein, Before the acquiring of the binding relationship between the target to be tracked and the parking space, the method further comprises: acquiring historical frame tracking data of the target to be tracked; determining whether there is a historical binding relationship between the target to be tracked and the parking space according to the historical frame tracking data; if no, determining the binding relationship between the target to be tracked and the parking space based on the stage corresponding to the parking scene at the current time again.

3. The method according to any one of claims 1 or 2, characterized in that, The method further comprises updating the binding relationship between the target to be tracked and the parking space in the current frame image based on the stage corresponding to the parking scene.

4. The method of claim 3, wherein, The updating of the binding relationship between the target to be tracked and the parking space in the current frame image based on the stage corresponding to the parking scene comprises: when the parking scene corresponds to a parking stage, determining whether a target parking space and the binding parking space are a same parking space, and determining whether a second distance between the target parking space and the target to be tracked satisfies a preset condition; when the target parking space and the binding parking space are not the same parking space, and the second distance between the target parking space and the target to be tracked is less than the second distance threshold, replacing the binding parking space with the target parking space and updating the binding relationship between the target to be tracked and the parking space based on the target parking space; or, when the parking scene corresponds to a searching parking space stage, acquiring a target parking space closest to the target to be tracked; determining whether the target parking space and the binding parking space are a same parking space, and determining whether a third distance between the target parking space and the target to be tracked satisfies a preset condition; When the target parking space and the binding parking space are not the same parking space, and a third distance between the target parking space and the to-be-tracked target satisfies the preset condition, the target parking space is replaced for the binding parking space, and the binding relationship between the to-be-tracked target and the parking space is updated based on the target parking space.

5. The method of claim 1, wherein, The method further includes: acquiring a historical frame image of the parking scene; matching a target in the current frame image with a tracking target in the historical frame image; when the matching is consistent, taking the target in the current frame image as the to-be-tracked target.

6. The method of claim 5, wherein, The method further includes: when the matching is consistent, updating parameter information of the to-be-tracked target in the current frame image according to parameter information of the tracking target in the historical frame image and parameter information of the to-be-tracked target in the current frame image.

7. The method of claim 5, wherein, The method further includes: when the matching is inconsistent, taking the target in the current frame image as the to-be-tracked target and constructing a binding relationship between the to-be-tracked target and a parking space based on a stage corresponding to the parking scene at a current moment.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the target tracking method in any one of claims 1 to 7.

9. A driving apparatus characterized by comprising: The driving device includes a driving device body, a memory, and a processor, the memory stores a computer program, and the computer program is executed by the processor to implement the target tracking method in any one of claims 1 to 7.

Images