WO2023088201A1 - 异常车辆的检测方法、装置及设备 - Google Patents

异常车辆的检测方法、装置及设备 Download PDF

Info

Publication number
WO2023088201A1
WO2023088201A1 PCT/CN2022/131621 CN2022131621W WO2023088201A1 WO 2023088201 A1 WO2023088201 A1 WO 2023088201A1 CN 2022131621 W CN2022131621 W CN 2022131621W WO 2023088201 A1 WO2023088201 A1 WO 2023088201A1
Authority
WO
WIPO (PCT)
Prior art keywords
point
time
location
data
adjacent
Prior art date
Application number
PCT/CN2022/131621
Other languages
English (en)
French (fr)
Inventor
陈钦
丁玲德
楼剑豪
张豪
陈泽群
Original Assignee
杭州海康威视数字技术股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 杭州海康威视数字技术股份有限公司 filed Critical 杭州海康威视数字技术股份有限公司
Publication of WO2023088201A1 publication Critical patent/WO2023088201A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection

Definitions

  • the present application relates to the field of intelligent transportation, in particular to a detection method, device and equipment for abnormal vehicles.
  • a vehicle with a set of license plates is called a set plate vehicle, and the set of license plates is forged with reference to the real license plate, so its license plate logo is the same as the real license plate.
  • the license plate vehicle is a kind of abnormal vehicle, therefore, it is necessary to detect and effectively manage the license plate vehicles driving on the road in time.
  • a large number of cameras (such as analog cameras or network cameras, etc.) are usually deployed, and images of vehicles driving on the road can be collected through these cameras, and the license plate identification of the vehicle can be analyzed based on the image.
  • the vehicle is confirmed as the same vehicle, so as to manage and regulate the road driving behavior of the vehicle.
  • the license plate identification of the real license plate of the normal vehicle is the same as that of the set license plate of the abnormal vehicle, it is possible to identify the normal vehicle and the abnormal vehicle with the same license plate identification as the same vehicle. Thus, if the management means for abnormal vehicles is applied to normal vehicles, management errors will be caused.
  • the present application provides a method for detecting an abnormal vehicle, the method comprising: if the target vehicle travels from a first position point to a second position point, obtaining a first time point when the target vehicle is at the first position point, the The target vehicle is at the second time point at the second position point, the number of target position points passed by the target vehicle from the first position point to the second position point; if the second time point is the same as the first time point If the difference between the first location point and the second location point is less than the minimum passing time between the first location point and the second location point, it is determined that the target vehicle is an abnormal vehicle; if the second time point and the first time point The difference between is not less than the minimum transit time, and the minimum number of location points between the first location point and the second location point is greater than the sum of the number of target location points and the preset number threshold, then It is determined that the target vehicle is an abnormal vehicle.
  • the present application provides a detection device for an abnormal vehicle, the device includes: an acquisition module, configured to acquire the first time point when the target vehicle is at the first position point if the target vehicle travels from the first position point to the second position point , the second time point when the target vehicle is at the second location point, the number of target location points passed by the target vehicle from the first location point to the second location point; the determination module is used to determine if the second time point point and the first time point is less than the minimum transit time between the first location point and the second location point, it is determined that the target vehicle is an abnormal vehicle; if the second time The difference between the point and the first time point is not less than the minimum travel time, and the minimum number of location points between the first location point and the second location point is greater than the number of target location points and The sum of the preset quantity thresholds determines that the target vehicle is an abnormal vehicle.
  • the present application provides a detection device for an abnormal vehicle, including: a processor and a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions that can be executed by the processor; wherein, the processor uses It is used to execute the machine-executable instructions to realize the above-mentioned abnormal vehicle detection method.
  • the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-executable instructions that can be executed by a processor; wherein the processor is configured to execute the machine-executable Instructions to implement the above abnormal vehicle detection method.
  • the present application provides a computer program, the computer program is stored in a machine-readable storage medium, and when the processor executes the computer program, the processor is prompted to implement the above abnormal vehicle detection method.
  • the target vehicle based on the minimum transit time between the first location point and the second location point and the minimum number of location points, it can be determined whether the target vehicle is an abnormal vehicle. Specifically, if the difference between the second time point when the target vehicle is at the second location point and the first time point when the target vehicle is at the first location point is less than the minimum transit time, it is determined that the target vehicle is an abnormal vehicle, if the The difference is not less than the minimum transit time, and the minimum number of location points is greater than the sum of the number of target location points (that is, the number of location points passed by the target vehicle from the first location point to the second location point) and the preset number threshold, Then it is determined that the target vehicle is an abnormal vehicle.
  • the above method can identify whether the target vehicle is an abnormal vehicle (that is, a fake vehicle with a fake license plate), thereby distinguishing a normal vehicle with a real license plate and an abnormal vehicle with a fake license plate. That is to say, although the license plate identification of the real license plate of the normal vehicle is the same as the license plate identification of the set license plate of the abnormal vehicle, the normal vehicle and the abnormal vehicle can also be identified as different vehicles. Therefore, when the vehicle is managed, the occurrence of management errors can be reduced, for example, the number of times of applying the management means for abnormal vehicles to normal vehicles can be reduced.
  • FIG. 1 is a schematic diagram of a network topology of a location point in an embodiment of the present application
  • FIG. 2 is a schematic flow diagram of a method for detecting an abnormal vehicle in an embodiment of the present application
  • FIG. 3 is a schematic flow diagram of a method for detecting an abnormal vehicle in an embodiment of the present application
  • Fig. 4 is a schematic diagram of filtering historical data in an embodiment of the present application.
  • FIG. 5 is a schematic diagram of a network topology in an implementation manner of the present application.
  • FIG. 6 is a schematic flow diagram of a method for detecting an abnormal vehicle in an embodiment of the present application.
  • Fig. 7 is a schematic structural diagram of an abnormal vehicle detection device in an embodiment of the present application.
  • Fig. 8 is a hardware structural diagram of an abnormal vehicle detection device in an embodiment of the present application.
  • first, second, and third may be used in the embodiment of the present application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of the present application, first information may also be called second information, and similarly, second information may also be called first information. Depending on the context, furthermore, the use of the word “if” could be interpreted as “at” or “when” or "in response to a determination.”
  • Location point The location used to deploy the camera.
  • a large number of cameras such as analog cameras or network cameras, etc.
  • images of vehicles driving on the road can be collected by these cameras.
  • Each camera corresponds to a location point, that is, the location of the camera is recorded as a location point, and the location point may also be called a bayonet point. That is, cameras can be deployed at a large number of points to capture images of vehicles traveling on the road.
  • FIG. 1 it is a schematic diagram of a network topology of a location point in an example.
  • the network topology can be configured by the user based on experience, or can be learned by using an algorithm. This application does not limit this, as long as the network topology can be obtained.
  • the network topology may include all location points, that is, all location points form the network topology.
  • 6 location points (such as location point A, location point B, location point C, location point D, location point E and location point F, etc.) are taken as an example. In practical applications, the number of location points Much larger than 6.
  • a camera is deployed at each location point, and the camera can collect images of vehicles passing by the location point, and analyze the license plate logo, vehicle color, vehicle appearance, etc. based on the image.
  • the camera can also send vehicle data to a storage device, which stores the vehicle data in a historical database.
  • the vehicle data stored in the historical database is called historical data.
  • Historical data includes a large number of data records, and each data record is a piece of vehicle data, which can include license plate identification, vehicle characteristics (such as vehicle color, vehicle appearance, etc.), vehicle images, and acquisition time (indicating that the image of the vehicle is captured by the camera It is collected at any time, and indicates that the vehicle is at the location point where the camera is deployed at the collection moment), the information of the location point (which can be the identification of the location point, such as location point A, location point B, etc., or the location point Longitude and latitude coordinates, the embodiment of the present application uses the identification of the location point as an example for illustration).
  • the data record may also include other content, which is not limited in this application.
  • Normal vehicles and abnormal vehicles vehicles with real license plates are called normal vehicles, and vehicles with fake license plates are called abnormal vehicles.
  • the abnormal vehicles in this embodiment refer to fake vehicles.
  • Fake license plates refer to fake license plates that are installed on a fake license plate with the same license plate logo as the real license plate with reference to the real license plate.
  • Specified statistical period refers to the specified time period for statistical vehicle data, for example, 24 hours a day (such as 0:00-24:00) can be used as a specified statistical period, or 7*24 hours a week can be used as a specified statistical period , this application does not limit it.
  • a method for detecting abnormal vehicles is proposed.
  • This method can be applied to the management device.
  • the management device and the storage device can be deployed on the same device, that is, the management device directly obtains vehicle data from the historical database, and analyzes the vehicle data based on the vehicle data. Whether it is an abnormal vehicle.
  • the management device and the storage device can also be deployed in different devices, that is, the management device is connected to the storage device, and the management device can obtain vehicle data from the historical database of the storage device, and analyze whether the vehicle is an abnormal vehicle based on the vehicle data.
  • the method may include steps 201 to 204 .
  • Step 201 If the target vehicle travels from the first location point to the second location point, obtain the first time point when the target vehicle is at the first location point, the second time point when the target vehicle is at the second location point, and the time point when the target vehicle is at the second location point. The number of target location points passed by from one location point to the second location point.
  • the target vehicle is any vehicle
  • the first location point is any location point among all location points
  • the second location point is any location point among all location points
  • the second location point is the same as the first location point something different.
  • the target vehicle is vehicle s1 (that is, the license plate is identified as s1)
  • the first location point is location point A
  • the second location point is location point D
  • all data records corresponding to vehicle s1 are obtained from the historical database, and each piece of data
  • the records include license plate identification s1, collection time, and location point identification.
  • the location point identifier in the first data record is location point A
  • the location point identifier in the second data record is location point B
  • the third data record The location point identifier in the record is location point D
  • the location point identifier in the fourth data record is location point C
  • the first time point is the collection time in the first data record
  • the second time point is At the collection time in the third data record, the number of target location points passed by vehicle s1 when driving from location point A to location point D is 2, that is, when driving from location point A to location point D, vehicle s1 passes through location points in turn Point B and location point D are 2 location points.
  • Step 202 Determine whether the difference between the second time point and the first time point is smaller than the minimum travel time between the first location point and the second location point. If not, go to step 203 ; if yes, go to step 204 .
  • the minimum travel time between the first location point and the second location point may be determined first, and the minimum travel time may be configured based on experience or obtained by using a certain algorithm, which is not limited in the present application.
  • a difference between the second time point and the first time point may be calculated. If the difference is not less than the minimum passing time, step 203 may be performed, and if the difference is less than the minimum passing time, step 204 may be performed.
  • Step 203 determine whether the minimum number of location points between the first location point and the second location point is greater than the sum of the target number of location points and a preset number threshold, and if yes, step 204 may be performed.
  • the minimum number of location points between the first location point and the second location point can be determined, and the minimum number of location points can be configured according to experience, or can be obtained by using a certain algorithm, which is not limited in this application.
  • the number of target location points and the preset number threshold can be calculated (can be configured according to experience, for this preset number
  • the threshold is not limited, such as the sum of 2, 3, etc.). If the minimum number of location points is greater than the sum of the number of target location points and the preset number threshold, step 204 may be executed.
  • Step 204 determine that the target vehicle is an abnormal vehicle (ie a licensed vehicle).
  • the target vehicle is an abnormal vehicle.
  • the difference between the second time point and the first time point is not less than the minimum travel time between the first location point and the second location point, and the minimum travel time between the first location point and the second location point. If the number of points is greater than the sum of the number of points at the target location and the preset number threshold, it can be determined that the target vehicle is an abnormal vehicle.
  • step 203 if the judgment result is no, that is, the minimum number of location points is not greater than the sum of the target number of location points and the preset number threshold, it may also include: determining that the target vehicle is normal Vehicle, i.e. the target vehicle is not a deck vehicle. This step is optional and not shown in Figure 2.
  • the target vehicle is a normal vehicle.
  • the minimum passing time is the minimum passing time corresponding to the specified statistical period
  • the minimum number of location points is the minimum number of location points corresponding to the specified statistical period, that is, for the specified statistical period, determine the first location point and the second The minimum travel time between two locations and the minimum number of locations.
  • the minimum passing time is the minimum passing time corresponding to the target time period
  • the minimum number of location points is the minimum number of location points corresponding to the target time period, that is, for the target time period, determine the first location point and The minimum travel time and the minimum number of locations between the second locations.
  • the specified statistical cycle can be divided into multiple time periods.
  • multiple time periods can be divided arbitrarily. For example, divide the specified statistical period (such as 0:00-24:00) into 4 time periods on average, time period 1 (0:00-6:00], time period 2 (6:00-12:00], time period 3 (12:00 -18 o'clock], time period 4 (18 o'clock-24 o'clock].
  • the specified statistical period (such as 0 o'clock-24 o'clock) is divided into 5 time periods, and time period 1 (0 o'clock-7 o'clock ], time period 2 (7 o'clock-9 o'clock], time period 3 (9 o'clock-17 o'clock], time period 4 (17 o'clock-19 o'clock], time period 5 (19 o'clock-24 o'clock], of which time period 2 and time period 4 are traffic peaks.
  • time period 1 (0 o'clock-7 o'clock ]
  • time period 2 (7 o'clock-9 o'clock
  • time period 3 (9 o'clock-17 o'clock
  • time period 4 (17 o'clock-19 o'clock
  • time period 5 (19 o'clock-24 o'clock] of which time period 2 and time period 4 are traffic peaks.
  • the above is just an example of the division method.
  • time period 1 corresponds to the minimum passing time t1 and the minimum number of location points n1
  • time period 2 corresponds to the minimum passing time t2 and the minimum number of location points n2
  • time period 3 corresponds to The minimum passing time t3 and the minimum number of location points n3
  • time period 4 corresponds to the minimum passing time t4 and the minimum number of location points n4.
  • the target is selected from the multiple time periods based on the first time point or the second time point period. For example, if a target time period is selected from multiple time periods based on the first time point, the time period where the first time point is located is used as the target time period. If the target time period is selected from multiple time periods based on the second time point, the time period where the second time point is located is taken as the target time period. After the target time period is obtained, the minimum transit time and the minimum number of location points corresponding to the target time period can be determined. In this case, in step 201 to step 204, the minimum passing time refers to the minimum passing time corresponding to the target time period, and the minimum number of location points refers to the minimum number of location points corresponding to the target time period.
  • the minimum passing time and the minimum number of location points between the first location point and the second location point can be determined, that is, the specified statistical period is not divided into multiple time periods, and the second location point is determined.
  • the minimum travel time and the minimum number of location points corresponding to the specified statistical period between a location point and a second location point can be determined based on all data within the specified statistical period.
  • the minimum transit time and the minimum number of position points between the first location point and the second location point corresponding to the time period may be determined. In this case, the minimum travel time and the minimum number of location points can be determined based on all the data in the time period.
  • the minimum passing time and the minimum number of location points between the first location point and the second location point corresponding to the statistical time period can be determined, and the statistical time period can be the complete time period of the specified statistical period, or specify
  • the statistical time period is any one of the multiple time periods. For example, when the statistical period is a specified statistical period, the minimum transit time and the minimum number of location points corresponding to the specified statistical period can be determined based on all the data within the specified statistical period. When the statistical time period is a certain time period, the minimum passing time and the minimum number of location points corresponding to the time period can be determined based on all the data in the time period.
  • the following method can be adopted: determine the distance between the first location point and the second location point At least one adjacent point pair passed by the target path, based on the transit time and the number of adjacent point pairs passed by the target path, determine the corresponding statistical time period between the first location point and the second location point The minimum passing time and the minimum number of location points; where, the passing time of adjacent point pairs is determined based on the sample data corresponding to the statistical time period in the historical database, the sample data includes the sample vehicle in the network topology during the statistical time period The collection time of each location point in , the adjacent point pair includes two adjacent location points, and the travel time of the adjacent point pair is determined based on the collection time when the sample vehicle is at the two adjacent location points.
  • the process of determining the transit time of adjacent point pairs based on the sample data corresponding to the statistical time period in the historical database may include but not limited to: for any adjacent point pair, obtain M (M is a positive integer) data pairs, each data pair includes the acquisition time when the sample vehicle is in the adjacent two position points of the adjacent point pair within the statistical time period; for each data pair, based on The two collection moments in the data pair determine the transit time; the minimum value among the transit durations corresponding to the M data pairs is determined as the transit duration of the adjacent point pair.
  • determine the minimum passing time and the minimum number of location points corresponding to the statistical time period between the first location point and the second location point may include but not limited to: based on the sum of the travel time of all adjacent point pairs passed by each path between the first location point and the second location point, select the path with the smallest sum of travel time as the target path, and the target The sum of the passing times of all adjacent point pairs passed by the path is determined as the minimum passing time corresponding to the statistical time period (that is, the minimum passing time between the first location point and the second location point), and the target path passes through The total number of all adjacent point pairs is determined as the minimum number of location points corresponding to the statistical time period (that is, the minimum number of location points between the first location point and the second location point).
  • the historical data corresponding to the statistical time period can be selected from the historical database, the historical data includes the collection time when the sample vehicle is at each location point within the statistical time period;
  • the historical data is filtered, and the historical data remaining after filtering is determined as sample data.
  • filtering the historical data may include but not limited to at least one of the following: for any sample vehicle, if the passing time of the sample vehicle passing through two adjacent location points is less than the preset duration threshold, filter the sample vehicle passing through the corresponding The historical data of two adjacent location points.
  • the data pairs include the sample vehicle passing through the corresponding The historical data of two adjacent location points.
  • For two adjacent location points obtain all data pairs corresponding to two adjacent location points; based on the passage time corresponding to each data pair, filter X1 data pairs with a small passage time (for example, start from the data pair with the smallest passage time , sequentially select X1 data pairs with small transit times), and filter X2 data pairs with large transit times (for example, start from the data pair with the largest transit time, and select X2 data pairs with long transit times), X1 and X2 All are positive integers. For two location points, if the total number of abnormal vehicles passing between the two location points is greater than the abnormal number threshold, all data pairs corresponding to the two location points are filtered. Certainly, the foregoing is only an example of a filtering manner, which is not limited in the present application.
  • determining the minimum passing time and the minimum number of location points corresponding to the statistical time period between the first location point and the second location point may include but is not limited to: determining whether the data update condition has been met; if so, determining The minimum passing time and the minimum number of location points corresponding to the statistical time period between the first location point and the second location point.
  • the data update condition if the total number of abnormal vehicles passing between the first location point and the second location point is greater than the abnormal number of times threshold, it is determined that the data update condition has been met; or, if the current time point and the last data update time point If the duration reaches the preset update duration, it is determined that the data update condition has been met, and the last data update time point is the last time point when the minimum passing time and the minimum number of location points were determined.
  • the target vehicle based on the minimum transit time between the first location point and the second location point and the minimum number of location points, it can be determined whether the target vehicle is an abnormal vehicle. Specifically, if the difference between the second time point when the target vehicle is at the second location point and the first time point when the target vehicle is at the first location point is less than the minimum transit time, it is determined that the target vehicle is an abnormal vehicle, if the The difference is not less than the minimum transit time, and the minimum number of location points is greater than the sum of the number of target location points (that is, the number of location points passed by the target vehicle from the first location point to the second location point) and the preset number threshold, Then it is determined that the target vehicle is an abnormal vehicle.
  • the above method can identify whether the target vehicle is an abnormal vehicle (that is, a fake vehicle with a fake license plate), thereby distinguishing a normal vehicle with a real license plate and an abnormal vehicle with a fake license plate. That is to say, although the license plate identification of the real license plate of the normal vehicle is the same as the license plate identification of the set license plate of the abnormal vehicle, the normal vehicle and the abnormal vehicle can also be identified as different vehicles. Therefore, when the vehicle is managed, the occurrence of management errors can be reduced, for example, the number of times of applying the management means for abnormal vehicles to normal vehicles can be reduced.
  • time period 1 For the convenience of description, time period 1 will be used as an example for illustration later, and the implementation manners of other time periods are similar, and details will not be repeated in this embodiment.
  • steps 301 to 303 may be used.
  • Step 301 select the historical data corresponding to time period 1 from the historical database, the historical data includes the collection time when the sample vehicle is in each location point in the network topology in time period 1; Filtering, determining the historical data remaining after filtering as the sample data corresponding to time period 1, that is, obtaining the sample data corresponding to time period 1.
  • the sample data may include collection times when the sample vehicle is at each location point in the network topology within the time period 1 .
  • the collection time in the historical data or the sample data refers to the collection time within the time period 1.
  • the historical database can store a large amount of historical data
  • the historical data includes a large number of data records
  • each data record can be a piece of vehicle data
  • the data records can include license plate identification, collection time, location point identification and other content. See Table 1 for an example of historical data.
  • license plate identification collection time position mark etc.
  • license plate logo s1 pt16 location point F license plate logo s2 pt21 Point A license plate logo s2 pt22 location point C license plate logo s2 pt23 location point E license plate logo s2 pt24 location point F ... ... ...
  • the historical data shown in Table 1 corresponds to time period 1
  • the vehicle corresponding to the license plate in the historical data is called a sample vehicle
  • the historical data includes the collection of the sample vehicle s1 (that is, the license plate is s1) at each position time, the collection time when the sample vehicle s2 is at each position point, and so on.
  • the historical data may be filtered, for example, the historical data may be filtered in at least one of the following manners.
  • use method 1 to filter historical data and use the filtered remaining historical data as sample data, or use method 2 to filter historical data and use the filtered remaining historical data as sample data, or use method 3 to filter historical data will filter the remaining historical data as sample data, or use method 1 and method 2 to filter historical data, and filter the remaining historical data as sample data, or use method 1 and method 3 to filter historical data, and filter the remaining historical data as sample data, or use method 2 and method 3 to filter historical data, and filter the remaining historical data as sample data, or use method 1, method 2 and method 3 to filter historical data, and filter the remaining Historical data is used as sample data, and there is no restriction on the filtering method. After the filtering is completed, the remaining historical data is used as sample data.
  • Method 1 For any sample vehicle, if the passing time of the sample vehicle passing through two adjacent location points is less than the preset duration threshold, filter the historical data of the sample vehicle passing through two adjacent location points.
  • the sample vehicle s1 corresponds to multiple data records, as shown in Table 1.
  • the location points can be sorted according to the collection time corresponding to each data record, such as sorting the location points according to the collection time from small to large, or sorting the location points according to the collection time from large to small, assuming
  • the sorting result is position point A, position point B, position point C, position point D, position point E, position point F, then position point A and position point B are two adjacent position points, position point B and position point C are two adjacent location points, location point C and location point D are two adjacent location points, location point D and location point E are two adjacent location points, location point E and location point F are two adjacent location points location point.
  • method 1 can be used to filter the historical data of the sample vehicle s1 passing through two adjacent location points.
  • method 1 can be used to filter other sample vehicles (such as sample vehicle s2, sample vehicle s3, etc.) It is filtered through the historical data of two adjacent location points, and will not be repeated here.
  • Mode 2 For two adjacent location points, if the total number of data pairs corresponding to the two adjacent location points is less than the preset times threshold, all data pairs corresponding to the two adjacent location points are filtered. Wherein, for each data pair, the data pair may include historical data of the sample vehicle passing through the two adjacent location points.
  • the position points can be sorted according to the collection time, and the sorted adjacent two position points form an adjacent point pair (that is, the adjacent point pair includes two adjacent points location points), so as to obtain multiple adjacent point pairs, for example, adjacent point pair 1 (location point A and location point B), adjacent point pair 2 (location point B and location point C), adjacent point pair 3 (location point C and location point D), adjacent point pair 4 (location point D and location point E), adjacent point pair 5 (location point E and location point F).
  • adjacent point pair 1 location point A and location point B
  • adjacent point pair 2 location point B and location point C
  • adjacent point pair 3 location point C and location point D
  • adjacent point pair 4 location point D and location point E
  • adjacent point pair 5 location point E and location point F
  • the position points can be sorted according to the collection time, and the sorted adjacent two position points form an adjacent point pair, for example, adjacent point pair 6 (position point A and location point C), adjacent point pair 7 (location point C and location point E), adjacent point pair 5 (location point E and location point F).
  • adjacent point pair 5 determined based on the historical data of the sample vehicle s2 is the same as the adjacent point pair 5 determined based on the historical data of the sample vehicle s1, and they are the same adjacent point pair.
  • the position points can be sorted based on the historical data of all sample vehicles, so that multiple adjacent point pairs can be obtained, and the adjacent point pairs determined based on the historical data of different sample vehicles may be repeated. No longer.
  • a plurality of adjacent point pairs can be obtained, and each adjacent point pair includes two adjacent position points.
  • the adjacent point pair For each adjacent point pair, the adjacent point pair includes two adjacent position points, and the total number of data pairs corresponding to the adjacent point pair can be counted. For example, taking adjacent point pair 1 as an example, adjacent point pair 1 includes position point A and position point B, if the sample vehicle s1 travels from position point A to position point B (that is, passes through position point A and position point B in sequence , without passing other location points between location point A and location point B), then the historical data of sample vehicle s1 at location point A and the historical data of location point B are a data pair of adjacent point pair 1.
  • the sample vehicle s1 may travel from the location point A to the location point B multiple times (for example, K times, K is a positive integer greater than 1), that is, the historical data of the sample vehicle s1 includes K historical data at the location point A data and K historical data at location point B, these historical data correspond to K data pairs.
  • the historical data of sample vehicle s2 at location point A and the historical data of location point B are a data pair of adjacent point pair 1.
  • the total number of data pairs corresponding to adjacent point pair 1 can be counted based on historical data (after sorting the historical data of sample vehicles according to the collection time, if the historical data includes position point A The historical data of the location point B and the historical data of the location point B, then the historical data of the location point A and the historical data of the location point B correspond to a data pair).
  • the total number of data pairs corresponding to adjacent point pair 1 can be obtained.
  • the data pair includes the historical data of the sample vehicle at point A and the history of the sample vehicle at point B data.
  • the preset number of times threshold can be configured according to experience
  • filter all data pairs corresponding to adjacent point pair 1 for example, filter sample vehicle s1 in The historical data of location point A and location point B, that is, delete the data record "license plate logo s1+pt11+ location point A" and data record "license plate logo s1+pt12+ location point B", and filter the sample vehicle s2 at location point A and location point B's historical data, and so on.
  • the preset times threshold can be configured according to experience
  • the data pair includes the historical data of location point A and the historical data of location point B.
  • the historical data here needs to be the history of two adjacent location points after sorting according to the collection time data. For example, after sorting the historical data of the sample vehicles according to the collection time, if the historical data of the location point A and the historical data of the location point B are included in sequence, then the historical data of the location point A and the historical data of the location point B are adjacent.
  • the historical data of two location points is a data pair.
  • the historical data of the location point A, the historical data of the location point C and the historical data of the location point B are included in sequence, then the historical data of the location point A and the historical data of the location point B
  • the historical data of is not the historical data of two adjacent position points, that is, it is not the data pair for adjacent point pair 1.
  • method 2 can be used to filter the historical data of all data pairs of adjacent point pair 1, or retain the historical data of all data pairs of adjacent point pair 1.
  • method 2 can be used to filter other adjacent point pairs Filter the historical data of all data pairs (such as adjacent point pair 2, adjacent point pair 3, etc.), or keep the historical data of all data pairs of other adjacent point pairs, which will not be repeated here.
  • the historical data of all data pairs of adjacent point pairs can be filtered, that is, when the number of passing times of adjacent point pairs is small, the adjacent point pair is filtered
  • the historical data of all data pairs in no longer retain the data pairs corresponding to this adjacent point pair, thereby reducing the interference of unreasonable data.
  • Method 3 For two adjacent location points, obtain all data pairs corresponding to two adjacent location points; based on the passage time corresponding to each data pair, filter the X1 data pairs with a small passage time, and filter the X2 data pairs with a long passage time data pairs.
  • sorting each data pair based on the corresponding travel time of each data pair you can sort each data pair according to the order of travel time from small to large, or sort each data pair according to the order of travel time from large to small, In the following, it is taken as an example to sort each data pair in ascending order of transit time.
  • the first X1 data pairs can be filtered, and the latter X2 data pairs can be filtered.
  • Both X1 and X2 are positive integers.
  • X2 data pairs keep the data pairs whose transit time is in the middle.
  • adjacent point pairs can be obtained based on the historical data of all sample vehicles.
  • Each adjacent point pair includes two adjacent location points.
  • method 2 For the method of obtaining adjacent point pairs, refer to method 2, which will not be repeated here. .
  • adjacent point pair 1 For each adjacent point pair, all data pairs corresponding to the adjacent point pair can be counted based on historical data.
  • adjacent point pair 1 includes position point A and position point B.
  • the data pair corresponding to adjacent point pair 1 includes the historical data of the sample vehicle at location point A and the historical data of the sample vehicle at location point B.
  • acquisition method of the data pair corresponding to each adjacent point pair refer to method 2, which will not be repeated here. .
  • the travel time corresponding to each data pair can be determined, that is, the collection time when the sample vehicle is at point B (based on the sample vehicle being at position B The difference between the historical data of point B) and the collection time when the sample vehicle is at point A (based on the historical data of the sample vehicle at point A).
  • the data pair corresponding to the sample vehicle s1 includes "license plate logo s1+pt11+location point A" and "license plate logo s1+pt12+location point B", and the corresponding travel time of this data pair is pt12 and pt11 difference.
  • the data pair After obtaining the transit time corresponding to each data pair, all data pairs can be sorted in ascending order of transit time. Based on the sorting result, the first X1 data pairs can be filtered, and the latter X2 data pairs can be filtered.
  • the data pair may include the historical data of the sample vehicle at location point A and the historical data of the sample vehicle at location point B, that is to say, it is necessary to filter the historical data of the sample vehicle at location point A and The historical data of the sample vehicle at location point B.
  • X1 can be configured based on experience, such as 1, 2, 3, etc., and can also be determined based on the transit time of all data pairs.
  • This application does not limit this
  • X2 can be configured based on experience, such as 1, 2, 3, etc. , can also be determined based on the transit time of all data pairs, which is not limited in this application.
  • t3 is the first passing time when the passing time presents a stable linear growth
  • t4 and t3 is not greater than the preset threshold
  • the difference between t7 and t6 is greater than the preset threshold, and the difference between t6 and t5 is not greater than the preset threshold, it means that t6 is the last passing time when the passing time shows a steady linear growth, and the data corresponding to t7 needs to be filtered Yes, that is, the value of X2 is 1. If the difference between t6 and t5 is greater than the preset threshold, and the difference between t5 and t4 is not greater than the preset threshold, it means that t5 is the last passing time when the passing time shows a steady linear growth, and the corresponding time between t7 and t6 needs to be filtered The data pair, that is, the value of X2 is 2, and so on.
  • method 3 can be used to filter the data pairs of adjacent point pair 1, that is, to filter the X1 data pairs with a small transit time, and to filter the X2 data pairs with a large transit time, and to keep the data in the middle of the transit time pair (that is, to filter the remaining data pairs), similarly, method 3 can be used to filter the data pairs of other adjacent point pairs (such as adjacent point pair 2, adjacent point pair 3, etc.), which will not be repeated here.
  • the interference of invalid data pairs can be removed, that is, the data pairs with a small transit time and a large transit time All are invalid data pairs, thereby reducing the interference of unreasonable data and retaining the most suitable data pairs.
  • the historical data can be filtered.
  • other methods can also be used to filter the historical data, which is not limited in this application.
  • the remaining historical data after filtering is determined as sample data corresponding to time period 1, and subsequent steps are performed based on the sample data.
  • Step 302 determine the travel time of each adjacent point pair based on the sample data (that is, the sample data corresponding to time period 1), the adjacent point pair may include two adjacent location points, and the sample data may be included in time period 1
  • the collection time when the inner sample vehicle is at each location point, the travel time can be determined based on the collection time when the sample vehicle is at two adjacent location points.
  • M data pairs corresponding to the adjacent point pair can be obtained, and for each data pair, the data pair can include The collection moments adjacent to the two position points, that is, the data pair includes two collection moments (these two collection moments are both collection moments within the time period 1).
  • the transit duration is determined based on the two collection moments in the data pair, that is, the transit duration corresponding to the data pair. In an example, the minimum value among the transit times corresponding to the M data pairs may be determined as the transit duration of the adjacent point pair.
  • each adjacent point pair can be obtained based on the sample data of all sample vehicles, and each adjacent point pair includes two adjacent location points.
  • the method of obtaining adjacent point pairs refer to step 301, and replace the historical data with The sample data is sufficient, and will not be repeated here.
  • all data pairs corresponding to the adjacent point pair can be determined based on the sample data of all sample vehicles, that is, the M data corresponding to the adjacent point pair Yes, each data pair may include the collection time when the sample vehicle is at two adjacent position points in the adjacent point pair.
  • adjacent point pair 1 includes position point A and position point B
  • M data pairs corresponding to adjacent point pair 1 can be obtained, and the data pairs include the sample vehicle at position point A
  • the sample data (such as the collection time) of the sample vehicle and the sample data (such as the collection time) of the sample vehicle at the position point B, the acquisition method of the data pair refers to step 301, and will not be repeated here.
  • each adjacent point pair after obtaining the M data pairs corresponding to the adjacent point pair, for each data pair, determine the passage corresponding to the data pair based on the two collection times in the data pair duration, and determine the minimum value among the transit durations corresponding to the M data pairs as the transit duration of the adjacent point pair.
  • adjacent point pair 1 for each of the M data pairs, calculate the difference between the collection time when the sample vehicle in the data pair is at point B and the collection time when the sample vehicle in the data pair is at point A The difference between them is the transit time corresponding to the data pair, so that M transit durations can be obtained, and then the minimum value of the M transit durations is used as the transit duration of the adjacent point pair 1.
  • step 302 the travel time of each adjacent point pair can be obtained.
  • Step 303 for any two location points in the network topology (which may be adjacent location points, or may not be adjacent location points), determine at least one adjacent point pair passed by the target path between these two location points, And based on the travel time and the number of adjacent point pairs passed by the target path, the minimum travel time and the minimum number of location points corresponding to time period 1 between the two location points are determined.
  • the path with the smallest sum of travel times can be selected as the target path, the sum of the travel time of all adjacent point pairs passed by the target path is determined as the minimum travel time, and the total number of all adjacent point pairs passed by the target path is determined as the minimum number of location points.
  • the network topology includes location point A, location point B, location point C, location point D, location point E, and location point F.
  • the network topology shown in Figure 5 can be constructed. The network topology shown is used to display all adjacent point pairs and the transit time of adjacent point pairs.
  • two position points with a direct connection relationship form an adjacent point pair, and the adjacent point pair has a direction.
  • A-position point B that is, driving from position point A to position point B
  • the direction of the arrow is when position point B points to position point A, indicating that the adjacent point pair is "position point B-position point A", that is, from position point B Drive to point A.
  • t1 is the transit time of the adjacent point pair "position point A-position point B”
  • t2 is the transit time length of the adjacent point pair "position point B-position point C”
  • t3 is the transit time of the adjacent point pair
  • t4 is the travel time of the adjacent point pair "point E-point D”, and so on.
  • location point A-location point B (or location point C, location point D, location point E, location point F), location point B-location point A (or location point C , location point D, location point E, location point F), location point C-location point A (or location point B, location point D, location point E, location point F), location point D-location point A (or location point Point B, Point C, Point E, Point F), Point E-Point A (or Point B, Point C, Point D, Point F), Point F-Point A ( Or location point B, location point C, location point D, location point E), and then take location point A-location point D as an example, the minimum travel time and the minimum number of location points can be obtained in the following ways:
  • the path corresponding to the location point A-location point D includes A-B-C-D and A-E-D.
  • the sum of the travel time corresponding to the route A-B-C-D is t1+t2+t3, and the sum of the travel time corresponding to the route A-E-D is t7+t4. If t1+t2+t3 is less than t7+t4, the target path between location point A and location point D is A-B-C-D, if t1+t2+t3 is greater than t7+t4, then the target path between location point A and location point D The path is A-E-D.
  • the minimum travel time is determined based on the sum of the travel time of all adjacent point pairs passed by the target path.
  • the minimum travel time can be t1+t2+t3, and the minimum travel time is also It can be (t1+t2+t3)*w, where w is a value greater than 0 and less than 1.
  • the total number of 3 determines the minimum number of location points, for example, the minimum number of location points can be 3.
  • the minimum transit time is determined based on the sum of the transit durations of all adjacent point pairs passed by the target path, such as the minimum transit duration can be t7+t4, and the minimum transit duration can also be (t7+t4) *w.
  • determine the minimum number of location points based on the total number 2 of all adjacent point pairs (including adjacent point pairs "location point A-location point E" and "location point E-location point D") passed by the target path, such as Say, the minimum number of location points can be 2.
  • the path with the smallest sum of travel time is usually the path with the fastest driving speed, but the actual road situation is that the fewer the number of traffic lights, that is, the fewer the number of location points, the faster the driving speed. quick.
  • the total number of all adjacent point pairs passed by the path with the smallest sum of travel time may be determined as the minimum number of location points.
  • the minimum travel time and the minimum number of location points of any two locations in the network topology can be determined; for time period 2, the distance between any two locations in the network topology can be determined. The minimum travel time and the minimum number of location points, and so on, will not be repeated here.
  • the minimum passing time and the minimum number of location points can be periodically determined to ensure that the minimum passing time and the minimum number of location points are closer to each other.
  • the actual road environment that is, steps 301 to 303 may be performed periodically.
  • the preset update time length such as one month, two months etc.
  • the travel time and travel trajectory of abnormal vehicles are very random.
  • the probability of abnormal vehicles appearing between two location points is very small, almost 0, that is, the number of abnormal vehicles between two location points should be less than the threshold of abnormal times. If the number of abnormal vehicles is greater than the threshold of abnormal times, it means that the road environment between two location points has changed (such as road conditions become better, or road conditions become worse, etc.), and the minimum travel time between two location points needs to be re-determined and the minimum number of location points to avoid misjudgment.
  • the method 1 to method 3 can be used. At least one method is used to filter historical data. On this basis, mode 4 can also be used to filter historical data. After the filtering is completed, the remaining historical data is used as sample data.
  • Method 4 For two location points (which can be two adjacent location points or two non-adjacent location points, this application does not limit this), if the distance between these two location points If the total number of abnormal vehicles is greater than the threshold of abnormal times, all data pairs corresponding to these two location points are filtered.
  • the probability of abnormal vehicles appearing between two location points is very small, almost 0. If the number of abnormal vehicles is greater than the threshold of abnormal times, it means that the road environment between two location points has changed, that is, Say, the data between these two location points may be invalid data. Therefore, all data pairs corresponding to these two location points can be obtained, and all data pairs corresponding to these two location points can be filtered. For the manner of acquiring all data pairs corresponding to two location points, refer to step 301, which will not be repeated here.
  • the minimum passing duration and the minimum number of location points corresponding to each time period can detect whether the vehicle is an abnormal vehicle.
  • the vehicle to be detected is called the target vehicle, as shown in Figure 6, the following steps are used to detect whether the target vehicle is an abnormal vehicle based on the minimum passing time and the minimum number of location points corresponding to each time period.
  • Step 601. If the target vehicle travels from the first location point to the second location point, obtain the first time point when the target vehicle is at the first location point, the second time point when the target vehicle is at the second location point, and the time point when the target vehicle is at the second location point. The number of target location points passed by from one location point to the second location point.
  • all data records corresponding to the target vehicle can be obtained from the historical database, each data record includes the license plate identification s1, collection time, location point identification, and these data records are sorted according to the order of collection time from front to back , assuming that the sorting result is: license plate logo s1+pt1+position point A, license plate logo s1+pt2+position point B, license plate logo s1+pt3+position point F, license plate logo s1+pt4+position point D, license plate logo s1+pt5+position point E , License plate mark s1+pt6+position point C.
  • the first position point is any position point in all position points, such as position point B
  • the second position point is any position point in all position points, such as position point F
  • the target vehicle is at the position of the first position point
  • the first time point is pt2
  • the second time point when the target vehicle is at the second position point is pt3
  • the number of target position points passed by the target vehicle from the first position point to the second position point is 1.
  • the vehicle characteristics corresponding to the target vehicle (such as vehicle color, vehicle model, vehicle appearance, etc.) Select the normal vehicle feature corresponding to the license plate identification of the target vehicle. If the vehicle feature corresponding to the target vehicle is different from the normal vehicle feature, such as the color of the vehicle, it is directly determined that the target vehicle is an abnormal vehicle, and step 601 is no longer performed. If the target vehicle If the corresponding vehicle characteristics are the same as the normal vehicle characteristics, step 601 and subsequent steps are performed to determine whether the target vehicle is an abnormal vehicle.
  • Step 602 Select a target time period from all time periods based on the first time point, for example, take the time period where the first time point is located as the target time period.
  • the target time period is selected from all time periods based on the second time point, for example, the time period of the second time point is used as the target time period.
  • Step 603 Determine the minimum transit time and the minimum number of location points corresponding to the target time period between the first location point and the second location point. For example, if the target time period is time period 1, then the minimum transit time and the minimum number of location points corresponding to time period 1 between location point B and location point F are determined.
  • Step 604 based on the first time point, the second time point, the number of target locations, the minimum travel time and the minimum number of locations, determine whether the target vehicle is an abnormal vehicle or a normal vehicle.
  • the target vehicle is at position B at the first time point pt2, and at position point F at the second time point pt3, the minimum travel time from position point B to position point F is known, if the difference between pt3 and pt2 is less than
  • the minimum travel time indicates that the target vehicle cannot travel from point B to point F within this period of time, that is to say, the target vehicle travels from point B to point F within an unreasonable time.
  • the vehicle at point B is not the same vehicle as the vehicle at point F, that is, the target vehicle is an abnormal vehicle.
  • the target vehicle is at position point B at the first time point pt2, and is at position point F at the second time point pt3, the number of target position points passed by from position point B to position point F is 1, and position point B arrives at The minimum travel time of location point F is known, and the minimum number of location points from location point B to location point F is known. If the difference between pt3 and pt2 is not less than the minimum travel time, the number of location points needs to be analyzed.
  • the minimum number of location points from location point B to location point F is 5, which means that at least 5 location points can be traveled from location point B to location point F
  • the preset number threshold is 2 (can be configured based on experience)
  • big data analysis can be performed on historical data in advance, and the minimum travel time and the minimum number of location points between any two location points that are in line with the actual use of vehicles can be calculated without manual data management.
  • the passing records related to the previously analyzed abnormal vehicles will be filtered out to reduce the impact of unreasonable passing data of abnormal vehicles on the calculation of the route by big data.
  • the results of time-space analysis and geospatial analysis are matched again, and the number of abnormal vehicles at the same location within a certain period of time is greater than the threshold. Filtering will be performed to reduce the change in passing time caused by road environment adjustments and other factors. impact on the accuracy of the analysis results.
  • a detection device for abnormal vehicles is proposed in the embodiment of the present application, as shown in Figure 7, which is a schematic structural diagram of the device, and the device may include:
  • An acquisition module 71 configured to acquire a first time point when the target vehicle is at the first position point and a second time point when the target vehicle is at the second position point if the target vehicle travels from the first position point to the second position point .
  • a determining module 72 configured to determine the The target vehicle is an abnormal vehicle; if the difference between the second time point and the first time point is not less than the minimum transit time, and the distance between the first location point and the second location point If the minimum number of location points is greater than the sum of the number of target location points and a preset number threshold, it is determined that the target vehicle is an abnormal vehicle.
  • the determining module 72 is further configured to: if the difference between the second time point and the first time point is not smaller than the first position point and the second time point The minimum passing time between the location points, and the minimum number of location points between the first location point and the second location point is not greater than the sum of the target location point number and the preset number threshold, then determine the The target vehicle is a normal vehicle.
  • the obtaining module 71 is further configured to: if the specified statistical period is divided into multiple time periods, based on the first time point or the second time point, from the multiple A target time period is selected from time periods; wherein, the minimum transit time is the minimum transit duration corresponding to the target time period; the minimum number of location points is the minimum number of location points corresponding to the target time period.
  • the determining module 72 is further configured to: determine the minimum passing time and the minimum number of location points between the first location point and the second location point corresponding to the statistical time period; Wherein, the statistical time period is the complete time period of the specified statistical period, or, when the specified statistical period is divided into multiple time periods, the statistical time period is any one of the multiple time periods;
  • the determining module determines the minimum passing time and the minimum number of location points corresponding to the statistical time period between the first location point and the second location point it is specifically used to: determine the distance between the first location point and the second location point At least one adjacent point pair that the target path passes through, based on the transit time and the number of adjacent point pairs that the target path passes through at least one adjacent point pair, determine the distance between the first location point and the second location point The minimum passing time corresponding to the statistical time period and the minimum number of location points; wherein, the passing time of the at least one adjacent point pair is determined based on the sample data corresponding to the statistical time period in the historical database, and the sample
  • the determination module 72 is specifically used to determine the transit time of the at least one adjacent point pair based on the sample data corresponding to the statistical time period in the historical database: for any adjacent point Yes, obtain M data pairs corresponding to the adjacent point pair, each data pair includes the acquisition time when the sample vehicle is in the adjacent two position points of the adjacent point pair within the statistical time period; for each For a data pair, the transit time is determined based on the two collection moments in the data pair; and the minimum value among the transit durations corresponding to the M data pairs is determined as the transit duration of the adjacent point pair.
  • the determining module 72 determines the first location point and the second location based on the travel time of the at least one adjacent point pair passed by the target path and the number of adjacent point pairs.
  • the minimum passing time and the minimum number of location points corresponding to the statistical time period between points are specifically used for: based on all adjacent point pairs passed by each path between the first location point and the second location point
  • the sum of transit times select the path with the smallest sum of transit durations as the target path, determine the sum of transit durations of all adjacent point pairs passed by the target path as the minimum transit duration corresponding to the statistical time period, and
  • the total number of all adjacent point pairs passed by the target path is determined as the minimum number of location points corresponding to the statistical time period.
  • the determination module 72 is further configured to: select historical data corresponding to the statistical time period from the historical database, the historical data including the sample vehicle within the statistical time period At the collection time of each location point; filter the historical data, and determine the remaining historical data after filtering as the sample data; wherein, filtering the historical data includes at least one of the following: for any sample vehicle, if the The passing time of the sample vehicle passing through two adjacent location points is less than the preset duration threshold, then filter the historical data of the sample vehicle passing through the two adjacent location points; for the two adjacent location points, if the two adjacent location points The total number of data pairs corresponding to each location point is less than the preset number of times threshold, then filter all data pairs corresponding to the two adjacent location points; wherein, the data pair includes the sample vehicle passing through the two adjacent location points historical data; for two adjacent location points, obtain all data pairs corresponding to the two adjacent location points; based on the passage time corresponding to each data pair, filter the X1 data pairs with a small passage time, and filter the passage
  • the determination module 72 is specifically used to determine the minimum passing time and the minimum number of location points between the first location point and the second location point and corresponding to the statistical time period: Determine whether the data update condition has been satisfied; if yes, then determine the minimum passing time and the minimum number of location points corresponding to the statistical time period between the first location point and the second location point; wherein, if the first location point If the total number of abnormal vehicles passing between a position point and the second position point is greater than the abnormal number of times threshold, it is determined that the data update condition has been met; or, if the duration between the current time point and the last data update time point reaches If the update duration is preset, it is determined that the data update condition has been met, and the last data update time point is the last time point when the minimum transit time length and the minimum number of location points were determined.
  • the detection device for an abnormal vehicle may include: a processor 81 and a machine-readable storage A medium 82, the machine-readable storage medium 82 stores machine-executable instructions that can be executed by the processor 81; the processor 81 is used to execute the machine-executable instructions, so as to realize the abnormal vehicle disclosed in the above examples of the present application detection method.
  • the embodiment of the present application also provides a machine-readable storage medium, on which several computer instructions are stored, and when the computer instructions are executed by a processor, the present invention can be realized. Apply the abnormal vehicle detection method disclosed in the above example.
  • the above-mentioned machine-readable storage medium may be any electronic, magnetic, optical or other physical storage device, which may contain or store information, such as executable instructions, data, and so on.
  • the machine-readable storage medium can be: RAM (Radom Access Memory, random access memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid state drive, any type of storage disk (such as CD, DVD, etc.), or similar storage media, or a combination of them.
  • the embodiment of the present application also provides a computer program, the computer program is stored in a machine-readable storage medium, and when the processor executes the computer program, the processor is prompted to implement the above-mentioned examples of the present application.
  • a disclosed detection method for abnormal vehicles is provided.
  • a typical implementing device is a computer, which may take the form of a personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant, media player, navigation device, e-mail device, game control device, etc. desktops, tablets, wearables, or any combination of these.
  • embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • these computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means,
  • the instruction means implements the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operational steps are performed on the computer or other programmable equipment to produce computer-implemented processing, so that the information executed on the computer or other programmable equipment
  • the instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Traffic Control Systems (AREA)

Abstract

本申请提供一种异常车辆的检测方法、装置及设备。该方法包括：若目标车辆从第一位置点行驶到第二位置点，获取目标车辆处于第一位置点的第一时间点、目标车辆处于第二位置点的第二时间点、目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量；若第二时间点与第一时间点之间的差值小于第一位置点与第二位置点之间的最小通行时长，确定目标车辆是异常车辆；若第二时间点与第一时间点之间的差值不小于最小通行时长，且第一位置点与第二位置点之间的最小位置点数量大于目标位置点数量与预设数量阈值之和，确定目标车辆是异常车辆。

Description

异常车辆的检测方法、装置及设备 技术领域
本申请涉及智能交通领域，尤其是一种异常车辆的检测方法、装置及设备。
背景技术
安装有套牌车牌的车辆称为套牌车辆，套牌车牌参照真实车牌伪造，因此其车牌标识与真实车牌相同。套牌车辆是一种异常车辆，因此，需要及时发现并有效管理道路上行驶的套牌车辆。
针对道路等特定场所，通常会部署大量摄像机(如模拟摄像机或者网络摄像机等)，可以通过这些摄像机采集道路上行驶的车辆的图像，并基于图像分析出车辆的车牌标识，将具有同一车牌标识的车辆确认为同一车辆，借此对该车辆的道路行驶行为进行管理和规范。
但是，由于正常车辆的真实车牌的车牌标识与异常车辆的套牌车牌的车牌标识相同，因此，可能将具有同一车牌标识的正常车辆和异常车辆识别为同一车辆。这样，如果将针对异常车辆的管理手段应用到正常车辆，就会造成管理错误。
发明内容
本申请提供一种异常车辆的检测方法，所述方法包括：若目标车辆从第一位置点行驶到第二位置点，则获取所述目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点、所述目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量；若所述第二时间点与所述第一时间点之间的差值小于所述第一位置点与所述第二位置点之间的最小通行时长，则确定所述目标车辆是异常车辆；若所述第二时间点与所述第一时间点之间的差值不小于所述最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是异常车辆。
本申请提供一种异常车辆的检测装置，所述装置包括：获取模块，用于若目标车辆从第一位置点行驶到第二位置点，则获取目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点、所述目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量；确定模块，用于若所述第二时间点与所述第一时间点之间的差值小于所述第一位置点与所述第二位置点之间的最小通行时长，则确定所述目标车辆是异常车辆；若所述第二时间点与所述第一时间点之间的差值不小于所述最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是异常车辆。
本申请提供一种异常车辆的检测设备，包括：处理器和机器可读存储介质，所述机器可读存储介质存储有能够被所述处理器执行的机器可执行指令；其中，所述处理器用于执行所述机器可执行指令，以实现上述的异常车辆的检测方法。
本申请提供一种非暂时性机器可读存储介质，所述非暂时性机器可读存储介质存储有能够被处理器执行的机器可执行指令；其中，所述处理器用于执行所述机器可执行指令，以实现上述的异常车辆的检测方法。
本申请提供一种计算机程序，所述计算机程序存储于机器可读存储介质，当处理器执行所述计算机程序时，促使处理器实现上述的异常车辆的检测方法。
由以上技术方案可见，本申请实施例中，基于第一位置点与第二位置点之间的最小通行时长以及最小位置点数量，可以确定目标车辆是否是异常车辆。具体地，若目标车 辆处于第二位置点的第二时间点与目标车辆处于第一位置点的第一时间点之间的差值小于该最小通行时长，则确定目标车辆是异常车辆，若该差值不小于该最小通行时长，且该最小位置点数量大于目标位置点数量(即目标车辆从第一位置点行驶到第二位置点所经过的位置点数量)与预设数量阈值之和，则确定目标车辆是异常车辆。上述方式可以识别出目标车辆是否为异常车辆(即安装有套牌车牌的套牌车辆)，从而区分出安装有真实车牌的正常车辆和安装有套牌车牌的异常车辆。也就是说，虽然正常车辆的真实车牌的车牌标识与异常车辆的套牌车牌的车牌标识相同，也能够将正常车辆和异常车辆识别为不同的车辆。因此，在对车辆进行管理时，可以减少管理错误的发生，比如说，减少将针对异常车辆的管理手段应用到正常车辆的次数。
附图说明
为了更加清楚地说明本申请实施例中的技术方案，下面将对本申请实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅用于辅助说明本申请中记载的一些实施例，对于本领域普通技术人员来讲，还可以根据本申请实施例的这些附图获得其他的附图。
图1是本申请一种实施方式中的位置点的网络拓扑示意图；
图2是本申请一种实施方式中的异常车辆的检测方法的流程示意图；
图3是本申请一种实施方式中的异常车辆的检测方法的流程示意图；
图4是本申请一种实施方式中的历史数据的过滤示意图；
图5是本申请一种实施方式中的网络拓扑的示意图；
图6是本申请一种实施方式中的异常车辆的检测方法的流程示意图；
图7是本申请一种实施方式中的异常车辆的检测装置的结构示意图；
图8是本申请一种实施方式中的异常车辆的检测设备的硬件结构图。
具体实施方式
在本申请实施例使用的术语仅仅是出于描述特定实施例的目的，而非限制本申请。本申请和权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式，除非上下文清楚地表示其它含义。还应当理解，本文中使用的术语“和/或”是指包含一个或多个相关联的列出项目的任何或所有可能组合。
应当理解，尽管在本申请实施例可能采用术语第一、第二、第三等来描述各种信息，但这些信息不应限于这些术语。这些术语仅用来将同一类型的信息彼此区分开。例如，在不脱离本申请范围的情况下，第一信息也可以被称为第二信息，类似地，第二信息也可以被称为第一信息。取决于语境，此外，所使用的词语“如果”可以被解释成为“在……时”或“当……时”或“响应于确定”。
在介绍本申请实施例的技术方案之前，先介绍与本申请有关的技术术语。
位置点：用于部署摄像机的位置。针对道路上的特定场所(如高速收费站、交通检查站、公路等场所)，通常会部署大量摄像机(如模拟摄像机或者网络摄像机等)，可以通过这些摄像机采集道路上行驶的车辆的图像。每个摄像机对应一个位置点，即将摄像机所在位置记为位置点，位置点也可以称为卡口点。也就是说，可以在大量位置点部署摄像机，以采集道路上行驶的车辆的图像。
参见图1所示，为一示例中的位置点的网络拓扑示意图。网络拓扑可以是用户根据经验配置，也可以是采用算法学习，本申请对此不做限制，只要能够得到网络拓扑即可。网络拓扑可以包括所有位置点，即所有位置点组成网络拓扑。在图1中，是以6个位置点(如位置点A、位置点B、位置点C、位置点D、位置点E和位置点F等)为例，在实际应用中，位置点的数量远远大于6个。
每个位置点部署有摄像机，摄像机可以采集经过该位置点的车辆的图像，并基于图像分析出车牌标识、车辆颜色、车辆外观等。摄像机还可以将车辆数据发送给存储设备，由存储设备将车辆数据存储在历史数据库中。将历史数据库中存储的车辆数据称为历史数据。历史数据包括大量数据记录，每条数据记录是一条车辆数据，可以包括车牌标识、车辆特征(如车辆颜色、车辆外观等)、车辆的图像、采集时刻(表示该车辆的图像是摄像机在该采集时刻采集，并表示该车辆在该采集时刻处于该摄像机所部署的位置点)、该位置点的信息(可以是位置点的标识，如位置点A、位置点B等，也可以是位置点的经纬度坐标，本申请实施例以位置点的标识为例进行说明)。另外，该数据记录还可以包括其它内容，本申请对此不做限制。
正常车辆和异常车辆：将具有真实车牌的车辆称为正常车辆，将具有套牌车牌的车辆称为异常车辆，本实施例中的异常车辆是指套牌车辆。套牌车牌是指参照真实车牌，被安装到套牌车辆的与真实车牌的车牌标识相同的假车牌。
指定统计周期：指指定的统计车辆数据的时间周期，例如，可以将一天的24小时(如0时-24时)作为一个指定统计周期，也可以将一周的7*24小时作为一个指定统计周期，本申请对此不做限制。
本申请实施例中提出一种异常车辆的检测方法，该方法可以应用于管理设备，管理设备与存储设备可以部署在同一设备，即管理设备直接从历史数据库中获取车辆数据，基于车辆数据分析车辆是否为异常车辆。管理设备与存储设备也可以部署在不同设备，即管理设备与存储设备连接，管理设备可以从存储设备的历史数据库中获取车辆数据，基于车辆数据分析车辆是否为异常车辆。
参见图2所示，为该异常车辆的检测方法的流程示意图，该方法可以包括步骤201至步骤204。
步骤201、若目标车辆从第一位置点行驶到第二位置点，则获取目标车辆处于第一位置点的第一时间点、目标车辆处于第二位置点的第二时间点、目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量。
示例性的，目标车辆是任一车辆，第一位置点是所有位置点中的任一位置点，第二位置点是所有位置点中的任一位置点，且第二位置点与第一位置点不同。假设目标车辆是车辆s1(即车牌标识为s1)，第一位置点是位置点A，第二位置点是位置点D，则从历史数据库中获取与车辆s1对应的所有数据记录，每条数据记录均包括车牌标识s1、采集时刻、位置点标识。按照采集时刻从前到后的顺序对这些数据记录进行排序，假设第一条数据记录中的位置点标识是位置点A，第二条数据记录中的位置点标识是位置点B，第三条数据记录中的位置点标识是位置点D，第四条数据记录中的位置点标识是位置点C，以此类推，第一时间点是第一条数据记录中的采集时刻，第二时间点是第三条数据记录中的采集时刻，则车辆s1从位置点A行驶到位置点D时所经过的目标位置点数量是2，即从位置点A行驶到位置点D时，车辆s1依次经过位置点B和位置点D2个位置点。
步骤202、确定第二时间点与第一时间点之间的差值是否小于第一位置点与第二位置点之间的最小通行时长。若否，执行步骤203，若是，执行步骤204。
示例性的，可以先确定第一位置点与第二位置点之间的最小通行时长，该最小通行时长可以根据经验配置，也可以采用某种算法得到，本申请对此不做限制。
示例性的，在得到该第二时间点和该第一时间点之后，可以计算该第二时间点与该第一时间点之间的差值。若该差值不小于该最小通行时长，则可以执行步骤203，若该差值小于该最小通行时长，则可以执行步骤204。
步骤203、确定第一位置点与第二位置点之间的最小位置点数量是否大于目标位置点数量与预设数量阈值之和，如果是，则可以执行步骤204。
示例性的，可以确定第一位置点与第二位置点之间的最小位置点数量，该最小位置 点数量可以根据经验配置，也可以采用某种算法得到，本申请对此不做限制。
示例性的，在得到目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量之后，可以计算目标位置点数量与预设数量阈值(可以根据经验配置，对此预设数量阈值不做限制，如2、3等)之和。若该最小位置点数量大于目标位置点数量与预设数量阈值之和，则可以执行步骤204。
步骤204、确定目标车辆是异常车辆(即套牌车辆)。
综上可以看出，若第二时间点与第一时间点之间的差值小于第一位置点与第二位置点之间的最小通行时长，则可以确定目标车辆是异常车辆。或者，若第二时间点与第一时间点之间的差值不小于第一位置点与第二位置点之间的最小通行时长，且第一位置点与第二位置点之间的最小位置点数量大于目标位置点数量与预设数量阈值之和，则可以确定目标车辆是异常车辆。
在一种可能的实施方式中，针对步骤203，如果判断结果为否，即，该最小位置点数量不大于目标位置点数量与预设数量阈值之和，则还可以包括：确定目标车辆是正常车辆，即目标车辆不是套牌车辆。此步骤为可选步骤，未在图2中示出。
综上，若第二时间点与第一时间点之间的差值不小于第一位置点与第二位置点之间的最小通行时长，且第一位置点与第二位置点之间的最小位置点数量不大于目标位置点数量与预设数量阈值之和，则确定目标车辆是正常车辆。
在一种可能的实施方式中，最小通行时长是指定统计周期对应的最小通行时长，最小位置点数量是指定统计周期对应的最小位置点数量，即针对指定统计周期，确定第一位置点与第二位置点之间的最小通行时长和最小位置点数量。
在另一种可能的实施方式中，最小通行时长是目标时间段对应的最小通行时长，最小位置点数量是目标时间段对应的最小位置点数量，即针对目标时间段，确定第一位置点与第二位置点之间的最小通行时长和最小位置点数量。
比如说，针对指定统计周期，可以将指定统计周期划分为多个时间段，对此划分方式不做限制，可以任意划分多个时间段。例如，将指定统计周期(如0时-24时)平均划分为4个时间段，时间段1(0时-6时]、时间段2(6时-12时]、时间段3(12时-18时]、时间段4(18时-24时]。又例如，按照行车高峰将指定统计周期(如0时-24时)划分为5个时间段，时间段1(0时-7时]、时间段2(7时-9时]、时间段3(9时-17时]、时间段4(17时-19时]、时间段5(19时-24时]，其中时间段2和时间段4是行车高峰。当然，上述只是划分方式的示例。
在将指定统计周期划分为多个时间段后，针对每个时间段(如时间段1、时间段2、时间段3或时间段4)，确定第一位置点与第二位置点之间的与该时间段对应的最小通行时长和最小位置点数量，如时间段1对应最小通行时长t1和最小位置点数量n1，时间段2对应最小通行时长t2和最小位置点数量n2，时间段3对应最小通行时长t3和最小位置点数量n3，时间段4对应最小通行时长t4和最小位置点数量n4。
在此基础上，在获取第一时间点和第二时间点之后，若指定统计周期被划分为多个时间段，则基于第一时间点或者第二时间点从多个时间段中选取出目标时间段。比如说，若基于第一时间点从多个时间段中选取出目标时间段，则将第一时间点所处的时间段作为目标时间段。若基于第二时间点从多个时间段中选取出目标时间段，则将第二时间点所处的时间段作为目标时间段。在得到目标时间段之后，就可以确定目标时间段对应的的最小通行时长和最小位置点数量。这种情况下，在步骤201至步骤204中，最小通行时长是指与目标时间段对应的最小通行时长，最小位置点数量是指与目标时间段对应的最小位置点数量。
在一种可能的实施方式中，可以确定第一位置点与第二位置点之间的最小通行时长和最小位置点数量，也就是说，未将指定统计周期划分为多个时间段，确定第一位置点与第二位置点之间的与指定统计周期对应的最小通行时长和最小位置点数量。在该情况 下，可以基于指定统计周期内的所有数据确定最小通行时长和最小位置点数量。或者，若指定统计周期被划分为多个时间段，针对各时间段，可以确定第一位置点与第二位置点之间的与该时间段对应的最小通行时长和最小位置点数量。在该情况下，可以基于该时间段内的所有数据确定最小通行时长和最小位置点数量。
综上所述，可以确定第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，统计时间段可以为指定统计周期的完整时间段，或者，指定统计周期被划分为多个时间段时，该统计时间段为多个时间段中的任一时间段。比如说，统计时间段是指定统计周期时，就可以基于指定统计周期内的所有数据确定与指定统计周期对应的最小通行时长和最小位置点数量。统计时间段是某个时间段时，就可以基于该时间段内的所有数据确定与该时间段对应的最小通行时长和最小位置点数量。
示例性的，为了确定第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，可以采用如下方式：确定第一位置点与第二位置点之间的目标路径经过的至少一个相邻点对，基于该目标路径经过的相邻点对的通行时长及相邻点对数量，确定第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，相邻点对的通行时长是基于历史数据库中的与统计时间段对应的样本数据确定，该样本数据包括在该统计时间段内样本车辆处于网络拓扑中各位置点的采集时刻，相邻点对包括相邻两个位置点，相邻点对的通行时长是基于样本车辆处于相邻两个位置点的采集时刻确定。
示例性的，基于历史数据库中的与统计时间段对应的样本数据确定相邻点对的通行时长的过程，可以包括但不限于：针对任一相邻点对，获取与该相邻点对匹配的M(M为正整数)个数据对，各数据对均包括在统计时间段内样本车辆处于该相邻点对中相邻两个位置点的采集时刻；针对每个数据对来说，基于该数据对中的两个采集时刻确定通行时长；将M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
示例性的，基于目标路径经过的相邻点对的通行时长及相邻点对数量，确定第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，可以包括但不限于：基于第一位置点与第二位置点之间的各路径经过的所有相邻点对的通行时长之和，选取通行时长之和最小的路径作为目标路径，将该目标路径经过的所有相邻点对的通行时长之和确定为与统计时间段对应的最小通行时长(即第一位置点与第二位置点之间的最小通行时长)，并将该目标路径经过的所有相邻点对的总数量确定为与统计时间段对应的最小位置点数量(即第一位置点与第二位置点之间的最小位置点数量)。
示例性的，为了得到与统计时间段对应的样本数据，可以从历史数据库中选取与统计时间段对应的历史数据，该历史数据包括在该统计时间段内样本车辆处于各位置点的采集时刻；对历史数据进行过滤，将过滤剩余的历史数据确定为样本数据。其中，对历史数据进行过滤可以包括但不限于以下至少一种：针对任一样本车辆，若该样本车辆经过相邻两个位置点的通行时长小于预设时长阈值，则过滤该样本车辆经过相邻两个位置点的历史数据。针对相邻两个位置点，若相邻两个位置点对应的数据对的总数量小于预设次数阈值，则过滤相邻两个位置点对应的所有数据对，该数据对包括样本车辆经过相邻两个位置点的历史数据。针对相邻两个位置点，获取相邻两个位置点对应的所有数据对；基于各数据对对应的通行时长，过滤通行时长小的X1个数据对(例如，从通行时长最小的数据对开始，依次选取通行时长小的X1个数据对)，并过滤通行时长大的X2个数据对(例如，从通行时长最大的数据对开始，依次选取通行时长大的X2个数据对)，X1和X2均为正整数。针对两个位置点，若两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则过滤两个位置点对应的所有数据对。当然，上述只是过滤方式的示例，本申请对此不做限制。
示例性的，确定第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，可以包括但不限于：确定是否已满足数据更新条件；若是，则确定 第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量。其中，若第一位置点与第二位置点之间经过的异常车辆的总数量大于异常次数阈值，则确定已满足数据更新条件；或，若当前时间点与上一次数据更新时间点之间的时长达到预设更新时长，则确定已满足数据更新条件，上一次数据更新时间点是上一次确定最小通行时长和最小位置点数量的时间点。
由以上技术方案可见，本申请实施例中，基于第一位置点与第二位置点之间的最小通行时长以及最小位置点数量，可以确定目标车辆是否是异常车辆。具体地，若目标车辆处于第二位置点的第二时间点与目标车辆处于第一位置点的第一时间点之间的差值小于该最小通行时长，则确定目标车辆是异常车辆，若该差值不小于该最小通行时长，且该最小位置点数量大于目标位置点数量(即目标车辆从第一位置点行驶到第二位置点所经过的位置点数量)与预设数量阈值之和，则确定目标车辆是异常车辆。上述方式可以识别出目标车辆是否为异常车辆(即安装有套牌车牌的套牌车辆)，从而区分出安装有真实车牌的正常车辆和安装有套牌车牌的异常车辆。也就是说，虽然正常车辆的真实车牌的车牌标识与异常车辆的套牌车牌的车牌标识相同，也能够将正常车辆和异常车辆识别为不同的车辆。因此，在对车辆进行管理时，可以减少管理错误的发生，比如说，减少将针对异常车辆的管理手段应用到正常车辆的次数。
以下结合具体应用场景，对本申请实施例的技术方案进行说明。
若指定统计周期被划分为多个时间段，针对每个时间段(如时间段1、时间段2、时间段3和时间段4)，可以确定任意两个位置点之间的与该时间段对应的最小通行时长和最小位置点数量。为了方便描述，后面将以时间段1为例进行说明，其它时间段的实现方式与此类似，本实施例中不再赘述。
当然，在实际应用中，也可以确定任意两个位置点之间的与指定统计周期对应的最小通行时长和最小位置点数量，这种情况下，将与时间段1对应的历史数据替换为所有历史数据(即不用根据所处时间段区分历史数据)即可，本实施例中不再赘述。
参见图3所示，为了确定任意两个位置点之间的与时间段1对应的最小通行时长和最小位置点数量，可以采用步骤301至步骤303。
步骤301、从历史数据库中选取与时间段1对应的历史数据，该历史数据包括在时间段1内样本车辆处于网络拓扑中各位置点的采集时刻；对与时间段1对应的该历史数据进行过滤，将过滤剩余的历史数据确定为与时间段1对应的样本数据，即得到与时间段1对应的样本数据。该样本数据可以包括在时间段1内样本车辆处于网络拓扑中各位置点的采集时刻。在后续过程中，历史数据或者样本数据中的采集时刻均是指位于时间段1内的采集时刻。
示例性的，历史数据库可以存储大量历史数据，该历史数据包括大量数据记录，每条数据记录可以是一条车辆数据，该数据记录可以包括车牌标识、采集时刻、位置点标识等内容。参见表1所示，为历史数据的一个示例。
表1
车牌标识 采集时刻 位置点标识等
车牌标识s1 pt11 位置点A
车牌标识s1 pt12 位置点B
车牌标识s1 pt13 位置点C
车牌标识s1 pt14 位置点D
车牌标识s1 pt15 位置点E
车牌标识s1 pt16 位置点F
车牌标识s2 pt21 位置点A
车牌标识s2 pt22 位置点C
车牌标识s2 pt23 位置点E
车牌标识s2 pt24 位置点F
… … …
假设表1所示的历史数据与时间段1对应，将历史数据中的车牌标识对应的车辆称为样本车辆，则该历史数据包括样本车辆s1(即车牌标识是s1)处于各位置点的采集时刻、样本车辆s2处于各位置点的采集时刻，以此类推。
在得到与时间段1对应的历史数据之后，可以对历史数据进行过滤，比如说，采用如下方式的至少一种对历史数据进行过滤。例如，采用方式1对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式2对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式3对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式1和方式2对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式1和方式3对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式2和方式3对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，或采用方式1、方式2和方式3对历史数据进行过滤，将过滤剩余的历史数据作为样本数据，对此过滤方式不做限制，在过滤完成后将剩余历史数据作为样本数据。
方式1、针对任一样本车辆，若该样本车辆经过相邻两个位置点的通行时长小于预设时长阈值，则过滤该样本车辆经过相邻两个位置点的历史数据。
比如说，针对样本车辆s1来说，样本车辆s1对应多个数据记录，参见表1所示。可以根据各数据记录对应的采集时刻对各位置点进行排序，如按照采集时刻从小到大的顺序对各位置点进行排序，或按照采集时刻从大到小的顺序对各位置点进行排序，假设排序结果为位置点A、位置点B、位置点C、位置点D、位置点E、位置点F，则位置点A和位置点B是相邻两个位置点，位置点B和位置点C是相邻两个位置点，位置点C和位置点D是相邻两个位置点，位置点D和位置点E是相邻两个位置点，位置点E和位置点F是相邻两个位置点。
确定样本车辆s1经过位置点A和位置点B的通行时长(即pt12与pt11的差值)，若该通行时长小于预设时长阈值(可以根据经验配置)，则过滤样本车辆s1处于位置点A和位置点B的历史数据，即删除数据记录“车牌标识s1+pt11+位置点A”和数据记录“车牌标识s1+pt12+位置点B”，若该通行时长不小于预设时长阈值，则保留样本车辆s1处于位置点A和位置点B的历史数据。
然后，确定样本车辆s1经过位置点B和位置点C的通行时长(即pt13与pt12的差值)，若该通行时长小于预设时长阈值，则过滤样本车辆s1处于位置点B和位置点C的历史数据，若该通行时长不小于预设时长阈值，则保留样本车辆s1处于位置点B和位置点C的历史数据，以此类推。
显然，针对样本车辆s1，可以采用方式1对样本车辆s1经过相邻两个位置点的历史数据进行过滤，同理，可以采用方式1对其它样本车辆(如样本车辆s2、样本车辆s3等)经过相邻两个位置点的历史数据进行过滤，在此不再赘述。
方式2、针对相邻两个位置点，若相邻两个位置点对应的数据对的总数量小于预设次数阈值，则过滤相邻两个位置点对应的所有数据对。其中，针对每个数据对来说，该数据对可以包括样本车辆经过这相邻两个位置点的历史数据。
比如说，针对样本车辆s1，参见表1所示，可以根据采集时刻对各位置点进行排序， 排序后的相邻两个位置点组成一个相邻点对(即相邻点对包括相邻两个位置点)，从而得到多个相邻点对，例如，相邻点对1(位置点A和位置点B)、相邻点对2(位置点B和位置点C)、相邻点对3(位置点C和位置点D)、相邻点对4(位置点D和位置点E)、相邻点对5(位置点E和位置点F)。
针对样本车辆s2，参见表1所示，可以根据采集时刻对各位置点进行排序，排序后的相邻两个位置点组成一个相邻点对，例如，相邻点对6(位置点A和位置点C)、相邻点对7(位置点C和位置点E)、相邻点对5(位置点E和位置点F)。显然，基于样本车辆s2的历史数据确定的相邻点对5与基于样本车辆s1的历史数据确定的相邻点对5相同，二者是同一个相邻点对。
以此类推，基于所有样本车辆的历史数据均可以对各位置点进行排序，从而能够得到多个相邻点对，基于不同样本车辆的历史数据确定的相邻点对可能存在重复，对此过程不再赘述。综上所述，在基于所有样本车辆的历史数据进行上述处理后，可以得到多个相邻点对，每个相邻点对均包括相邻两个位置点。
针对每个相邻点对，该相邻点对包括相邻两个位置点，可以统计该相邻点对对应的数据对的总数量。比如说，以相邻点对1为例，相邻点对1包括位置点A和位置点B，若样本车辆s1从位置点A行驶到位置点B(即依次经过位置点A和位置点B，在位置点A和位置点B之间未经过其它位置点)，则样本车辆s1处于位置点A的历史数据和处于位置点B的历史数据是相邻点对1的一个数据对。在实际应用中，样本车辆s1可能多次(如K次，K为大于1的正整数)从位置点A行驶到位置点B，即样本车辆s1的历史数据包括K个处于位置点A的历史数据和K个处于位置点B的历史数据，这些历史数据对应K个数据对。
同理，若样本车辆s2从位置点A行驶到位置点B，则样本车辆s2处于位置点A的历史数据和处于位置点B的历史数据是相邻点对1的一个数据对。
以此类推，针对相邻点对1，可以基于历史数据统计相邻点对1对应的数据对的总数量(按照采集时刻对样本车辆的历史数据进行排序后，若历史数据依次包括位置点A的历史数据和位置点B的历史数据，则位置点A的历史数据和位置点B的历史数据对应一个数据对)。
综上所述，可以得到相邻点对1对应的数据对的总数量，针对每个数据对来说，该数据对包括样本车辆处于位置点A的历史数据和样本车辆处于位置点B的历史数据。在此基础上，若相邻点对1对应的数据对的总数量小于预设次数阈值(可以根据经验配置)，则过滤相邻点对1对应的所有数据对，例如，过滤样本车辆s1处于位置点A和位置点B的历史数据，即删除数据记录“车牌标识s1+pt11+位置点A”和数据记录“车牌标识s1+pt12+位置点B”，过滤样本车辆s2处于位置点A和位置点B的历史数据，以此类推。若相邻点对1对应的数据对的总数量不小于预设次数阈值，则保留相邻点对1对应的所有数据对。
需要说明的是，针对数据对来说，该数据对包括位置点A的历史数据和位置点B的历史数据，这里的历史数据需要是根据采集时刻进行排序后相邻的两个位置点的历史数据。比如说，按照采集时刻对样本车辆的历史数据进行排序后，若依次包括位置点A的历史数据和位置点B的历史数据，则位置点A的历史数据和位置点B的历史数据就是相邻两个位置点的历史数据，是一个数据对。但是，按照采集时刻对样本车辆的历史数据进行排序后，若依次包括位置点A的历史数据、位置点C的历史数据和位置点B的历史数据，则位置点A的历史数据和位置点B的历史数据不是相邻两个位置点的历史数据，也就不是针对相邻点对1的数据对。
显然，可以采用方式2对相邻点对1的所有数据对的历史数据进行过滤，或者保留相邻点对1的所有数据对的历史数据，同理，可以采用方式2对其它相邻点对(如相邻点对2、相邻点对3等)的所有数据对的历史数据进行过滤，或者保留其它相邻点对的 所有数据对的历史数据，在此不再赘述。
综上所述，通过设置预设次数阈值，可以对相邻点对的所有数据对的历史数据进行过滤，即，当相邻点对的过车次数较少时，就过滤这个相邻点对的所有数据对的历史数据，不再保留这个相邻点对对应的数据对，从而减少不合理数据的干扰。
方式3、针对相邻两个位置点，获取相邻两个位置点对应的所有数据对；基于各数据对对应的通行时长，过滤通行时长小的X1个数据对，并过滤通行时长大的X2个数据对。
例如，基于各数据对对应的通行时长对各数据对进行排序，可以按照通行时长从小到大的顺序对各数据对进行排序，或者按照通行时长从大到小的顺序对各数据对进行排序，后续以按照通行时长从小到大的顺序对各数据对进行排序为例。基于排序结果，可以过滤前面X1个数据对，并过滤后面X2个数据对，X1和X2均为正整数，也就是说，可以过滤通行时长较小的X1个数据对，过滤通行时长较大的X2个数据对，保留通行时长处于中间的数据对。
比如说，可以基于所有样本车辆的历史数据得到多个相邻点对，每个相邻点对均包括相邻两个位置点，相邻点对的获取方式参见方式2，在此不再赘述。
针对每个相邻点对来说，可以基于历史数据统计该相邻点对对应的所有数据对，以相邻点对1为例，相邻点对1包括位置点A和位置点B，相邻点对1对应的数据对包括样本车辆处于位置点A的历史数据和样本车辆处于位置点B的历史数据，各相邻点对对应的数据对的获取方式参见方式2，在此不再赘述。
在得到相邻点对(以相邻点对1为例)对应的所有数据对之后，可以确定每个数据对对应的通行时长，即样本车辆处于位置点B的采集时刻(基于样本车辆处于位置点B的历史数据获知)与样本车辆处于位置点A的采集时刻(基于样本车辆处于位置点A的历史数据获知)之间的差值。比如说，参见表1所示，样本车辆s1对应的数据对包括“车牌标识s1+pt11+位置点A”和“车牌标识s1+pt12+位置点B”，这个数据对对应的通行时长是pt12与pt11的差值。
在得到每个数据对对应的通行时长之后，就可以按照通行时长从小到大的顺序对所有数据对进行排序，基于排序结果，可以过滤前面X1个数据对，并过滤后面X2个数据对。针对待过滤的每个数据对，该数据对可以包括样本车辆处于位置点A的历史数据和样本车辆处于位置点B的历史数据，也就是说，需要过滤样本车辆处于位置点A的历史数据和样本车辆处于位置点B的历史数据。
示例性的，X1可以根据经验配置，如1、2、3等，也可以基于所有数据对的通行时长确定，本申请对此不做限制，X2可以根据经验配置，如1、2、3等，也可以基于所有数据对的通行时长确定，本申请对此不做限制。在基于所有数据对的通行时长确定X1和X2的取值时，参见图4所示，可以采用如下方式实现：
对所有数据对的通行时长进行排序，依次得到t1、t2、t3、t4、t5、t6、t7，若t2与t1的差值大于预设阈值，且t3与t2的差值不大于预设阈值，则表示t2是通行时长呈现稳定的线性增长时的第一个通行时长，需要过滤的就是t1对应的数据对，即X1的取值为1。若t3与t2的差值大于预设阈值，且t4与t3的差值不大于预设阈值，则表示t3是通行时长呈现稳定的线性增长时的第一个通行时长，需要过滤的就是t1和t2对应的数据对，即X1的取值为2，以此类推。
若t7与t6的差值大于预设阈值，且t6与t5的差值不大于预设阈值，则表示t6是通行时长呈现稳定线性增长时的最后一个通行时长，需要过滤的是t7对应的数据对，即X2的取值为1。若t6与t5的差值大于预设阈值，且t5与t4的差值不大于预设阈值，则表示t5是通行时长呈现稳定线性增长时的最后一个通行时长，需要过滤的是t7和t6对应的数据对，即X2的取值为2，以此类推。
显然，可以采用方式3对相邻点对1的数据对进行过滤，即过滤通行时长较小的X1 个数据对，并过滤通行时长较大的X2个数据对，并保留通行时长位于中间的数据对(即过滤剩余的数据对)，同理，可以采用方式3对其它相邻点对(如相邻点对2、相邻点对3等)的数据对进行过滤，在此不再赘述。
综上所述，通过过滤通行时长较小的X1个数据对，并过滤通行时长较大的X2个数据对，可以去除无效数据对的干扰，即通行时长较小和通行时长较大的数据对均是无效数据对，从而减少不合理数据的干扰，保留最合适的数据对。
综上所述，基于方式1、方式2和方式3，均可以对历史数据进行过滤，当然，也可以采用其它方式对历史数据进行过滤，本申请对此不做限制。将过滤剩余的历史数据确定为与时间段1对应的样本数据，基于样本数据执行后续步骤。
步骤302、基于样本数据(即与时间段1对应的样本数据)确定各相邻点对的通行时长，该相邻点对可以包括相邻两个位置点，该样本数据可以包括在时间段1内样本车辆处于各位置点的采集时刻，该通行时长可以是基于样本车辆处于相邻两个位置点的采集时刻确定。
示例性的，针对任一相邻点对来说，可以获取与该相邻点对对应的M个数据对，针对每个数据对，该数据对可以包括样本车辆处于该相邻点对中相邻两个位置点的采集时刻，即该数据对包括两个采集时刻(这两个采集时刻均是位于时间段1内的采集时刻)。针对每个数据对，基于该数据对中的两个采集时刻确定通行时长，即该数据对对应的通行时长。在一个示例中，可以将M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
比如说，可以基于所有样本车辆的样本数据得到多个相邻点对，每个相邻点对均包括相邻两个位置点，相邻点对的获取方式参见步骤301，将历史数据替换为样本数据即可，在此不再赘述。在得到多个相邻点对后，针对每个相邻点对，可以基于所有样本车辆的样本数据确定该相邻点对对应的所有数据对，即与该相邻点对对应的M个数据对，每个数据对可以包括样本车辆处于该相邻点对中相邻两个位置点的采集时刻。例如，以相邻点对1为例，相邻点对1包括位置点A和位置点B，可以获取与相邻点对1对应的M个数据对，且数据对包括样本车辆处于位置点A的样本数据(如采集时刻)和样本车辆处于位置点B的样本数据(如采集时刻)，数据对的获取方式参见步骤301，在此不再赘述。
示例性的，针对每个相邻点对，在得到该相邻点对对应的M个数据对之后，针对每个数据对，基于该数据对中的两个采集时刻确定该数据对对应的通行时长，并将M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
以相邻点对1为例，针对M个数据对中的每个数据对，计算该数据对中样本车辆处于位置点B的采集时刻与该数据对中样本车辆处于位置点A的采集时刻之间的差值，该差值就是该数据对对应的通行时长，从而能够得到M个通行时长，然后，将M个通行时长中的最小值作为相邻点对1的通行时长。
综上所述，在步骤302中，可以得到每个相邻点对的通行时长。
步骤303、针对网络拓扑中的任意两个位置点(可以是相邻位置点，也可以不是相邻位置点)，确定这两个位置点之间的目标路径经过的至少一个相邻点对，并基于该目标路径经过的相邻点对的通行时长及相邻点对数量，确定这两个位置点之间的与时间段1对应的最小通行时长和最小位置点数量。
示例性的，针对网络拓扑中任意两个位置点，基于这两个位置点之间的各路径经过的所有相邻点对的通行时长之和，可以选取出通行时长之和最小的路径作为目标路径，将目标路径经过的所有相邻点对的通行时长之和确定为最小通行时长，并将目标路径经过的所有相邻点对的总数量确定为最小位置点数量。
比如说，参见图1所示，网络拓扑包括位置点A、位置点B、位置点C、位置点D、位置点E和位置点F，在得到所有相邻点对之后，可以构建图5所示的网络拓扑，该网 络拓扑用于显示所有相邻点对，及相邻点对的通行时长。
参见图5所示，具有直接连接关系的两个位置点组成相邻点对，且相邻点对具有方向，如箭头方向是位置点A指向位置点B时，表示相邻点对“位置点A-位置点B”，即从位置点A行驶到位置点B，箭头方向是位置点B指向位置点A时，表示相邻点对“位置点B-位置点A”，即从位置点B行驶到位置点A。
参见图5所示，t1是相邻点对“位置点A-位置点B”的通行时长，t2是相邻点对“位置点B-位置点C”的通行时长，t3是相邻点对“位置点C-位置点D”的通行时长，t4是相邻点对“位置点E-位置点D”的通行时长，以此类推。
针对网络拓扑中的任意两个位置点，如位置点A-位置点B(或位置点C、位置点D、位置点E、位置点F)、位置点B-位置点A(或位置点C、位置点D、位置点E、位置点F)、位置点C-位置点A(或位置点B、位置点D、位置点E、位置点F)、位置点D-位置点A(或位置点B、位置点C、位置点E、位置点F)、位置点E-位置点A(或位置点B、位置点C、位置点D、位置点F)、位置点F-位置点A(或位置点B、位置点C、位置点D、位置点E)，后续以位置点A-位置点D为例，可以采用如下方式获取最小通行时长和最小位置点数量：
参见图5所示，位置点A-位置点D对应的路径包括A-B-C-D和A-E-D。路径A-B-C-D对应的通行时长之和是t1+t2+t3，路径A-E-D对应的通行时长之和是t7+t4。若t1+t2+t3小于t7+t4，则位置点A与位置点D之间的目标路径是A-B-C-D，若t1+t2+t3大于t7+t4，则位置点A与位置点D之间的目标路径是A-E-D。
比如说，若目标路径是A-B-C-D，则基于该目标路径经过的所有相邻点对的通行时长之和确定该最小通行时长，比如说，最小通行时长可以是t1+t2+t3，最小通行时长也可以是(t1+t2+t3)*w，w为大于0且小于1的数值。以及，基于该目标路径经过的所有相邻点对(包括相邻点对“位置点A-位置点B”、“位置点B-位置点C”和“位置点C-位置点D”)的总数量3确定最小位置点数量，比如说，最小位置点数量可以为3。若目标路径是A-E-D，则基于该目标路径经过的所有相邻点对的通行时长之和确定该最小通行时长，如最小通行时长可以是t7+t4，最小通行时长也可以是(t7+t4)*w。以及，基于该目标路径经过的所有相邻点对(包括相邻点对“位置点A-位置点E”和“位置点E-位置点D”)的总数量2确定最小位置点数量，比如说，最小位置点数量可以为2。
需要说明的是，在实际应用中，考虑到通行时长之和最小的路径，通常是行车速度最快的路径，而实际道路情况是红绿灯数量越少时，即位置点数量越少，行车速度越快。基于上述情况，本实施例中，可以将通行时长之和最小的路径经过的所有相邻点对的总数量确定为最小位置点数量。
综上所述，针对时间段1，可以确定网络拓扑中的任意两个位置点的最小通行时长和最小位置点数量，针对时间段2，可以确定网络拓扑中的任意两个位置点之间的最小通行时长和最小位置点数量，以此类推，在此不再重复赘述。
在一种可能的实施方式中，考虑到道路环境的变化，会影响车辆的行驶情况，因此，可以周期性确定最小通行时长和最小位置点数量，以保证最小通行时长和最小位置点数量更贴近实际的道路环境，也就是说，步骤301至步骤303可以周期性执行。在此基础上，若当前时间点与上一次数据更新时间点(即上一次确定最小通行时长和最小位置点数量的时间点)之间的时长达到预设更新时长(如一个月、两个月等)，则确定已满足数据更新条件，需要重新执行步骤301至步骤303，得到更新后的最小通行时长和最小位置点数量。
在一种可能的实施方式中，异常车辆(即套牌车辆)的出行时间和出行轨迹具有非常大的随机性，理论上，在两个位置点之间出现异常车辆的概率非常小，几乎为0，也就是说，在两个位置点之间出现异常车辆的次数应该小于异常次数阈值。若出现异常车辆的次数大于异常次数阈值，则表示两个位置点之间的道路环境发生变化(如路况变好、 或路况变差等)，需要重新确定两个位置点之间的最小通行时长和最小位置点数量，以避免误判的情况发生。在此基础上，若两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则确定已满足数据更新条件，需要重新执行步骤301至步骤303，得到更新后的最小通行时长和最小位置点数量。
需要说明的是，若两个位置点之间经过的异常车辆的总数量大于异常次数阈值，在重新确定最小通行时长和最小位置点数量时，针对步骤301，可以采用方式1至方式3中的至少一种对历史数据进行过滤，在此基础上，也可以采用方式4对历史数据进行过滤，过滤完成后将剩余历史数据作为样本数据。
方式4、针对两个位置点(可以是相邻的两个位置点，也可以是不相邻的两个位置点，本申请对此不做限制)，若这两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则过滤这两个位置点对应的所有数据对。
参见上述描述，在两个位置点之间出现异常车辆的概率非常小，几乎为0，若出现异常车辆的次数大于异常次数阈值，则表示两个位置点之间的道路环境发生变化，也就是说，这两个位置点之间的数据可能为无效数据。因此，可以获取这两个位置点对应的所有数据对，并过滤这两个位置点对应的所有数据对。关于两个位置点对应的所有数据对的获取方式，可以参见步骤301，在此不再赘述。
在一种可能的实施方式中，基于每个时间段对应的最小通行时长和最小位置点数量(如时间段1对应的最小通行时长和最小位置点数量、时间段2对应的最小通行时长和最小位置点数量、以此类推)，可以检测车辆是否为异常车辆。将待检测的车辆称为目标车辆，参见图6所示，基于每个时间段对应的最小通行时长和最小位置点数量采用如下步骤检测目标车辆是否为异常车辆。
步骤601、若目标车辆从第一位置点行驶到第二位置点，则获取目标车辆处于第一位置点的第一时间点、目标车辆处于第二位置点的第二时间点、目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量。
示例性的，可以从历史数据库中获取与目标车辆对应的所有数据记录，每条数据记录均包括车牌标识s1、采集时刻、位置点标识，按照采集时刻从前到后的顺序对这些数据记录进行排序，假设排序结果为：车牌标识s1+pt1+位置点A、车牌标识s1+pt2+位置点B、车牌标识s1+pt3+位置点F、车牌标识s1+pt4+位置点D、车牌标识s1+pt5+位置点E、车牌标识s1+pt6+位置点C。
第一位置点是所有位置点中的任一位置点，如位置点B，第二位置点是所有位置点中的任一位置点，如位置点F，那么，目标车辆处于第一位置点的第一时间点就是pt2，目标车辆处于第二位置点的第二时间点就是pt3，目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量就是1。
在一种可能的实施方式中，步骤601之前，可以获取目标车辆对应的车辆特征(如车辆颜色、车辆型号、车辆外观等)，从特征管理库(用于存储正常车辆的正常车辆特征)中选取与目标车辆的车牌标识对应的正常车辆特征，若目标车辆对应的车辆特征与该正常车辆特征不同，如车辆颜色不同，则直接确定目标车辆是异常车辆，不再执行步骤601，若目标车辆对应的车辆特征与该正常车辆特征相同，则执行步骤601及后续步骤，确定目标车辆是否为异常车辆。
步骤602、基于第一时间点从所有时间段中选取出目标时间段，比如说，将第一时间点所处的时间段作为目标时间段。或者，基于第二时间点从所有时间段中选取出目标时间段，比如说，将第二时间点所处的时间段作为目标时间段。
步骤603、确定第一位置点与第二位置点之间的与目标时间段对应的最小通行时长和最小位置点数量。比如说，目标时间段是时间段1，则确定位置点B与位置点F之间的与时间段1对应的最小通行时长和最小位置点数量。
步骤604、基于该第一时间点、该第二时间点、该目标位置点数量、该最小通行时 长和该最小位置点数量，确定目标车辆是异常车辆还是正常车辆。
情况1、若该第二时间点与该第一时间点之间的差值小于第一位置点与第二位置点之间的最小通行时长，则可以确定目标车辆是异常车辆。
比如说，目标车辆在第一时间点pt2处于位置点B，在第二时间点pt3处于位置点F，位置点B到位置点F的最小通行时长为已知，若pt3与pt2的差值小于该最小通行时长，则说明目标车辆无法在这段时间内从位置点B行驶到位置点F，也就是说，目标车辆从位置点B行驶到位置点F是在不合理的时长内到达，位置点B的车辆与位置点F的车辆不是同一车辆，即目标车辆是异常车辆。
情况2、若该第二时间点与该第一时间点之间的差值不小于第一位置点与第二位置点之间的最小通行时长，且第一位置点与第二位置点之间的最小位置点数量大于目标位置点数量与预设数量阈值之和，则确定目标车辆是异常车辆。
比如说，目标车辆在第一时间点pt2处于位置点B，在第二时间点pt3处于位置点F，从位置点B行驶到位置点F所经过的目标位置点数量是1，位置点B到位置点F的最小通行时长为已知，位置点B到位置点F的最小位置点数量为已知，若pt3与pt2的差值不小于该最小通行时长，则需要分析位置点数量。
假设位置点B到位置点F的最小位置点数量是5，表示最少经过5个位置点，才能够从位置点B行驶到位置点F，且假设预设数量阈值是2(可以根据经验配置)，那么，由于从位置点B行驶到位置点F的目标位置点数量(也即实际位置点数量)是1，且最小位置点数量5大于目标位置点数量1与预设数量阈值2之和，因此，说明目标车辆途径的位置点在地理空间上是不符合实际情况的，位置点B的车辆与位置点F的车辆不是同一车辆，即目标车辆是异常车辆。
情况3、若该第二时间点与该第一时间点之间的差值不小于第一位置点与第二位置点之间的最小通行时长，且第一位置点与第二位置点之间的最小位置点数量不大于目标位置点数量与预设数量阈值之和，则确定目标车辆是正常车辆。
由以上技术方案可见，本申请实施例中，可以预先对历史数据进行大数据分析，推算任意两个位置点之间符合实际用车的最小通行时长和最小位置点数量，无需人工治理数据。分别从高峰、平峰、低峰等时间段来推算最小通行时长和最小位置点数量，最大可能的贴近车辆在该段时间内行车环境与行车速度，解决潮汐车道、早晚高峰时间差异大、分时单循环等一系列影响车辆行车速度与轨迹的问题。通过对历史数据进行大数据分析，推算在任意两个位置点之间通过时需要经过的最小位置点数量，通过比较实际过车时两个位置点之间途径的目标位置点数量与大数据推算出的最小位置点数量的差异(是基于地理空间的比较)，解决套牌车场景中正常车辆和异常车辆不在相同时间段内出现在路上的情况。时空分析和地理空间分析所使用的参照数据(即最小通行时长和最小位置点数量)均是大数据定期推算所得，会随着道路环境等变化自动调整数据，更贴近实际的行车情况。在使用数据层面，会过滤掉与之前分析出的异常车辆相关的过车记录，减少异常车辆不合理的过车数据对大数据推算路径时的影响。对时空分析、地理空间分析的结果再进行二次匹配，针对一定时间段内相同位置点出现异常车辆次数大于阈值的情况，会进行过滤，减少因为道路环境调整等因素导致的过车时间发生变化而对分析结果的准确度造成的影响。
基于与上述方法同样的申请构思，本申请实施例中提出一种异常车辆的检测装置，参见图7所示，为所述装置的结构示意图，所述装置可以包括：
获取模块71，用于若目标车辆从第一位置点行驶到第二位置点，则获取目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点、所述目标车辆从第一位置点行驶到第二位置点所经过的目标位置点数量；
确定模块72，用于若所述第二时间点与所述第一时间点之间的差值小于所述第一位置点与所述第二位置点之间的最小通行时长，则确定所述目标车辆是异常车辆；若所 述第二时间点与所述第一时间点之间的差值不小于所述最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是异常车辆。
在一种可能的实施方式中，所述确定模块72还用于：若所述第二时间点与所述第一时间点之间的差值不小于所述第一位置点与所述第二位置点之间的最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量不大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是正常车辆。
在一种可能的实施方式中，所述获取模块71还用于：若指定统计周期被划分为多个时间段，则基于所述第一时间点或者所述第二时间点，从所述多个时间段中选取出目标时间段；其中，所述最小通行时长是与所述目标时间段对应的最小通行时长；所述最小位置点数量是与所述目标时间段对应的最小位置点数量。
在一种可能的实施方式中，所述确定模块72还用于：确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，所述统计时间段为指定统计周期的完整时间段，或者，指定统计周期被划分为多个时间段时，所述统计时间段为所述多个时间段中的任一时间段；所述确定模块确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量时具体用于：确定第一位置点与第二位置点之间的目标路径经过的至少一个相邻点对，基于所述目标路径经过的至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，所述至少一个相邻点对的通行时长是基于历史数据库中的与统计时间段对应的样本数据确定，所述样本数据包括在所述统计时间段内样本车辆处于网络拓扑中各位置点的采集时刻，相邻点对包括相邻两个位置点，相邻点对的通行时长是基于样本车辆处于相邻两个位置点的采集时刻确定。
在一种可能的实施方式中，所述确定模块72基于历史数据库中的与统计时间段对应的样本数据确定所述至少一个相邻点对的通行时长时具体用于：针对任一相邻点对，获取与该相邻点对对应的M个数据对，各数据对包括在所述统计时间段内所述样本车辆处于该相邻点对中相邻两个位置点的采集时刻；针对各数据对，基于该数据对中的两个采集时刻确定通行时长；将M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
在一种可能的实施方式中，所述确定模块72基于所述目标路径经过的所述至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量时具体用于：基于所述第一位置点与所述第二位置点之间的各路径经过的所有相邻点对的通行时长之和，选取通行时长之和最小的路径作为目标路径，将所述目标路径经过的所有相邻点对的通行时长之和确定为与所述统计时间段对应的最小通行时长，并将所述目标路径经过的所有相邻点对的总数量确定为与所述统计时间段对应的最小位置点数量。
在一种可能的实施方式中，所述确定模块72还用于：从历史数据库中选取与所述统计时间段对应的历史数据，所述历史数据包括在所述统计时间段内所述样本车辆处于各位置点的采集时刻；对历史数据进行过滤，将过滤剩余的历史数据确定为所述样本数据；其中，对所述历史数据进行过滤包括以下至少一种：针对任一样本车辆，若该样本车辆经过相邻两个位置点的通行时长小于预设时长阈值，则过滤该样本车辆经过所述相邻两个位置点的历史数据；针对相邻两个位置点，若所述相邻两个位置点对应的数据对的总数量小于预设次数阈值，则过滤所述相邻两个位置点对应的所有数据对；其中，所述数据对包括样本车辆经过所述相邻两个位置点的历史数据；针对相邻两个位置点，获取所述相邻两个位置点对应的所有数据对；基于各数据对对应的通行时长，过滤通行时长小的X1个数据对，并过滤通行时长大的X2个数据对，所述X1和所述X2均为正整 数；针对两个位置点，若所述两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则过滤所述两个位置点对应的所有数据对。
在一种可能的实施方式中，所述确定模块72确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量时具体用于：确定是否已满足数据更新条件；如果是，则确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，若所述第一位置点与所述第二位置点之间经过的异常车辆的总数量大于异常次数阈值，则确定已满足数据更新条件；或，若当前时间点与上一次数据更新时间点之间的时长达到预设更新时长，则确定已满足数据更新条件，所述上一次数据更新时间点是上一次确定最小通行时长和最小位置点数量的时间点。
基于与上述方法同样的申请构思，本申请实施例中提出一种异常车辆的检测设备(如管理设备)，参见图8所示，异常车辆的检测设备可以包括：处理器81和机器可读存储介质82，所述机器可读存储介质82存储有能够被所述处理器81执行的机器可执行指令；所述处理器81用于执行机器可执行指令，以实现本申请上述示例公开的异常车辆的检测方法。
基于与上述方法同样的申请构思，本申请实施例还提供一种机器可读存储介质，所述机器可读存储介质上存储有若干计算机指令，所述计算机指令被处理器执行时，能够实现本申请上述示例公开的异常车辆的检测方法。
其中，上述机器可读存储介质可以是任何电子、磁性、光学或其它物理存储装置，可以包含或存储信息，如可执行指令、数据，等等。例如，机器可读存储介质可以是：RAM(Radom Access Memory，随机存取存储器)、易失存储器、非易失性存储器、闪存、存储驱动器(如硬盘驱动器)、固态硬盘、任何类型的存储盘(如光盘、dvd等)，或者类似的存储介质，或者它们的组合。
基于与上述方法同样的申请构思，本申请实施例还提供一种计算机程序，所述计算机程序存储于机器可读存储介质，当处理器执行所述计算机程序时，促使处理器实现本申请上述示例公开的异常车辆的检测方法。
上述实施例阐明的系统、装置、模块或单元，具体可以由计算机或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机，计算机的具体形式可以是个人计算机、膝上型计算机、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件收发设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任意几种设备的组合。
为了描述的方便，描述以上装置时以功能分为各种单元分别描述。当然，在实施本申请时可以把各单元的功能在同一个或多个软件和/或硬件中实现。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、装置、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可以由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其它可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其它可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
而且，这些计算机程序指令也可以存储在能引导计算机或其它可编程数据处理设备 以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或者多个流程和/或方框图一个方框或者多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其它可编程数据处理设备上，使得在计算机或者其它可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其它可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
以上所述仅为本申请的实施例而已，并不用于限制本申请。对于本领域技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本申请的权利要求范围之内。

Claims (17)

  1. 一种异常车辆的检测方法，其特征在于，所述方法包括：
    若目标车辆从第一位置点行驶到第二位置点，则获取所述目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点、所述目标车辆从所述第一位置点行驶到所述第二位置点所经过的目标位置点数量；
    若所述第二时间点与所述第一时间点之间的差值小于所述第一位置点与所述第二位置点之间的最小通行时长，则确定所述目标车辆是异常车辆；
    若所述第二时间点与所述第一时间点之间的差值不小于所述最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是异常车辆。
  2. 根据权利要求1所述的方法，其特征在于，
    所述获取所述目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点之后，所述方法还包括：
    若指定统计周期被划分为多个时间段，则基于所述第一时间点或者所述第二时间点，从所述多个时间段中选取出目标时间段；
    其中，所述最小通行时长是与所述目标时间段对应的最小通行时长；所述最小位置点数量是与所述目标时间段对应的最小位置点数量。
  3. 根据权利要求1或2所述的方法，其特征在于，
    所述获取所述目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点之前，所述方法还包括：
    确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，所述统计时间段为指定统计周期的完整时间段，或者，所述指定统计周期被划分为多个时间段时，所述统计时间段为所述多个时间段中的任一时间段；
    其中，所述确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，包括：
    确定所述第一位置点与所述第二位置点之间的目标路径经过的至少一个相邻点对，基于所述目标路径经过的所述至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量；
    其中，相邻点对的通行时长是基于历史数据库中的与所述统计时间段对应的样本数据确定，所述样本数据包括在所述统计时间段内样本车辆处于网络拓扑中各位置点的采集时刻，所述相邻点对包括相邻两个位置点，所述相邻点对的通行时长是基于所述样本车辆处于相邻两个位置点的采集时刻确定。
  4. 根据权利要求3所述的方法，其特征在于，基于历史数据库中的与所述统计时间段对应的样本数据确定所述至少一个相邻点对的通行时长的过程包括：
    针对任一相邻点对，获取与该相邻点对对应的M个数据对，各数据对包括在所述统计时间段内所述样本车辆处于该相邻点对中相邻两个位置点的采集时刻；
    针对各数据对，基于该数据对中的两个采集时刻确定通行时长；
    将所述M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
  5. 根据权利要求3所述的方法，其特征在于，所述基于所述目标路径经过的所述至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量，包括：
    基于所述第一位置点与所述第二位置点之间的各路径经过的所有相邻点对的通行时长之和，选取通行时长之和最小的路径作为目标路径；
    将所述目标路径经过的所有相邻点对的通行时长之和确定为与所述统计时间段对 应的最小通行时长，并将所述目标路径经过的所有相邻点对的总数量确定为与所述统计时间段对应的最小位置点数量。
  6. 根据权利要求3所述的方法，其特征在于，所述方法还包括：
    从历史数据库中选取与所述统计时间段对应的历史数据，所述历史数据包括在所述统计时间段内所述样本车辆处于各位置点的采集时刻；
    对所述历史数据进行过滤，将过滤剩余的历史数据确定为所述样本数据；其中，对所述历史数据进行过滤包括以下至少一种：
    针对任一样本车辆，若该样本车辆经过相邻两个位置点的通行时长小于预设时长阈值，则过滤该样本车辆经过所述相邻两个位置点的历史数据；
    针对相邻两个位置点，若所述相邻两个位置点对应的数据对的总数量小于预设次数阈值，则过滤所述相邻两个位置点对应的所有数据对；其中，所述数据对包括样本车辆经过所述相邻两个位置点的历史数据；
    针对相邻两个位置点，获取所述相邻两个位置点对应的所有数据对；基于各数据对对应的通行时长，过滤通行时长小的X1个数据对，并过滤通行时长大的X2个数据对，所述X1和所述X2均为正整数；
    针对两个位置点，若所述两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则过滤所述两个位置点对应的所有数据对。
  7. 根据权利要求3所述的方法，其特征在于，确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量，包括：
    确定是否已满足数据更新条件；
    如果是，则确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量；
    其中，若所述第一位置点与所述第二位置点之间经过的异常车辆的总数量大于异常次数阈值，则确定已满足数据更新条件；或，若当前时间点与上一次数据更新时间点之间的时长达到预设更新时长，则确定已满足所述数据更新条件，所述上一次数据更新时间点是上一次确定最小通行时长和最小位置点数量的时间点。
  8. 一种异常车辆的检测装置，其特征在于，所述装置包括：
    获取模块，用于若目标车辆从第一位置点行驶到第二位置点，则获取目标车辆处于第一位置点的第一时间点、所述目标车辆处于第二位置点的第二时间点、所述目标车辆从所述第一位置点行驶到所述第二位置点所经过的目标位置点数量；
    确定模块，用于若所述第二时间点与所述第一时间点之间的差值小于所述第一位置点与所述第二位置点之间的最小通行时长，则确定所述目标车辆是异常车辆；若所述第二时间点与所述第一时间点之间的差值不小于所述最小通行时长，且所述第一位置点与所述第二位置点之间的最小位置点数量大于所述目标位置点数量与预设数量阈值之和，则确定所述目标车辆是异常车辆。
  9. 根据权利要求8所述的装置，其特征在于，所述获取模块还用于：
    若指定统计周期被划分为多个时间段，则基于所述第一时间点或者所述第二时间点，从所述多个时间段中选取出目标时间段；其中，所述最小通行时长是与所述目标时间段对应的最小通行时长；所述最小位置点数量是与所述目标时间段对应的最小位置点数量。
  10. 根据权利要求8或9所述的装置，其特征在于，所述确定模块还用于：
    确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量；其中，所述统计时间段为指定统计周期的完整时间段，或者，所述指定统计周期被划分为多个时间段时，所述统计时间段为所述多个时间段中的任一时间段；
    所述确定模块确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量时具体用于：
    确定所述第一位置点与所述第二位置点之间的目标路径经过的至少一个相邻点对，基于所述目标路径经过的所述至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量；
    其中，相邻点对的通行时长是基于历史数据库中的与所述统计时间段对应的样本数据确定，所述样本数据包括在所述统计时间段内样本车辆处于网络拓扑中各位置点的采集时刻，所述相邻点对包括相邻两个位置点，所述相邻点对的通行时长是基于所述样本车辆处于相邻两个位置点的采集时刻确定。
  11. 根据权利要求10所述的装置，其特征在于，所述确定模块基于历史数据库中的与所述统计时间段对应的样本数据确定所述至少一个相邻点对的通行时长时具体用于：
    针对任一相邻点对，获取与该相邻点对对应的M个数据对，各数据对包括在所述统计时间段内所述样本车辆处于该相邻点对中相邻两个位置点的采集时刻；
    针对各数据对，基于该数据对中的两个采集时刻确定通行时长；
    将所述M个数据对对应的通行时长中的最小值确定为该相邻点对的通行时长。
  12. 根据权利要求10所述的装置，其特征在于，所述确定模块基于所述目标路径经过的所述至少一个相邻点对的通行时长及相邻点对数量，确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量时具体用于：
    基于所述第一位置点与所述第二位置点之间的各路径经过的所有相邻点对的通行时长之和，选取通行时长之和最小的路径作为目标路径；
    将所述目标路径经过的所有相邻点对的通行时长之和确定为与所述统计时间段对应的最小通行时长，并将所述目标路径经过的所有相邻点对的总数量确定为与所述统计时间段对应的最小位置点数量。
  13. 根据权利要求10所述的装置，其特征在于，所述确定模块还用于：
    从历史数据库中选取与所述统计时间段对应的历史数据，所述历史数据包括在所述统计时间段内所述样本车辆处于各位置点的采集时刻；
    对所述历史数据进行过滤，将过滤剩余的历史数据确定为所述样本数据；其中，对所述历史数据进行过滤包括以下至少一种：
    针对任一样本车辆，若该样本车辆经过相邻两个位置点的通行时长小于预设时长阈值，则过滤该样本车辆经过所述相邻两个位置点的历史数据；
    针对相邻两个位置点，若所述相邻两个位置点对应的数据对的总数量小于预设次数阈值，则过滤所述相邻两个位置点对应的所有数据对；其中，所述数据对包括样本车辆经过所述相邻两个位置点的历史数据；
    针对相邻两个位置点，获取所述相邻两个位置点对应的所有数据对；基于各数据对对应的通行时长，过滤通行时长小的X1个数据对，并过滤通行时长大的X2个数据对，所述X1和所述X2均为正整数；
    针对两个位置点，若所述两个位置点之间经过的异常车辆的总数量大于异常次数阈值，则过滤所述两个位置点对应的所有数据对。
  14. 根据权利要求10所述的装置，其特征在于，
    所述确定模块确定所述第一位置点与所述第二位置点之间的与统计时间段对应的最小通行时长和最小位置点数量时具体用于：
    确定是否已满足数据更新条件；
    如果是，则确定所述第一位置点与所述第二位置点之间的与所述统计时间段对应的最小通行时长和最小位置点数量；
    其中，若所述第一位置点与所述第二位置点之间经过的异常车辆的总数量大于异常 次数阈值，则确定已满足数据更新条件；或，若当前时间点与上一次数据更新时间点之间的时长达到预设更新时长，则确定已满足所述数据更新条件，所述上一次数据更新时间点是上一次确定最小通行时长和最小位置点数量的时间点。
  15. 一种异常车辆的检测设备，其特征在于，所述异常车辆的检测设备包括：处理器和机器可读存储介质，所述机器可读存储介质存储有能够被所述处理器执行的机器可执行指令；其中，所述处理器用于执行所述机器可执行指令，以实现权利要求1-7任一项所述的方法步骤。
  16. 一种非暂时性机器可读存储介质，其特征在于，所述非暂时性机器可读存储介质存储有能够被处理器执行的机器可执行指令；其中，所述处理器用于执行所述机器可执行指令，以实现权利要求1-7任一项所述的方法步骤。
  17. 一种计算机程序，其特征在于，所述计算机程序存储于机器可读存储介质，当处理器执行所述计算机程序时实现权利要求1-7任一项所述的方法步骤。
PCT/CN2022/131621 2021-11-19 2022-11-14 异常车辆的检测方法、装置及设备 WO2023088201A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111410662.2A CN114140780A (zh) 2021-11-19 2021-11-19 一种异常车辆的检测方法、装置及设备
CN202111410662.2 2021-11-19

Publications (1)

Publication Number Publication Date
WO2023088201A1 true WO2023088201A1 (zh) 2023-05-25

Family

ID=80391658

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/131621 WO2023088201A1 (zh) 2021-11-19 2022-11-14 异常车辆的检测方法、装置及设备

Country Status (2)

Country Link
CN (1) CN114140780A (zh)
WO (1) WO2023088201A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114140780A (zh) * 2021-11-19 2022-03-04 杭州海康威视数字技术股份有限公司 一种异常车辆的检测方法、装置及设备
CN114495031A (zh) * 2022-03-31 2022-05-13 青岛海信网络科技股份有限公司 一种车牌信息校正方法、设备及装置
CN115601974B (zh) * 2022-09-29 2023-06-16 广州天长信息技术有限公司 一种高速路套牌车辆确定方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107204114A (zh) * 2016-03-18 2017-09-26 中兴通讯股份有限公司 一种车辆异常行为的识别方法及装置
CN111369801A (zh) * 2019-08-27 2020-07-03 杭州海康威视系统技术有限公司 车辆识别方法、装置、设备和存储介质
CN111402574A (zh) * 2018-12-13 2020-07-10 阿里巴巴集团控股有限公司 车辆检测方法、装置、设备和存储介质
CN111767776A (zh) * 2019-12-28 2020-10-13 西安宇视信息科技有限公司 一种异常车牌推选方法及装置
CN114140780A (zh) * 2021-11-19 2022-03-04 杭州海康威视数字技术股份有限公司 一种异常车辆的检测方法、装置及设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107204114A (zh) * 2016-03-18 2017-09-26 中兴通讯股份有限公司 一种车辆异常行为的识别方法及装置
CN111402574A (zh) * 2018-12-13 2020-07-10 阿里巴巴集团控股有限公司 车辆检测方法、装置、设备和存储介质
CN111369801A (zh) * 2019-08-27 2020-07-03 杭州海康威视系统技术有限公司 车辆识别方法、装置、设备和存储介质
CN111767776A (zh) * 2019-12-28 2020-10-13 西安宇视信息科技有限公司 一种异常车牌推选方法及装置
CN114140780A (zh) * 2021-11-19 2022-03-04 杭州海康威视数字技术股份有限公司 一种异常车辆的检测方法、装置及设备

Also Published As

Publication number Publication date
CN114140780A (zh) 2022-03-04

Similar Documents

Publication Publication Date Title
WO2023088201A1 (zh) 异常车辆的检测方法、装置及设备
Wang et al. A simple baseline for travel time estimation using large-scale trip data
CN111402580B (zh) 车辆行驶轨迹预测方法、装置及电子设备
WO2020206861A1 (zh) 基于YOLO v3的针对交通枢纽关键物体的检测方法
Mallikarjuna et al. Traffic data collection under mixed traffic conditions using video image processing
JP6819118B2 (ja) 移動物体の計数装置及び方法
US9754178B2 (en) Long-term static object detection
CN103646257B (zh) 一种基于视频监控图像的行人检测和计数方法
CN104252627A (zh) Svm分类器训练样本获取方法、训练方法及其系统
CN104408406A (zh) 基于帧差法和减背景法的人员离岗检测方法
CN110348143A (zh) 一种轨迹合并方法、装置及存储介质
WO2024067210A1 (zh) 一种交通状态确定方法、装置及设备
CN114428937A (zh) 一种基于时空图神经网络的时间序列数据预测方法
CN114647684A (zh) 基于stacking算法的流量预测方法、装置及相关设备
CN107862019B (zh) 一种基于大数据分析昼伏夜出车辆的方法及装置
CN109727268A (zh) 目标跟踪方法、装置、计算机设备和存储介质
CN108898617A (zh) 一种目标对象的跟踪方法和装置
CN111784742A (zh) 一种行人跨镜头追踪方法及装置
CN113656625B (zh) 一种确定人体空间域的方法、装置及电子设备
Ganesh et al. An intelligent video surveillance framework with big data management for Indian road traffic system
Tišljarić et al. Traffic state estimation using speed profiles and convolutional neural networks
Juhász et al. How do volunteer mappers use crowdsourced Mapillary street level images to enrich OpenStreetMap
CN111177465A (zh) 一种确定类别的方法及装置
Li et al. Relieve the congestion by shuttle bus in rush hours using aggregation clustering algorithm on group travel pattern
CN110097583A (zh) 图片筛选方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22894737

Country of ref document: EP

Kind code of ref document: A1

WD Withdrawal of designations after international publication