A vehicle tracking method, device, computer equipment and storage medium

Abstract

The application discloses a vehicle tracking method and device, computer equipment and a storage medium. The method comprises the following steps: receiving first vehicle information of a target vehicle sent by a current tracking device; when the target vehicle drives to a target position, acquiring second vehicle information; predicting driving decision information of the target vehicle based on the second vehicle information and the first vehicle information; and determining a target tracking device based on the driving decision information, so as to track the target vehicle through the target tracking device. Through the prediction of multi-dimensional vehicle information, the driving decision information of the vehicle is obtained, so that the corresponding tracking device is called to track the vehicle, the relay tracking capability across tracking devices is realized, the accuracy and real-time performance of vehicle monitoring are improved, and the user experience is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the video monitoring technical field, and particularly relates to a vehicle tracking method and device, computer equipment and a storage medium. BACKGROUND

[0002] With the continuous development of unattended parking lots, vehicle recognition devices are widely used in various parking lot entrances and in parking lots for continuously detecting and tracking vehicles to guide and monitor the driving routes of the vehicles.

[0003] However, due to complex environmental conditions in parking lots, mobile network is poor, navigation of underground parking lots is difficult to be accurate, information transmission has certain time delay, and the recognition range of the vehicle recognition device is limited, so that the existing vehicle recognition device is difficult to accurately monitor and guide the entire driving process of the vehicle. Therefore, the accuracy and real-time performance of vehicle control cannot achieve satisfactory results, resulting in poor user experience. SUMMARY

[0004] Therefore, it is necessary to provide a vehicle tracking method, device, computer equipment and storage medium to solve at least one problem in the prior art.

[0005] In a first aspect, the embodiments of the present application are implemented as follows, a vehicle tracking method is provided, applied to a candidate tracking device, the candidate tracking device is determined by a vehicle driving direction of a target vehicle collected by a current tracking device and a pre-constructed inter-tracking device topology relationship graph, the method comprises:

[0006] receiving first vehicle information of the target vehicle collected by the current tracking device;

[0007] when detecting that the target vehicle drives to a target position, acquiring second vehicle information;

[0008] based on the first vehicle information and the second vehicle information, predicting driving decision information of the target vehicle, the driving decision information representing a driving intention of the target vehicle;

[0009] based on the driving decision information, determining a target tracking device to track the target vehicle through the target tracking device.

[0010] In a second aspect, a vehicle tracking device is provided, applied to a candidate tracking device, the candidate tracking device is determined by a vehicle driving direction of a target vehicle collected by a current tracking device and a pre-constructed inter-tracking device topology relationship graph, the device comprises:

[0011] a first vehicle information receiving unit configured to receive first vehicle information of a target vehicle collected by the current tracking device;

[0012] a second vehicle information obtaining unit configured to obtain second vehicle information when the target vehicle drives to a target position;

[0013] a driving decision information predicting unit configured to predict driving decision information of the target vehicle based on the first vehicle information and the second vehicle information;

[0014] a vehicle tracking unit configured to determine a target tracking device based on the driving decision information, so as to track the target vehicle by the target tracking device.

[0015] In a third aspect, a computer device is provided, which includes a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, and the processor implements the steps of the vehicle tracking method as described above when executing the computer readable instructions.

[0016] In a fourth aspect, a readable storage medium is provided, which stores computer readable instructions, and the computer readable instructions implement the steps of the vehicle tracking method as described above when executed by a processor.

[0017] The vehicle tracking method, device, computer device, and storage medium described above implement the method, which includes: receiving first vehicle information sent by the current tracking device; obtaining second vehicle information when the target vehicle drives to a target position; predicting driving decision information of the target vehicle based on the first vehicle information and the second vehicle information; and determining a target tracking device based on the driving decision information, so as to track the target vehicle by the target tracking device. In the embodiments of the present application, multi-dimensional vehicle information is predicted by network prediction capability, driving decision information of the vehicle is obtained, i.e., driving intention of the vehicle is determined, so as to call corresponding tracking devices to track the vehicle, relay tracking capability across tracking devices can be achieved, accuracy and real-time performance of vehicle monitoring can be improved, and user experience can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 is an application environment diagram of a vehicle tracking method in an embodiment of the present application Figure 1;

[0020] Figure 2 is an application environment diagram of a vehicle tracking method in an embodiment of the present application Figure 2 ;

[0021] Figure 3 is a flow diagram of a vehicle tracking method in an embodiment of the present application Figure 1 ;

[0022] Figure 4 is a region division scenario diagram of a tracking device in an embodiment of the present application

[0023] Figure 5 is a flow diagram of a target tracking device determination method in an embodiment of the present application

[0024] Figure 6 is a flow diagram of a vehicle tracking method in an embodiment of the present application Figure 2 ;

[0025] Figure 7 is a tracking device and candidate tracking device distribution scenario diagram when the road type is a straight road in an embodiment of the present application

[0026] Figure 8 is a tracking device and candidate tracking device distribution scenario diagram when the road type is an L-shaped road in an embodiment of the present application

[0027] Figure 9 is a tracking device and candidate tracking device distribution scenario diagram when the road type is a T-shaped road in an embodiment of the present application

[0028] Figure 10 is a tracking device and candidate tracking device distribution scenario diagram when the road type is a crossroad in an embodiment of the present application

[0029] Figure 11 is a flow diagram of a vehicle tracking method in an embodiment of the present application Figure 4 ;

[0030] Figure 12 is a structure diagram of a vehicle tracking device in an embodiment of the present application

[0031] Figure 13 is a diagram of a computer device in an embodiment of the present application DETAILED DESCRIPTION

[0032] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts are within the scope of the present application.

[0033] The vehicle tracking method provided in the embodiments can be applied in an application environment as shown in Figure 1 For example, when the tracking device A detects a target vehicle, the tracking device A can send the vehicle information of the target vehicle to its adjacent tracking device B. The tracking device B can be at least one. If the tracking device B detects that the target vehicle drives into the tracking area of the tracking device B, the tracking device B can collect the vehicle information of the target vehicle, and predict the driving decision information of the target vehicle based on the vehicle information. Each candidate tracking device can send the predicted driving decision information to other candidate tracking devices. Each candidate tracking device can compare the driving decision information, and select the driving decision information with the highest confidence. The target vehicle is tracked by the candidate tracking device corresponding to the driving decision information with the highest confidence, so as to realize relay tracking across tracking devices. Without the participation of a background system, relay tracking can be realized only by the tracking devices, the accuracy of video navigation is realized, the vehicle information and the decision information are sent or broadcasted between cameras, the timeliness of information transmission is improved through process communication, accurate tracking of the vehicle is realized, and user experience is improved.

[0034] The vehicle tracking method provided in the embodiments can be applied in an application environment as shown in Figure 2 For example, when the tracking device A detects a target vehicle, the tracking device A can send the vehicle information of the target vehicle to its adjacent tracking device B. The tracking device B can be at least one. If the tracking device B detects that the target vehicle drives into the tracking area of the tracking device B, the tracking device B can collect the vehicle information of the target vehicle, and predict the driving decision information of the target vehicle based on the vehicle information. Each candidate tracking device can send the predicted driving decision information to other candidate tracking devices. Each candidate tracking device can compare the driving decision information, and select the driving decision information with the highest confidence. The target vehicle is tracked by the candidate tracking device corresponding to the driving decision information with the highest confidence, so as to realize relay tracking across tracking devices. Without the participation of a background system, relay tracking can be realized only by the tracking devices, the accuracy of video navigation is realized, the vehicle information and the decision information are sent or broadcasted between cameras, the timeliness of information transmission is improved through process communication, accurate tracking of the vehicle is realized, and user experience is improved.

[0035] It should be noted that the tracking device can be arranged on both sides of the road or above the road for detecting and tracking the vehicle driving in the road. The tracking device can be arranged in different arrangement manners according to different road types. For example, when the road type is a straight road, the tracking device can be arranged on both sides of the road, and the tracking devices can be arranged on the same side of the road or staggered on both sides. And a tracking device can be arranged every preset distance.

[0036] In an embodiment, as shown in Figure 3 A vehicle tracking method is provided, applied to a candidate tracking device, determined by a vehicle driving direction of a target vehicle collected by a current tracking device and a pre-constructed inter-tracking device topology relationship graph, including the following steps:

[0037] In step S110, the first vehicle information sent by the current tracking device is received.

[0038] In the embodiment of the present application, the current tracking device is a camera, a camera, etc. which can capture the to-be-recognized area in real time to obtain the to-be-detected video data. The to-be-detected video data is split into continuous video frames, and the target vehicle in each video frame is detected, for example, the YOLO algorithm or the re-identification algorithm can be used, so as to determine the running track of the target vehicle. The tracking device can also identify the first vehicle information of the target vehicle, so as to obtain the first vehicle information of the target vehicle.

[0039] The first vehicle information can include license plate number, vehicle ID, vehicle logo, vehicle color, vehicle model, vehicle type, vehicle driving direction, whether the photographed is the front or the tail of the vehicle, and other multi-dimensional information.

[0040] In the embodiment of the present application, the node in the inter-tracking device topology relationship graph represents the tracking device, and the edge represents the relay tracking relationship between two tracking devices. The topology relationship graph can be constructed according to the road type. For example, when the road type is a straight road, the adjacent tracking devices of tracking device A are tracking device B and tracking device C, and the topology relationship can be B-A-C.

[0041] Further, according to the topology relationship graph and the driving direction of the target vehicle, the candidate tracking device can be determined. For example, the topology relationship is B-A-C, and the driving direction of the vehicle is from the position of tracking device A to the position of tracking device B, so the candidate tracking device can be determined as tracking device B.

[0042] If the topological relationship is B-A-C-D, and one side of the tracking device A is a straight road, the other side is an L-shaped road, and the adjacent tracking device on the straight road side of the tracking device A is B, and the adjacent tracking device on the L-shaped road side of the tracking device A is C and D, at this time, if the driving direction of the target vehicle is from the position of the tracking device A to the position of the tracking device C, and from the position of the tracking device C to the position of the tracking device D, it can be determined that the candidate tracking device is the tracking device C.

[0043] In the embodiment of the application, the candidate tracking device can be one or more. After the candidate tracking device is determined, the first vehicle information can be sent to each candidate tracking device by the current tracking device, or the first vehicle information can also be sent to each candidate tracking device by the parking lot platform. For example, the first vehicle information can be broadcast to each candidate tracking device by Bluetooth broadcast, or the first vehicle information can be broadcast to each candidate tracking device by WiFi, 4G, 5G, etc. Network mode, so that the candidate tracking device can detect whether the target vehicle appears in the target position according to the first vehicle information.

[0044] It can be understood that the first vehicle information can include license plate number, vehicle ID, etc. By sending the first vehicle information to the candidate tracking device, the candidate tracking device can determine the target vehicle. Since there may be multiple vehicles passing by at the same time, the candidate tracking device cannot determine which vehicle needs to be tracked, so by using the first vehicle information, the tracking object can be determined, and the problem of tracking the wrong vehicle can be avoided.

[0045] In step S120, when the target vehicle drives to the target position, the second vehicle information is obtained.

[0046] Referring to Figure 4 In the embodiment of the application, each tracking device can include a blind area, a tracking area, and an identification area. The blind area refers to an area where the tracking device cannot detect the target vehicle, the tracking area refers to an area where the target vehicle is tracked, and the identification area refers to an area where the vehicle information of the target vehicle is identified. It should be noted that the tracking area of the current tracking device is the blind area of its adjacent tracking device. For example, the tracking area of the tracking device A is the blind area of the tracking device B, and the arrow direction is the driving direction of the target vehicle.

[0047] The target position refers to the identification area of the candidate tracking device.

[0048] In the embodiment of the application, the second vehicle information can include license plate number, vehicle ID, vehicle logo, vehicle body color, vehicle model, vehicle type, vehicle driving direction, whether the photographed part is the front or the rear of the vehicle, and other multi-dimensional information.

[0049] In step S130, driving decision information of the target vehicle is predicted based on the first vehicle information and the second vehicle information, the driving decision information representing a driving intention of the target vehicle.

[0050] In the embodiment of the present application, the driving decision information is used to represent the driving intention of the target vehicle, such as the next driving action, the driving direction or the driving speed, etc. Taking the driving direction as an example, for example, when the road type is an intersection, a candidate tracking device is set at each intersection, at this time, each candidate tracking device can predict the driving decision information of the target vehicle, so as to determine which direction the target vehicle will drive.

[0051] In the embodiment of the present application, a prediction model can be constructed, such as BP neural network, decision tree, random forest, etc. The prediction model is iteratively trained by training sample data, and when it meets the preset convergence condition, the trained prediction model is obtained. Based on the trained prediction model, the second vehicle information collected by each candidate tracking device and the first vehicle information of the tracking device are predicted to obtain the score corresponding to each information, and the weight obtained by the trained model is weighted and summed to obtain the driving decision score of each candidate tracking device.

[0052] In the embodiment of the present application, if there are multiple candidate tracking devices, when the target vehicle drives to the tracking area of the candidate tracking device, the running video data of the target vehicle in the tracking area can be collected, and a virtual identification line can be set in the tracking area in the running video data. If it is detected that the vehicle detection box of the target vehicle collides with the virtual identification line, i.e., it coincides with the virtual identification line, the prediction of the driving decision information can be triggered. It can be understood that if there are multiple candidate tracking devices, when the target vehicle drives to the overlapping area of the shooting angles of the multiple candidate tracking devices, a virtual identification line can be set in the overlapping area, and if it is detected that the vehicle detection box of the target vehicle collides with the virtual identification line, i.e., it coincides with the virtual identification line, the prediction of the driving decision information can be triggered.

[0053] In step S140, based on the driving decision information, the target tracking device is determined to track the target vehicle through the target tracking device.

[0054] In the embodiment of the present application, the driving decision information can include a driving decision score, and the driving decision scores of each candidate tracking device are compared to determine the candidate tracking device with the highest driving decision score as the target tracking device, and the target tracking device is tracked by relay.

[0055] It should be noted that the driving decision information can be predicted for each candidate tracking device respectively, or can be predicted by the parking lot platform. If the driving decision score of each candidate tracking device is predicted for each candidate tracking device, the driving decision score can be sent to other candidate tracking devices, for example, sent by the way of Bluetooth broadcast, WiFi, 4G, 5G network. If the driving decision score is predicted by the parking lot platform, the driving decision score can be sent to each candidate tracking device respectively for comparison, or the parking lot platform can directly compare to determine the driving decision score with the highest confidence, and send the tracking instruction to the tracking device corresponding to the driving decision score with the highest confidence, so as to relay track the target vehicle.

[0056] In the embodiment of the application, a vehicle tracking method is provided, including: receiving first vehicle information sent by a current tracking device; when detecting that a target vehicle drives to a target position, obtaining second vehicle information; predicting driving decision information of the target vehicle based on the first vehicle information and the second vehicle information; and determining a target tracking device based on the driving decision information, so as to track the target vehicle through the target tracking device. In the embodiment of the application, the network prediction capability is used to predict multi-dimensional vehicle information to obtain the driving decision information of the vehicle, that is, to determine the driving intention of the vehicle, so as to call the corresponding tracking device to track the vehicle. The relay tracking capability across tracking devices can be realized, the accuracy and real-time performance of vehicle monitoring can be improved, and the user experience can be improved.

[0057] Referring to Figure 5 In an embodiment of the application, the candidate tracking devices include a plurality of candidate tracking devices, and determining the target tracking device based on the driving decision information includes:

[0058] In step S210, each candidate tracking device sends the driving decision information predicted by itself to other candidate tracking devices, so that each candidate tracking device obtains the driving decision information corresponding to all candidate tracking devices.

[0059] In step S220, the driving decision information corresponding to all candidate tracking devices is compared to select the driving decision information with the highest confidence.

[0060] In step S230, the target tracking device is determined based on the driving decision information with the highest confidence.

[0061] Optionally, each candidate tracking device can predict the respective driving decision information, which can include a driving decision score, based on the collected second vehicle information and the received first vehicle information. Then the driving decision information is sent to other candidate tracking devices through Bluetooth, WiFi, 4G, 5G, etc. So that each candidate tracking device can obtain the driving decision scores predicted by all candidate tracking devices, and then compare them to determine the driving decision information with the highest confidence, i.e. the driving decision score with the highest score, and the candidate tracking device that predicts the highest driving decision score is taken as the target tracking device.

[0062] Optionally, the parking lot platform can obtain the first vehicle information and the second vehicle information collected by the tracking device and the candidate tracking device, predict the driving decision information of each candidate tracking device, such as the driving decision score, and compare them to obtain the driving decision information with the highest driving decision score as the driving decision information with the highest confidence, and the candidate tracking device that predicts the highest driving decision score is taken as the target tracking device.

[0063] Referring to Figure 6 In an embodiment of the present application, the second vehicle information includes multiple dimension information, and the driving decision information of the target vehicle is predicted, including:

[0064] In step S131, the score corresponding to each dimension information is determined;

[0065] In step S132, a corresponding weight value is assigned to each score, and the weight value is obtained from the pre-trained prediction model;

[0066] In step S133, the driving decision score is determined based on the weighted sum of each score, as the driving decision information.

[0067] The dimension information can include at least two of the license plate number, the license plate angle, the vehicle head direction, the vehicle tail direction, the vehicle body direction, the vehicle logo, the vehicle body color, and the vehicle re-identification result. It can be understood that the above-mentioned two dimension information can be combined or replaced by other dimension information according to actual conditions, which is not limited in the present application.

[0068] Optionally, taking the above-mentioned 8 dimension information as an example of model input information, the 8 dimension information is preprocessed, such as data cleaning, removing invalid values, and uniform format, and the key features can be extracted and input into the prediction model based on the training, to predict a score for each dimension information, and the weight obtained through the trained model is used to weight and sum the scores corresponding to each dimension information, thereby obtaining the driving decision score of each candidate tracking device.

[0069] Exemplarily, x1, x2, x3,..., x8 represent scores of 8-dimensional information, which can range from 0 to 1.0, and the weights can be represented as a, b, c, d, e, f, g, and h, which can range from 0 to 1.0. Then, the driving decision score score = ax1 + bx2 + cx3 + dx4 + ex5 + fx6 + gx7 + hx8, the score is set to 1.0 in the training phase, and the score range in prediction can be 0 to 1.0. The weights can be obtained by continuous training and optimization of the trained model in the training phase, and can be continuously updated.

[0070] The prediction model can be a pre-trained BP neural network, a decision tree, a random forest, or the like.

[0071] In an embodiment of the present application, the score corresponding to each dimension information is determined, including:

[0072] A license plate similarity score is obtained based on the similarity between the license plate number of the target vehicle identified by the candidate tracking device and the license plate number of the target vehicle identified by the tracking device.

[0073] A license plate angle score is obtained based on the angle between the license plate direction of the target vehicle and the first horizontal line.

[0074] A vehicle head direction score is obtained based on the angle between the vehicle head direction of the target vehicle and the second horizontal line.

[0075] A vehicle tail direction score is obtained based on the angle between the vehicle tail direction of the target vehicle and the third horizontal line.

[0076] A vehicle body direction score is obtained based on the angle between the vehicle body direction of the target vehicle and the fourth horizontal line.

[0077] A logo similarity score is obtained based on the similarity between the logo of the target vehicle identified by the candidate tracking device and the logo of the target vehicle identified by the tracking device.

[0078] A vehicle body color similarity score is obtained based on the similarity between the vehicle body color of the target vehicle identified by the candidate tracking device and the vehicle body color of the target vehicle identified by the tracking device.

[0079] A vehicle re-identification score is obtained based on the comparison result of the vehicle re-identification result of the target vehicle identified by the candidate tracking device and the vehicle re-identification result of the target vehicle identified by the tracking device.

[0080] Optionally, target vehicle detection can be performed on the driving video data of the target vehicle collected by the tracking device and the candidate tracking device respectively, and a license plate detection frame and a vehicle detection frame of the target vehicle are generated. Based on the license plate detection frame, character segmentation can be performed, and the segmented characters can be recognized to determine the license plate number of the target vehicle. Then, the two recognized license plate numbers are compared to determine whether they are exactly the same. If yes, the score is 1.0, otherwise, the score is determined according to the similarity, for example, if the similarity is 90%, the score is 0.9. For the framed license plate detection frame, the included angle between the license plate detection frame and a first horizontal line can be determined. The first horizontal line can be a virtual line pre-set in the current video frame. The smaller the included angle, the higher the score, for example, if the included angle is 10 degrees, the score is 0.8. For the vehicle head direction, the included angle between the vehicle head direction and a second horizontal line can be determined. Specifically, a pixel coordinate system can be established to determine the inclination angle of the vehicle head direction in the coordinate system, thereby determining the included angle. The smaller the angle, the higher the score, for example, if the included angle is 15 degrees, the score is 0.7. Similarly, the vehicle tail direction can also be determined based on the pixel coordinate system, and the smaller the angle, the higher the score. For the vehicle body direction, three-dimensional SLAM laser scanner (3D box) can be used to output real-time environmental three-dimensional laser point cloud data to determine the vehicle detection frame, or the chassis regression technology can be used to determine the vehicle detection frame, and then the included angle between the vehicle detection frame and the horizontal line is determined. It can also be determined by establishing a pixel coordinate system to determine the inclination angle, thereby obtaining the score. The smaller the angle, the higher the score. For the vehicle logo, the vehicle logos collected by the tracking device and the candidate tracking device can be compared to determine whether they are the same. If yes, the score is 1, otherwise, the score is 0. For the vehicle body color, the vehicle body colors collected by the tracking device and the candidate tracking device can be compared to determine whether they are the same. If yes, the score is 1, otherwise, the score is 0. For the re-identification result, the re-identification results collected by the tracking device and the candidate tracking device can be compared to determine the similarity, for example, by using the cosine similarity algorithm. The higher the similarity, the higher the score.

[0081] In the above manner, the scores of the license plate number, the license plate angle, the vehicle head direction, the vehicle tail direction, the vehicle body direction, the vehicle logo, the vehicle body color, and the vehicle re-identification feature can be calculated, i.e., data1, data2, data3, data4, data5, data6, data7, and data8. For example, if the above eight scores of the candidate tracking device A are 1, 1, 0, 1, 1, 1, 1, and 1 respectively, it can be considered that the target vehicle is driving in the direction of the candidate tracking device A. It should be noted that when five of the above eight dimensions are 1, it can be considered that the target vehicle has entered the tracking range of the candidate tracking device.

[0082] It can be understood that the license plate number information can be assigned the highest weight for different scores, and the head direction score and the tail direction score can be mutually exclusive.

[0083] In an embodiment of the present application, the topology relationship graph between tracking devices is constructed in the following way:

[0084] Determine the road type of the driving road;

[0085] Based on the road type, determine the candidate tracking device associated with each tracking device;

[0086] Establish the topology relationship graph between each tracking device and its associated candidate tracking device.

[0087] Optionally, different road types include different driving directions, so they can correspond to different numbers, and the candidate tracking devices in different settings, so the candidate tracking device associated with or adjacent to each tracking device can be determined based on the road type, and the topology relationship graph between the tracking device and its associated or adjacent candidate tracking device is established.

[0088] For example, see Figure 7 , a distribution scenario diagram of tracking devices and candidate tracking devices when the road type is a straight road is provided, wherein the target vehicle left side can be tracking device A, and the right side can be tracking device B and tracking device C, so the topology relationship graph is A-B-C, and the arrow direction can represent the vehicle driving direction, and the vehicle drives from tracking device A to its adjacent tracking device B, so the candidate tracking device is tracking device B, which relays tracking device A for relay tracking.

[0089] See Figure 8 , a distribution scenario diagram of tracking devices and candidate tracking devices when the road type is an L-shaped road is provided, wherein the target vehicle left side can be tracking device A, and the right side can be tracking device B, tracking device C and tracking device D, so the topology relationship can be A-B-C-D, and the arrow direction can represent the vehicle driving direction, and the target vehicle drives from tracking device A to its adjacent tracking device B, and then drives from tracking device B to tracking device C and tracking device D, so the candidate tracking device is tracking device B, which can relay the previous tracking device for relay tracking in the order of tracking device B, tracking device C and tracking device D.

[0090] See Figure 9, provides a distribution scene graph of tracking devices and candidate tracking devices when the road type is a T-shaped road, wherein the target vehicle left side can be tracking device A, the right side can be tracking device B, tracking device C and tracking device D, and the vertical direction road is tracking device E and tracking device F, so the topology relationship can be A-B-C-D, A-B-E-F, the arrow direction can represent the vehicle driving direction, the target vehicle drives from tracking device A to tracking device B, tracking device C and tracking device D, or the target vehicle drives from tracking device A to tracking device B, tracking device E and tracking device F. Therefore, the candidate tracking device can be tracking device B, tracking device C and tracking device E, when the target vehicle drives in the tracking area, such as the position of the dashed line in the figure, the vehicle information can be collected through tracking device B, tracking device C and tracking device E, and the driving decision score of tracking device B and tracking device C and tracking device E is predicted, for example, 1.0, 0.8, 0.5 respectively, 1.0>0.8>0.5, so tracking device B is considered as the target tracking device, which will relay tracking device A for tracking, when the target vehicle passes through tracking device B, it can be predicted again whether it drives to the direction of tracking device C or tracking device E, if it is predicted to drive to the direction of tracking device E, tracking device E can be used for relay tracking.

[0091] Referring to Figure 10, a distribution scenario of tracking devices and candidate tracking devices when the road type is an intersection is provided, wherein the tracking device A can be on the left side of the target vehicle, the tracking device B, the tracking device C and the tracking device D can be on the right side of the target vehicle, the tracking device E and the tracking device F can be above the vertical direction road, and the tracking device G and the tracking device H can be below the vertical direction road, so the topology relationship can be A-B-C-D, A-B-E-F, A-B-G-H, the arrow direction can represent the driving direction of the vehicle, the target vehicle drives from the tracking device A to the tracking device B, the tracking device C and the tracking device D, or the target vehicle drives from the tracking device A to the tracking device B, the tracking device E and the tracking device F, or the target vehicle drives from the tracking device A to the tracking device G and the tracking device H. Therefore, the candidate tracking devices can be the tracking device B, the tracking device C, the tracking device E and the tracking device G, when the target vehicle drives in the tracking area, such as the position of the dashed line in the figure, the vehicle information can be collected through the tracking device B, the tracking device C, the tracking device E and the tracking device G, and the driving decision scores of the tracking device B and the tracking device C, the tracking device E and the tracking device G are obtained through prediction, for example, 1.0, 0.8, 0.5, 0.3 respectively, 1.0>0.8>0.5>0.3, so the tracking device B is considered as the target tracking device, which will relay the tracking of the tracking device A, when the target vehicle passes through the tracking device B, it can be predicted again whether it drives to the direction of the tracking device C, the tracking device E or the tracking device G, if it is predicted to drive to the direction of the tracking device G, the tracking device G can be used for relay tracking.

[0092] Referring to Figure 11 In an embodiment of the present application, when it is detected that the target vehicle drives to the target position, the second vehicle information is obtained, including:

[0093] In step S121, when the target vehicle drives to the tracking area of the candidate tracking device, the running video data of the target vehicle in the tracking area is obtained, and a virtual identification line is arranged in the tracking area in the running video data;

[0094] In step S122, it is determined whether the target boundary line of the vehicle detection frame of the target vehicle coincides with the virtual identification line;

[0095] In step S123, if yes, the second vehicle information of the target vehicle is obtained.

[0096] Optionally, a virtual identification line can be configured in the video data collected by the tracking device, and a vehicle detection frame is framed for the detected target vehicle, and then it is determined whether the target boundary line of the vehicle detection frame collides with the virtual identification line, that is, whether the target boundary line coincides with the virtual identification line, if so, it indicates that the target vehicle travels to the tracking area, at this time, the second vehicle information of the target vehicle collected by the candidate tracking device can be triggered.

[0097] The second vehicle information can include multi-dimensional information such as license plate number, vehicle ID, vehicle logo, vehicle body color, vehicle model, vehicle type, vehicle driving direction, whether the photographed is the front or the tail of the vehicle, etc.

[0098] In the embodiments of the present application, the multi-dimensional vehicle information is predicted by the network prediction capability to obtain the driving decision information of the vehicle, that is, to determine the driving intention of the vehicle, so as to call the corresponding tracking device to track the vehicle, which can realize the relay tracking capability across tracking devices, improve the accuracy and real-time performance of vehicle monitoring, and improve the user experience.

[0099] It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0100] In an embodiment, a vehicle tracking device is provided, which is applied to a candidate tracking device determined by the vehicle driving direction of a target vehicle collected by a current tracking device and a pre-constructed inter-tracking device topology relationship graph. The vehicle tracking device corresponds to the vehicle tracking method in the above embodiments. As shown in the figure, the vehicle tracking device includes a first vehicle information receiving unit 10, a second vehicle information obtaining unit 20, a driving decision information prediction unit 30, and a vehicle tracking unit 40. The detailed description of each functional module is as follows: Figure 12

[0101] The first vehicle information obtaining unit 10 is configured to receive the first vehicle information of the target vehicle sent by the current tracking device;

[0102] The second vehicle information obtaining unit 20 is configured to obtain the second vehicle information when the target vehicle travels to the target position;

[0103] The driving decision information prediction unit 30 is configured to predict the driving decision information of the target vehicle based on the first vehicle information and the second vehicle information, and the driving decision information represents the driving intention of the target vehicle;

[0104] The vehicle tracking unit 40 is configured to determine a target tracking device based on the driving decision information, so as to track the target vehicle through the target tracking device.

[0105] ​In an embodiment of the present application, the vehicle tracking unit 40 is further configured to:

[0106] The candidate tracking devices include a plurality of tracking devices, and the target tracking device is determined based on the driving decision information.

[0107] Each candidate tracking device sends the respective predicted driving decision information to other candidate tracking devices, so that each candidate tracking device obtains the driving decision information corresponding to all candidate tracking devices.

[0108] The driving decision information corresponding to all candidate tracking devices is compared to select the driving decision information with the highest confidence.

[0109] The target tracking device is determined based on the driving decision information with the highest confidence.

[0110] In an embodiment of the present application, the second vehicle information includes a plurality of dimension information, and the driving decision information prediction unit 30 is further configured to:

[0111] The score corresponding to each dimension information is determined.

[0112] A corresponding weight value is assigned to each score, and the weight value is obtained from a pre-trained prediction model.

[0113] The driving decision score is determined based on the weighted sum of each score, and the driving decision score is used as the driving decision information.

[0114] In an embodiment of the present application, the dimension information includes at least two of the license plate number, the license plate angle, the vehicle head direction, the vehicle tail direction, the vehicle body direction, the vehicle logo, the vehicle body color, and the vehicle re-identification result.

[0115] In an embodiment of the present application, the driving decision information prediction unit 30 is further configured to:

[0116] The license plate similarity score is obtained based on the similarity between the license plate number of the target vehicle recognized by the candidate tracking device and the license plate number of the target vehicle recognized by the tracking device.

[0117] The license plate angle score is obtained based on the included angle between the license plate direction of the target vehicle and the first horizontal line.

[0118] The vehicle head direction score is obtained based on the included angle between the vehicle head direction of the target vehicle and the second horizontal line.

[0119] The vehicle tail direction score is obtained based on the included angle between the vehicle tail direction of the target vehicle and the third horizontal line.

[0120] The vehicle body direction score is obtained based on the included angle between the vehicle body direction of the target vehicle and the fourth horizontal line.

[0121] a logo similarity score is obtained based on a similarity between a logo of the target vehicle identified by the candidate tracking device and a logo of the target vehicle identified by the tracking device;

[0122] a body color similarity score is obtained based on a similarity between a body color of the target vehicle identified by the candidate tracking device and a body color of the target vehicle identified by the tracking device;

[0123] a vehicle re-identification score is obtained based on a comparison result between a vehicle re-identification result of the target vehicle identified by the candidate tracking device and a vehicle re-identification result of the target vehicle identified by the tracking device.

[0124] In an embodiment of the present application, the apparatus further comprises a topological relationship graph construction unit configured to:

[0125] determine a road type of the driving road;

[0126] determine, based on the road type, the candidate tracking device associated with each tracking device;

[0127] establish a topological relationship graph between each tracking device and the candidate tracking device associated therewith.

[0128] In an embodiment of the present application, the second vehicle information acquisition unit 30 is further configured to:

[0129] when the target vehicle drives into a tracking area of the candidate tracking device, acquire running video data of the target vehicle in the tracking area, wherein a virtual identification line is set in the tracking area in the running video data;

[0130] determine whether a target boundary line of a vehicle detection frame of the target vehicle coincides with the virtual identification line;

[0131] if yes, acquire the second vehicle information of the target vehicle.

[0132] In the embodiments of the present application, the multi-dimensional vehicle information is predicted by the network prediction capability to obtain driving decision information of the vehicle, i.e., to determine the driving intention of the vehicle, so as to call the corresponding tracking device to track the vehicle, which can realize relay tracking capability across tracking devices, improve the accuracy and real-time performance of vehicle monitoring, and improve user experience.

[0133] The specific limitations of the vehicle tracking apparatus can be referred to the limitations of the vehicle tracking method in the foregoing, which will not be repeated here. Each module in the vehicle tracking apparatus described above can be realized by software, hardware, and combinations thereof, in whole or in part. Each module described above can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to each module.

[0134] In one embodiment, a computer device is provided, which can be a terminal device, and an internal structure diagram of the computer device can be as shown in Figure 13 The computer device includes a processor, a memory, and a network interface connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a readable storage medium. The readable storage medium stores computer readable instructions. The network interface of the computer device is configured to communicate with an external terminal through a network connection. The computer readable instructions are executed by the processor to implement a vehicle tracking method. The readable storage medium provided in the embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0135] In the embodiments of the present application, a computer device is provided, which includes a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, and the processor executes the computer readable instructions to implement the steps of the vehicle tracking method described above.

[0136] In the embodiments of the present application, a readable storage medium is provided, which stores computer readable instructions, and the computer readable instructions are executed by the processor to implement the steps of the vehicle tracking method described above.

[0137] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by computer readable instructions instructing related hardware, and the computer readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium, and when executed, can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), etc.

[0138] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0139] The above examples are only used to illustrate the technical solutions of the present application, but not to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those ordinarily skilled in the art should understand: the technical solutions recorded in the foregoing examples can still be modified, or some technical features can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A vehicle tracking method, characterized in that: Applied to candidate tracking devices, the candidate tracking devices are determined by the vehicle driving direction of the target vehicle collected by the current tracking device and a pre-built topological relationship diagram between tracking devices, the method comprising: receiving first vehicle information of a target vehicle sent by the current tracking device; When detecting that the target vehicle has traveled to the target location, obtaining second vehicle information; Predicting driving decision information of the target vehicle based on the first vehicle information and the second vehicle information, wherein the driving decision information represents a driving intention of the target vehicle; Based on the driving decision information, a target tracking device is determined to track the target vehicle by the target tracking device, wherein the candidate tracking devices include a plurality of candidate tracking devices, and the determining of the target tracking device based on the driving decision information includes: Each candidate tracking device sends the driving decision information predicted by itself to other candidate tracking devices respectively, so that each candidate tracking device obtains the driving decision information corresponding to all candidate tracking devices; Comparing the driving decision information corresponding to all the candidate tracking devices to select the driving decision information with the highest confidence; The target tracking device is determined based on the driving decision information with the highest confidence.

2. The vehicle tracking method according to claim 1, wherein: The second vehicle information includes information of multiple dimensions, and the predicted driving decision information of the target vehicle includes: Determine the score corresponding to each dimension information; Assigning a corresponding weight value to each of the scores, wherein the weight value is obtained from a pre-trained prediction model; Based on the weighted sum of each of the scores, a driving decision score is determined as the driving decision information.

3. The vehicle tracking method according to claim 2, wherein: The dimensional information includes at least two of the license plate number, license plate angle, vehicle front direction, vehicle rear direction, vehicle body direction, vehicle logo, vehicle body color, and vehicle re-identification result.

4. The vehicle tracking method according to claim 3, wherein: Determining the score corresponding to each dimension information includes: Obtaining a license plate similarity score based on the similarity between the license plate number of the target vehicle identified by the candidate tracking device and the license plate number of the target vehicle identified by the current tracking device; Obtaining a license plate angle score based on an angle between the license plate direction of the target vehicle and the first horizontal line; Obtaining a vehicle head direction score based on an angle between the vehicle head direction and the second horizontal line; Obtaining a rear direction score based on an angle between the rear direction of the target vehicle and the third horizontal line; Obtaining a body direction score based on an angle between the body direction of the target vehicle and a fourth horizontal line; Obtaining a vehicle logo similarity score based on the similarity between the vehicle logo of the target vehicle identified by the candidate tracking device and the vehicle logo of the target vehicle identified by the current tracking device; Obtaining a body color similarity score based on a similarity between the body color of the target vehicle identified by the candidate tracking device and the body color of the target vehicle identified by the current tracking device; A vehicle re-identification score is obtained based on a comparison result of a vehicle re-identification result of the target vehicle identified by the candidate tracking device and a vehicle re-identification result of the target vehicle identified by the current tracking device.

5. The vehicle tracking method according to claim 1, wherein: The topological relationship diagram between the tracking devices is constructed in the following way: Determine the road type of the travel road; determining, based on the road type, a candidate tracking device associated with each current tracking device; A topological relationship diagram between each current tracking device and its associated candidate tracking devices is established.

6. The vehicle tracking method according to claim 1, wherein: When the target vehicle is detected to have traveled to the target location, obtaining second vehicle information includes: When the target vehicle travels into the tracking area of ​​the candidate tracking device, obtaining running video data of the target vehicle in the tracking area, wherein a virtual identification line is set in the tracking area in the running video data; determining whether a target boundary line of a vehicle detection frame of the target vehicle coincides with the virtual identification line; If so, obtain the second vehicle information of the target vehicle.

7. A vehicle tracking device, characterized in that: Applied to candidate tracking devices, the candidate tracking devices are determined by the vehicle driving direction of the target vehicle collected by the current tracking device and a pre-built topological relationship diagram between tracking devices, the device comprising: A first vehicle information receiving unit, configured to receive first vehicle information of a target vehicle collected by the current tracking device; A second vehicle information acquiring unit is configured to acquire second vehicle information when detecting that the target vehicle has traveled to a target location; a driving decision information prediction unit, configured to predict driving decision information of the target vehicle based on the first vehicle information and the second vehicle information; a vehicle tracking unit, configured to determine a target tracking device based on the driving decision information, so as to track the target vehicle through the target tracking device; The candidate tracking devices include a plurality of candidate tracking devices, and the vehicle tracking unit is further configured to: Each candidate tracking device sends the driving decision information predicted by itself to other candidate tracking devices respectively, so that each candidate tracking device obtains the driving decision information corresponding to all candidate tracking devices; Comparing the driving decision information corresponding to all the candidate tracking devices to select the driving decision information with the highest confidence; The target tracking device is determined based on the driving decision information with the highest confidence.

8. A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein: When the processor executes the computer-readable instructions, the steps of the vehicle tracking method according to any one of claims 1 to 6 are implemented.

9. A readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, the steps of the vehicle tracking method according to any one of claims 1 to 6 are implemented.

Images