A method and related device for judging false license plates based on license plate trajectories

Abstract

An embodiment of the present application discloses a method for judging a fake license plate, including: obtaining license plate data; determining a first license plate trajectory according to the license plate data; calculating a first trajectory feature value according to the first license plate trajectory; judging whether a preset number of the first trajectory feature values are within the corresponding feature range, where the feature range is the feature value range of the non-fake license plate trajectory obtained by training; if not satisfied, determining the license plate corresponding to the first license plate trajectory as a fake license plate, and the fake license plate is a license plate obtained by forging and imitating a license plate that can pass through the gate.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of video surveillance, and in particular to a method for judging fake license plates based on license plate trajectories and related devices. Background Art

[0002] In order to facilitate parking lot management, license plate recognition devices used to identify license plates have been widely used at the entrances and exits of parking lots. Currently, the mainstream license plate recognition devices are all equipped with cameras to collect data, and the gates are opened and closed by recognizing the license plates of the vehicles collected. After the vehicle enters and exits the parking lot, the parking lot charges according to the parking duration of the vehicle.

[0003] However, some vehicle owners will use fake license plates to attack in order to avoid payment, such as mobile phone license plates, paper license plates, and fabricated license plates, etc. By placing the fake license plate in front of the installed camera, when the license plate recognition device recognizes a fake license plate that imitates a real license plate that can pass through the gate, the gate will be opened, which does not meet the user's expectations. How to judge the authenticity of a license plate is an important issue. Summary of the Invention

[0004] The embodiments of the present application provide a method for judging fake license plates based on license plate trajectories and related devices.

[0005] A method for judging fake license plates based on license plate trajectories includes:

[0006] Obtain license plate data;

[0007] Determine a first license plate trajectory according to the license plate data;

[0008] Calculate a first trajectory feature value according to the first license plate trajectory;

[0009] Judge whether a preset number of the first trajectory feature values are within the corresponding feature range, where the feature range is the feature value range of non-fake license plate trajectories obtained through training;

[0010] If not, determine the license plate corresponding to the first license plate trajectory as a fake license plate, where the fake license plate is a license plate obtained by forging and imitating a license plate that can pass through the gate.

[0011] Optionally, determining a first license plate trajectory according to the license plate data includes:

[0012] Obtain the license plate frame of each frame according to the license plate data;

[0013] Judge whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same;

[0014] If so, extract the license plate frame of the previous frame in every two adjacent frames with the same license plate number;

[0015] Determine the first license plate trajectory based on the extracted license plate frame.

[0016] Optionally, after determining whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same and before determining the first license plate trajectory based on the extracted license plate frame, the method further includes:

[0017] If not, determine whether there is a situation where four or more consecutive digits are the same among the license plate numbers of every two adjacent frames with different license plate numbers;

[0018] If there is, extract the license plate frame of the previous frame among every two adjacent frames with four or more consecutive identical digits in the license plate number.

[0019] Optionally, after determining whether there is a situation where four or more consecutive digits are the same among the license plate numbers of every two adjacent frames with different license plate numbers and before determining the first license plate trajectory based on the extracted license plate frame, the method further includes:

[0020] If not, calculate the intersection over union (IoU) of every two adjacent frames with license plate numbers that do not meet the condition of having four or more consecutive identical digits;

[0021] Determine whether the intersection over union is greater than a preset threshold;

[0022] If it is greater, extract the license plate frame of the previous frame among every two adjacent frames with license plate numbers that do not meet the condition of having four or more consecutive identical digits.

[0023] Optionally, after the license plate trajectory is a pseudo license plate trajectory, the method further includes:

[0024] Upload the warning information and the license plate data to the server.

[0025] Optionally, before determining whether a preset number of the first trajectory feature values are within the corresponding feature ranges, the method further includes:

[0026] Obtain training data;

[0027] Determine a second license plate trajectory and a third license plate trajectory based on the training data, where the second license plate trajectory is multiple non - pseudo license plate trajectories in the training data, and the third license plate trajectory is multiple pseudo license plate trajectories in the training data;

[0028] Calculate a second trajectory feature value based on the second license plate trajectory and calculate a third trajectory feature value based on the third license plate trajectory;

[0029] Perform screening based on the second trajectory feature value and the third trajectory feature value to obtain the feature range.

[0030] Optionally, the feature range includes the average speed feature range of the trajectory in the x-axis direction, the variance feature range of the trajectory in the x-axis direction, the average speed feature range of the trajectory in the y-axis direction, the variance feature range of the trajectory in the y-axis direction, and the slope feature range of the trajectory.

[0031] A pseudo license plate judgment device includes:

[0032] An acquisition unit for acquiring license plate data;

[0033] A determination unit for determining a first license plate trajectory according to the license plate data;

[0034] A calculation unit for calculating a first trajectory eigenvalue according to the first license plate trajectory;

[0035] A judgment unit for judging whether a preset number of the first trajectory eigenvalues are within the corresponding feature ranges, where the feature ranges are the eigenvalue ranges of the non-pseudo license plate trajectories obtained through training;

[0036] The determination unit is further configured to, when the preset number is not satisfied, determine the license plate corresponding to the first license plate trajectory as a pseudo license plate, where the pseudo license plate is a license plate obtained by forging and imitating a license plate that can pass through the barrier gate.

[0037] A pseudo license plate judgment device includes:

[0038] A central processing unit, a memory, and an input / output interface;

[0039] The memory is a transient storage memory or a persistent storage memory;

[0040] The central processing unit is configured to communicate with the memory and execute the instruction operations in the memory to execute the foregoing method.

[0041] A computer-readable storage medium includes instructions that, when run on a computer, cause the computer to execute the foregoing method.

[0042] As can be seen from the above technical solutions, the embodiments of the present application have the following advantages:

[0043] After acquiring the license plate data, determine the first license plate trajectory, and then calculate the first trajectory eigenvalue. Then judge whether a preset number of the first trajectory eigenvalues are within the corresponding feature ranges. When the preset number is not satisfied, determine the license plate corresponding to the first license plate trajectory as a pseudo license plate. Based on the license plate trajectory, judge whether it is a pseudo license plate trajectory, that is, whether the license plate is a pseudo license plate, through the feature ranges of the non-pseudo license plate trajectories obtained through training, so as to accurately judge the authenticity of the license plate and meet the user's expectations. Description of the Drawings

[0044] Figure 1 Schematic diagram of an embodiment of the false license plate judgment method according to an embodiment of the present application;

[0045] Figure 2 Schematic diagram of another embodiment of the false license plate judgment method according to an embodiment of the present application;

[0046] Figure 3 Schematic diagram of an embodiment of the false license plate judgment device according to an embodiment of the present application;

[0047] Figure 4 Schematic diagram of another embodiment of the false license plate judgment device according to an embodiment of the present application. Detailed implementation manners

[0048] The embodiments of the present application provide a false license plate judgment method based on license plate trajectory and related devices.

[0049] The license plate can be said to be the "name" of the vehicle. By detecting the license plate, relevant information of the vehicle can be obtained. The existing license plate recognition devices at the entrance and exit of parking lots are attacked by false license plates of some vehicle owners to avoid tolls. To solve the above problems, the embodiments of the present application provide a method and related devices capable of judging false license plates.

[0050] The following describes the false license plate judgment method based on license plate trajectory and related devices according to the embodiments of the present application.

[0051] Please refer to Figure 1 , an embodiment of the false license plate judgment method according to an embodiment of the present application includes:

[0052] 101. Obtain license plate data;

[0053] Obtain license plate data through a camera device. Among them, the license plate data is video data of the time period from when the license plate first appears in the recognition area to when it disappears in the recognition area, that is, the video data of the license plate moving in the recognition area of the camera device. The most common camera device is a camera, which is used to monitor the entrance and exit of the parking lot and collect license plate data. When the camera body is offset due to external reasons, the swing of the camera can be controlled to capture clear and complete images for subsequent operations.

[0054] 102. Determine the first license plate trajectory according to the license plate data;

[0055] After obtaining the license plate data, determine the first license plate trajectory according to the license plate data. Specifically, after obtaining the video data of the license plate, divide the video data into data of different frames, and then identify the license plate data of each frame to obtain the license plate number, license plate frame, and the position of the license plate frame of the license plate. According to the obtained results for judgment and confirmation, the first license plate trajectory can be obtained. The first license plate trajectory is composed of the extracted license plate frames, which reflects the movement of the license plate in the recognition area of the camera during the video recording time.

[0056] 103. Calculate a first trajectory eigenvalue according to the first license plate trajectory;

[0057] Calculate a first trajectory eigenvalue according to the first license plate trajectory. After obtaining the first license plate trajectory, establish a plane coordinate system, mark the center point coordinates of the license plate frames forming the first license plate trajectory, and then calculate the relevant trajectory eigenvalues of the first license plate trajectory. Among them, the trajectory eigenvalues include values representing the characteristics of the trajectory such as the speed on the x-axis, the average speed on the x-axis, the variance on the x-axis, the speed on the y-axis, the average speed on the y-axis, the variance on the y-axis, the slope of the trajectory, the acceleration on the x-axis, and the acceleration on the y-axis.

[0058] 104. Determine whether a preset number of the first trajectory eigenvalues are within the corresponding eigenvalue ranges. If not, execute step 105; if so, execute step 106;

[0059] Determine whether a preset number of the first trajectory eigenvalues are within the corresponding eigenvalue ranges. Specifically, the preset number can be set according to requirements, such as set to 3, 5, etc. Among them, the eigenvalue range is the eigenvalue range of the non-fake license plate trajectory obtained through training. For example, assume that this step is to determine whether there are 3 or more first trajectory eigenvalues within the corresponding eigenvalue ranges. The eigenvalue range representing a non-fake license plate includes that the average speed eigenvalue range of the trajectory in the x-axis direction is 10 to 15, the variance eigenvalue range of the trajectory in the x-axis direction is 2 to 3, the average speed eigenvalue range of the trajectory in the y-axis direction is 11 to 16, the variance eigenvalue range of the trajectory in the y-axis direction is 3.5 to 5, and the slope eigenvalue range of the trajectory is 0.5 to 0.8. Assume that the first trajectory eigenvalues of the first license plate trajectory include the average speed on the x-axis is 11, the variance on the x-axis is 2.6, the average speed on the y-axis is 9, the variance on the y-axis is 5.1, and the slope of the trajectory is 0.6. After judgment, 3 of the first trajectory eigenvalues fall within the eigenvalue range, then it is determined that the first license plate trajectory is a non-fake license plate, that is, a genuine license plate.

[0060] 105. Determine the license plate corresponding to the first license plate trajectory as a fake license plate;

[0061] If there is no preset number of first trajectory eigenvalues within the corresponding eigenvalue ranges, then the license plate corresponding to the first license plate trajectory is determined to be a fake license plate. Here, a fake license plate is a license plate obtained by forging and imitating a license plate that can pass through the gate, which is the opposite concept of a genuine license plate.

[0062] 106. Determine the license plate corresponding to the first license plate trajectory as a non-fake license plate.

[0063] If there is a preset number of first trajectory eigenvalues within the corresponding eigenvalue ranges, then the license plate corresponding to the first license plate trajectory is determined to be a non-fake license plate.

[0064] In the embodiments of the present application, after obtaining license plate data, the first license plate trajectory is determined, and then the first trajectory eigenvalues are calculated. Then, it is determined whether a preset number of first trajectory eigenvalues are within the corresponding eigenvalue ranges. When the preset number is not satisfied, the license plate corresponding to the first license plate trajectory is determined to be a fake license plate. Based on the license plate trajectory, it is determined whether it is a fake license plate trajectory, that is, whether the license plate is a fake license plate, by using the eigenvalue range of the non-fake license plate trajectory trained. In this way, the authenticity of the license plate can be accurately judged, meeting the user's expectations.

[0065] Please refer to Figure 2 , another embodiment of the fake license plate judgment method in the embodiments of the present application includes:

[0066] 201. Obtain training data;

[0067] Obtain training data. Here, the training data includes non-fake license plate data and fake license plate data. The non-fake license plate data is obtained by cameras installed at the entrance and exit. The fake license plate data, including mobile phone license plates, paper license plates, and imitated license plates, is obtained by simulating the attack means that the vehicle owner may use. To ensure that the non-fake license plate data of real vehicles is representative, it is necessary to collect the passing data of multiple vehicles in multiple scenarios. The passing data of 100 vehicles' non-fake license plates can be collected to ensure the effectiveness of the data. In addition, to improve the accuracy, data can also be obtained during the day and at night respectively.

[0068] 202. Determine the second license plate trajectory and the third license plate trajectory according to the training data;

[0069] Determine the second license plate trajectory and the third license plate trajectory according to the training data, where the second license plate trajectory is multiple non-fake license plate trajectories in the training data, and the third license plate trajectory is multiple fake license plate trajectories in the training data. Specifically, the training data includes data of multiple license plates. After obtaining the training data, each license plate data in the training data is divided into data of different frames, and then the license plate number, license plate frame, and position of the license plate frame of each frame of license plate data are identified. Judgment and confirmation are performed according to the obtained results. The second license plate trajectory and the third license plate trajectory can be divided by performing a series of judgments on the license plate numbers and intersection over union ratios of adjacent two frames. For example, if the training data includes 100 non-fake license plate data and 100 fake license plate data, and one data forms one trajectory, then 100 true license plate trajectories and 100 fake license plate trajectories can be obtained.

[0070] 203. Calculate the second trajectory eigenvalue according to the second license plate trajectory, and calculate the third trajectory eigenvalue according to the third license plate trajectory;

[0071] Calculate the second trajectory eigenvalue and the third trajectory eigenvalue according to the second license plate trajectory and the third license plate trajectory. Calculate the trajectory eigenvalue of each true license plate trajectory and the trajectory eigenvalue of each fake license plate trajectory for subsequent operations. Specifically, after obtaining the trajectories, establish a plane coordinate system, mark the center point coordinates of the license plate frames corresponding to each license plate trajectory, and the coordinates (x1, y1), (x2, y2), (x3, y3), (x4, y4), etc. can be obtained. Then calculate the relevant trajectory eigenvalues of the license plate trajectory. Among them, the trajectory eigenvalues include values representing the characteristics of the trajectory such as the speed on the x-axis, the average speed on the x-axis, the variance on the x-axis, the speed on the y-axis, the average speed on the y-axis, the variance on the y-axis, the slope of the trajectory, the acceleration on the x-axis, and the acceleration on the y-axis. The speed, average speed, and variance on the x-axis, and the slope of the trajectory are described.

[0072] Calculate the speed of the license plate trajectory on the x-axis through the following formula:

[0073]

[0074] where v x is the speed between two certain frames, Δx is the speed difference between two certain frames, and ΔF is the frame interval number between two certain frames.

[0075] The average speed of the license plate trajectory on the x-axis can be obtained by taking the average value of multiple v x values.

[0076] Calculate the variance of the license plate trajectory on the x-axis through the following formula:

[0077]

[0078] Among them, s x is the variance on the x-axis, i represents the i-th coordinate, n represents the total number of coordinates, and v xi is the velocity of the i-th coordinate on the x-axis, and mean(v x ) represents the average velocity on the x-axis.

[0079] The calculation methods of the velocity and variance of the license plate trajectory on the y-axis are similar to the above content and will not be elaborated here.

[0080] The slope of the license plate trajectory is calculated by the following formula:

[0081]

[0082] k is the slope, Δx is the velocity difference on the x-axis, and Δy is the velocity difference on the y-axis.

[0083] 204. Screen according to the second trajectory eigenvalue and the third trajectory eigenvalue to obtain the feature range;

[0084] Screen according to the second trajectory eigenvalue and the third trajectory eigenvalue to obtain the feature range. Specifically, after obtaining multiple second trajectory eigenvalues and multiple third trajectory eigenvalues, divide the ranges respectively to obtain the ranges of multiple true license plate trajectory eigenvalues and the ranges of multiple false license plate trajectory eigenvalues. Set that the coincidence rate of the ranges of true license plate trajectory eigenvalues and false license plate trajectory eigenvalues cannot exceed a certain percentage, such as twenty percent. If it does not exceed, then determine the range of the corresponding true license plate trajectory eigenvalues as the feature range. For example, the range of the average velocity eigenvalue of the true license plate trajectory on the x-axis is 1 to 10, the range of the average velocity eigenvalue of the false license plate trajectory on the x-axis is 9 to 15, and the coincidence rate is ten percent. The range of the average velocity eigenvalue of the true license plate trajectory on the x-axis can be determined as the feature range.

[0085] Based on practice, the feature range includes the average velocity feature range of the trajectory in the x-axis direction, the variance feature range of the trajectory in the x-axis direction, the average velocity feature range of the trajectory in the y-axis direction, the variance feature range of the trajectory in the y-axis direction, and the slope feature range of the trajectory.

[0086] 205. Obtain license plate data;

[0087] Obtain license plate data through a camera device. Among them, the license plate data is the video data of the time period from when the license plate first appears in the recognition area to when it disappears in the recognition area, that is, the video data of the license plate moving in the recognition area of the camera device. The most common camera device is a camera, which is used to monitor the entrance and exit of a parking lot and collect license plate data. When the camera body is offset due to external reasons, the swing of the camera can be controlled to capture clear and complete images for subsequent operations.

[0088] 206. Obtain the license plate frame for each frame based on the license plate data;

[0089] After obtaining the license plate data, obtain the license plate frame for each frame based on the license plate data. Specifically, after obtaining the video data of the license plate, divide the video data into data for different frames, and then identify the license plate number, license plate frame, and the position of the license plate frame for each frame of the license plate data.

[0090] 207. Determine whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same. If so, execute step 208; if not, execute step 209;

[0091] Determine whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same. Specifically, that is, judge from the first frame to the second-to-last frame in the video data to detect whether the license plate numbers corresponding to the license plate frames of two adjacent frames are exactly the same. For example, first judge the 1st frame and the 2nd frame, then judge the 2nd frame and the 3rd frame, and so on. This will not be elaborated here.

[0092] 208. Extract the license plate frame of the previous frame in every two adjacent frames with the same license plate number;

[0093] Extract the license plate frame of the previous frame in every two adjacent frames with the same license plate number. Specifically, in every two adjacent frames, if the license plate numbers are the same, extract the license plate frame of the previous frame. For example, if the license plate numbers of the 1st frame and the 2nd frame are the same, then extract the license plate frame of the 1st frame.

[0094] 209. Determine whether there is a situation where four or more consecutive digits are the same in the license plate numbers of every two adjacent frames with different license plate numbers. If so, execute step 210; if not, execute step 211;

[0095] Determine whether there is a situation where four or more consecutive digits are the same in the license plate numbers of every two adjacent frames with different license plate numbers. Specifically, in every two adjacent frames, if the license plate numbers are not the same, then determine whether there is a situation where four or more consecutive digits are the same.

[0096] 210. Extract the license plate frame of the previous frame in every two adjacent frames with four or more consecutive identical digits in the license plate number;

[0097] Extract the license plate frame of the previous frame in every two adjacent frames where four or more consecutive digits of the license plate number are the same. Specifically, in every two adjacent frames, if the recognized license plate numbers are not completely the same but have four or more consecutive digits in common, extract the license plate frame of the previous frame. For example, if the license plate number of the second frame is recognized as 123456 and the license plate number of the third frame is recognized as 123465, and both have 1234, then extract the license plate frame of the second frame.

[0098] 211. Calculate the intersection over union (IoU) of every two adjacent frames where the license plate number does not satisfy the condition of having four or more consecutive digits the same;

[0099] Calculate the intersection over union (IoU) of every two adjacent frames where the license plate number does not satisfy the condition of having four or more consecutive digits the same. Specifically, in every two adjacent frames, if the recognized license plate numbers do not meet the condition of having four or more consecutive digits the same, calculate their IoU. For example, if the license plate number of the third frame is recognized as 123465 and the license plate number of the fourth frame is recognized as 123654, and only 123 is the same, then calculate the IoU of the two frames. Here, the intersection over union refers to the overlap rate of two frames, that is, the ratio of the intersection to the union of the two frames, which is commonly used in object detection.

[0100] 212. Determine whether the intersection over union is greater than a preset threshold. If so, execute step 213; if not, execute step 207;

[0101] After calculating the intersection over union, determine whether the intersection over union is greater than a preset threshold. Specifically, the preset threshold can be set according to experience and requirements, and is generally set to 0.5.

[0102] 213. Extract the license plate frame of the previous frame in every two adjacent frames where the license plate number does not satisfy the condition of having four or more consecutive digits the same;

[0103] Extract the license plate frame of the previous frame in every two adjacent frames where the license plate number does not satisfy the condition of having four or more consecutive digits the same. Specifically, in every two adjacent frames, if the license plate number does not satisfy the condition of having four or more consecutive digits the same and the intersection over union is greater than the preset threshold, extract the license plate frame of the previous frame. For example, if the license plate number of the third frame is recognized as 123465 and the license plate number of the fourth frame is recognized as 123654, and the IoU of the two license plate frames is 0.6, which is greater than 0.5, then extract the license plate frame of the third frame.

[0104] 214. Determine the first license plate trajectory based on the extracted license plate frames;

[0105] Determine the first license plate trajectory based on the extracted license plate frame. Specifically, extract the license plate frames in the previous frame among two adjacent frames that meet the conditions of having the same number of digits in the license plate number, having four or more consecutive digits in the license plate number that are the same, and having an intersection over union greater than a preset threshold. The license plate frame and its location can determine the first license plate trajectory.

[0106] 215. Calculate a first trajectory eigenvalue based on the first license plate trajectory;

[0107] Calculate a first trajectory eigenvalue based on the first license plate trajectory. After obtaining the first license plate trajectory, establish a plane coordinate system, mark the center point coordinates of the license plate frames forming the first license plate trajectory, and then calculate the relevant trajectory eigenvalues of the first license plate trajectory. Among them, the trajectory eigenvalues include values representing the characteristics of the trajectory such as the velocity on the x-axis, the average velocity on the x-axis, the variance on the x-axis, the velocity on the y-axis, the average velocity on the y-axis, the variance on the y-axis, the slope of the trajectory, the acceleration on the x-axis, and the acceleration on the y-axis. The specific calculation formulas can refer to the content related to step 203 and will not be elaborated here.

[0108] 216. Determine whether a preset number of the first trajectory eigenvalues are within the corresponding characteristic ranges. If not, execute step 217; if so, execute step 218;

[0109] Determine whether a preset number of the first trajectory eigenvalues are within the corresponding characteristic ranges. Specifically, the preset number can be set according to requirements, such as set to 3, 5, etc. Among them, the characteristic ranges are the eigenvalue ranges of the non-fake license plate trajectories obtained through training. For example, assume that in this step, it is determined whether there are 3 or more first trajectory eigenvalues within the corresponding characteristic ranges. The characteristic ranges indicating non-fake license plates include that the average velocity characteristic range of the trajectory in the x-axis direction is 10 to 15, the variance characteristic range of the trajectory in the x-axis direction is 2 to 3, the average velocity characteristic range of the trajectory in the y-axis direction is 11 to 16, the variance characteristic range of the trajectory in the y-axis direction is 3.5 to 5, and the slope characteristic range of the trajectory is 0.5 to 0.8. Assume that the first trajectory eigenvalues of the first license plate trajectory include the average velocity on the x-axis is 11, the variance on the x-axis is 2.6, the average velocity on the y-axis is 9, the variance on the y-axis is 5.1, and the slope of the trajectory is 0.6. After judgment, 3 of the first trajectory eigenvalues fall within the characteristic ranges, then determine that the first license plate trajectory is a non-fake license plate, that is, a genuine license plate.

[0110] 217. Determine the license plate corresponding to the first license plate trajectory as a fake license plate;

[0111] If there is no preset number of first trajectory eigenvalues within the corresponding eigenvalue ranges, then the license plate corresponding to the first license plate trajectory is determined to be a fake license plate. Here, a fake license plate is a license plate obtained by forging and imitating a license plate that can pass through the barrier gate, which is the opposite concept of a genuine license plate.

[0112] 218. Determine the license plate corresponding to the first license plate trajectory as a non-fake license plate;

[0113] If there is a preset number of first trajectory eigenvalues within the corresponding eigenvalue ranges, then the license plate corresponding to the first license plate trajectory is determined to be a non-fake license plate.

[0114] 219. Upload the warning information and the license plate data to the server.

[0115] After determining that the license plate in the license plate data is a fake license plate, upload the warning information and the license plate data to the server. The warning information can prompt the server operator to handle it, and the license plate data can be saved as evidence in the server. If a genuine license plate trajectory is misjudged as a fake license plate trajectory, the operator can learn this license plate trajectory through settings and update the eigenvalue ranges of the learned eigenvalues. By adding configuration items on the platform, the automatic update of the eigenvalue ranges can be achieved.

[0116] In this embodiment, first, the license plate trajectory is trained to obtain the eigenvalue ranges. After obtaining the license plate data, the first license plate trajectory is determined, and then the first trajectory eigenvalue is calculated. Then, it is judged whether a preset number of first trajectory eigenvalues are within the corresponding eigenvalue ranges. When the preset number is not satisfied, the license plate corresponding to the first license plate trajectory is determined to be a fake license plate. During the training process, three types of judgments are used to improve the accuracy of the judgment. In addition, based on the license plate trajectory, it is judged whether it is a fake license plate trajectory, that is, whether the license plate is a fake license plate, through the eigenvalue ranges of the non-fake license plate trajectories obtained by training. In this way, the authenticity of the license plate can be accurately judged. After judging a fake license plate, it can also be uploaded to the server for warning and evidence preservation, meeting the user's expectations.

[0117] Next, the fake license plate judgment device of the embodiment of the present application will be described. Please refer to Figure 3 , an embodiment of the fake license plate judgment device of the embodiment of the present application includes:

[0118] An acquisition unit 301, configured to acquire license plate data;

[0119] A determination unit 302, configured to determine a first license plate trajectory according to the license plate data;

[0120] A calculation unit 303, configured to calculate a first trajectory eigenvalue according to the first license plate trajectory;

[0121] A determination unit 304, configured to determine whether a preset number of the first trajectory feature values are within corresponding feature ranges, where the feature ranges are the feature value ranges of non-fake license plate trajectories obtained through training;

[0122] The determination unit 302 is further configured to, when the preset number is not satisfied, determine the license plate corresponding to the first license plate trajectory as a fake license plate, where the fake license plate is a license plate obtained by forging and imitating a license plate that can pass through a gate.

[0123] In an embodiment of the present application, after the acquisition unit 301 acquires license plate data, the determination unit 302 determines a first license plate trajectory, and then the calculation unit 303 calculates a first trajectory feature value. Then, the determination unit 304 determines whether a preset number of the first trajectory feature values are within the corresponding feature ranges. When the preset number is not satisfied, the license plate corresponding to the first license plate trajectory is determined as a fake license plate. Based on the license plate trajectory, it is determined whether it is a fake license plate trajectory, that is, whether the license plate is a fake license plate, through the feature range of the non-fake license plate trajectory obtained through training. In this way, the authenticity of the license plate can be accurately determined, meeting the user's expectations.

[0124] In this embodiment, the functions and processes executed by each unit in the fake license plate determination device are similar to those executed by the fake license plate determination device in the foregoing Figures 1 to 2 The functions and processes executed by the fake license plate determination device are not described herein again.

[0125] Figure 4 FIG. 15 is a schematic structural diagram of a fake license plate determination device provided by an embodiment of the present application. The fake license plate determination device 400 may include one or more central processing units (CPUs) 401 and a memory 405. One or more application programs or data are stored in the memory 405.

[0126] Among them, the memory 405 may be volatile storage or persistent storage. The program stored in the memory 405 may include one or more modules, and each module may include a series of instruction operations on the fake license plate determination device. Further, the central processor 401 may be configured to communicate with the memory 405 and execute a series of instruction operations in the memory 405 on the fake license plate determination device 400.

[0127] The fake license plate determination device 400 may further include one or more power supplies 402, one or more wired or wireless network interfaces 403, one or more input / output interfaces 404, and / or one or more operating systems, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0128] The central processing unit 401 can execute the operations performed by the pseudo license plate determination device in the foregoing Figures 1 to 2 embodiment shown, and the details are not described herein again.

[0129] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0130] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0131] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0132] In addition, the functional units in each embodiment of the present application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0133] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. And the foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, read-only memory), random access memories (RAM, random access memory), magnetic disks, or optical discs that can store program codes.

Claims

1. A method for judging false license plates based on license plate trajectories, characterized in that, Including: Obtain license plate data; Determine a first license plate trajectory according to the license plate data; Calculate a first trajectory eigenvalue according to the first license plate trajectory; Judge whether a preset number of the first trajectory eigenvalues are within corresponding eigenvalue ranges, where the eigenvalue ranges are eigenvalue ranges of non-counterfeit license plate trajectories obtained through training. Among them, the preset number is at least three, and the first trajectory eigenvalues include the average speed of the trajectory in the x-axis direction, the variance of the trajectory in the x-axis direction, the average speed of the trajectory in the y-axis direction, the variance of the trajectory in the y-axis direction, and the slope of the trajectory. The eigenvalue ranges include the average speed eigenvalue range of the trajectory in the x-axis direction, the variance eigenvalue range of the trajectory in the x-axis direction, the average speed eigenvalue range of the trajectory in the y-axis direction, the variance eigenvalue range of the trajectory in the y-axis direction, and the slope eigenvalue range of the trajectory; If not satisfied, determine the license plate corresponding to the first license plate trajectory as a counterfeit license plate, where the counterfeit license plate is a license plate forged by imitating a license plate that can pass through the barrier gate.

2. The pseudo license plate determination method according to claim 1, characterized in that, Determine a first license plate trajectory according to the license plate data, including: Obtain the license plate frame of each frame according to the license plate data; Judge whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same; If so, extract the license plate frame of the previous frame in every two adjacent frames with the same license plate number; Determine a first license plate trajectory according to the extracted license plate frames.

3. The method for judging a false license plate according to claim 2 is characterized in that, After judging whether the license plate numbers corresponding to the license plate frames of every two adjacent frames are the same and before determining a first license plate trajectory according to the extracted license plate frames, the method further includes: If not, judge whether there is a situation where four or more consecutive digits are the same among the license plate numbers of every two adjacent frames with different license plate numbers; If there is, extract the license plate frame of the previous frame in every two adjacent frames with four or more consecutive identical digits in the license plate numbers.

4. The pseudo license plate determination method according to claim 3, wherein After judging whether there is a situation where four or more consecutive digits are the same among the license plate numbers of every two adjacent frames with different license plate numbers and before determining a first license plate trajectory according to the extracted license plate frames, the method further includes: If not, calculate the intersection over union of every two adjacent frames with license plate numbers that do not satisfy four or more consecutive identical digits; Judge whether the intersection over union is greater than a preset threshold; If greater, extract the license plate frame of the previous frame in every two adjacent frames with license plate numbers that do not satisfy four or more consecutive identical digits.

5. The method for judging a false license plate according to claim 1, characterized in that After the license plate trajectory is a counterfeit license plate trajectory, the method further includes: Upload warning information and the license plate data to the server.

6. The method for judging a false license plate according to claim 1, characterized in that Before judging whether a preset number of the first trajectory eigenvalues are within corresponding eigenvalue ranges, the method further includes: Obtain training data; Determine a second license plate trajectory and a third license plate trajectory according to the training data, where the second license plate trajectory is multiple non-counterfeit license plate trajectories in the training data, and the third license plate trajectory is multiple counterfeit license plate trajectories in the training data; Calculate a second trajectory eigenvalue according to the second license plate trajectory, and calculate a third trajectory eigenvalue according to the third license plate trajectory; Perform screening according to the second trajectory eigenvalue and the third trajectory eigenvalue to obtain the eigenvalue range.

7. A pseudo license plate judgment device, characterized in that, Including: An acquisition unit for acquiring license plate data; A determination unit for determining a first license plate trajectory according to the license plate data; A calculation unit for calculating a first trajectory feature value according to the first license plate trajectory; A judgment unit for judging whether a preset number of the first trajectory feature values are within corresponding feature ranges, where the feature ranges are the feature value ranges of non-fake license plate trajectories obtained through training. Among them, the preset number is at least three, and the first trajectory feature values include the average speed of the trajectory in the x-axis direction, the variance of the trajectory in the x-axis direction, the average speed of the trajectory in the y-axis direction, the variance of the trajectory in the y-axis direction, and the slope of the trajectory. The feature ranges include the average speed feature range of the trajectory in the x-axis direction, the variance feature range of the trajectory in the x-axis direction, the average speed feature range of the trajectory in the y-axis direction, the variance feature range of the trajectory in the y-axis direction, and the slope feature range of the trajectory; The determination unit is further configured to, when the preset number is not satisfied, determine the license plate corresponding to the first license plate trajectory as a fake license plate, where the fake license plate is a license plate obtained by forging and imitating a license plate that can pass through the barrier gate.

8. A pseudo license plate determination device, characterized in that, Comprising: A central processing unit, a memory, and an input / output interface; The memory is a transient storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute the instruction operations in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Including instructions that, when run on a computer, cause the computer to execute the method according to any one of claims 1 to 6.

Images