Track correction method and device and medium

Abstract

The invention discloses a track correction method and device and a medium. The method comprises the following steps: acquiring a track to be corrected; determining at least one section of abnormal trajectory included in the trajectory to be corrected; correcting the at least one section of abnormal trajectory to obtain at least one section of corrected trajectory; and determining a target trajectory based on the at least one section of corrected trajectory and a normal trajectory included in the trajectory to be corrected. According to the method, the negative influence caused by adjacent correction of the track points through the filtering window in the related technology can be overcome, so that the track correction requirement in practical application can be met.

Description

Technical Field

[0001] The present application relates to the field of positioning technology, and in particular to a trajectory correction method, device and medium. Background Art

[0002] When correcting the Global Positioning System (GPS) trajectory, it is usually achieved by filtering the entire GPS data without adding map information. In this method, a filter window is set to perform weighted correction on the positions of the trajectory points covered by the filter window, and the GPS trajectory is corrected based on the results of the weighted correction. However, the number of trajectory points within the length of the filter window and the state fluctuations of the trajectory points within the filter window will have a direct impact on the correction results of the trajectory points within the filter window. Therefore, the above correction method cannot meet the actual trajectory correction needs. Summary of the invention

[0003] Based on the above problems, the embodiments of the present application provide a trajectory correction method, device and medium.

[0004] The technical solution provided by the embodiment of the present application is as follows:

[0005] The present application embodiment first provides a trajectory correction method, the method comprising:

[0006] Obtain the trajectory to be corrected;

[0007] Determining at least one abnormal trajectory included in the trajectory to be corrected;

[0008] Correcting the at least one abnormal trajectory to obtain at least one corrected trajectory;

[0009] A target trajectory is determined based on the at least one corrected trajectory and a normal trajectory included in the trajectory to be corrected.

[0010] In some embodiments, if the trajectory to be corrected includes multiple abnormal trajectories, the kth abnormal trajectory and the k+1th abnormal trajectory include the mth normal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the number threshold; k and m are both integers greater than or equal to 1; the determining at least one abnormal trajectory included in the trajectory to be corrected includes:

[0011] In the process of correcting the kth abnormal trajectory, the trajectory segments in the trajectory to be corrected except the kth abnormal trajectory are analyzed to determine the at least one abnormal trajectory.

[0012] In some embodiments, if the trajectory to be corrected includes multiple abnormal trajectories, the kth abnormal trajectory and the k+1th abnormal trajectory include the mth normal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the number threshold; k and m are both integers greater than or equal to 1; and the correcting of the at least one abnormal trajectory to obtain at least one corrected trajectory includes:

[0013] In the process of correcting the kth abnormal trajectory, the k+nth abnormal trajectory and / or the kxth abnormal trajectory are corrected; wherein n is an integer greater than or equal to 1; and x is a positive integer less than k.

[0014] In some embodiments, the determining at least one abnormal trajectory included in the trajectory to be corrected includes:

[0015] Determine a trajectory point set included in the trajectory to be corrected; wherein the trajectory points in the trajectory point set include point types of the trajectory points; the point type is used to characterize whether the trajectory point is an abnormal point;

[0016] The trajectory points in the trajectory point set are statistically divided based on the point type to determine the at least one abnormal trajectory.

[0017] In some embodiments, the method further comprises:

[0018] Obtaining adjacent trajectory points of the p-th trajectory point in the trajectory to be corrected; wherein the adjacent trajectory points include P trajectory points before the p-th trajectory point and P trajectory points after the p-th trajectory point; p and P are both integers greater than 1;

[0019] Determine motion feature information of the p-th trajectory point and the adjacent trajectory points; wherein the motion feature information includes at least one of instantaneous speed, average speed, direction angle, and static point ratio;

[0020] The point type of the p-th trajectory point is determined based on the motion feature information.

[0021] In some embodiments, the determining the point type of the p-th trajectory point based on the motion feature information includes:

[0022] If the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory point is greater than a gradient threshold, the p-th trajectory point is determined to be an abnormal point.

[0023] In some embodiments, the performing statistical division of the trajectory points in the trajectory point set based on the point type to determine the at least one abnormal trajectory includes:

[0024] If the N consecutive trajectory points adjacent to the p-th trajectory point are all normal points, and the p-1-th trajectory point is an abnormal point, and the p+N+1-th trajectory point is the abnormal point, the trajectory interval formed by the trajectory points between the p-th trajectory point and the p+N-th trajectory point is divided into a normal trajectory; wherein p is an integer greater than 1; and N is greater than or equal to the quantity threshold;

[0025] Determine a trajectory segment in the trajectory to be corrected excluding the normal trajectory as the abnormal trajectory.

[0026] In some embodiments, the correcting the at least one abnormal trajectory includes:

[0027] Determine the abnormal type of the kth abnormal trajectory; wherein k is an integer greater than or equal to 1;

[0028] Determining a kth correction strategy based on the abnormal type of the kth abnormal trajectory;

[0029] The kth abnormal trajectory is corrected based on the kth correction strategy.

[0030] In some embodiments, determining the abnormal type of the kth abnormal trajectory includes:

[0031] The motion feature information of the trajectory points in the k-th abnormal trajectory is correlated and analyzed by a decision tree model to determine the abnormal type of the k-th abnormal trajectory.

[0032] In some embodiments, determining the abnormal type of the kth abnormal trajectory includes:

[0033] Acquire sample data; wherein the sample data includes motion feature information of trajectory points in a normal trajectory and motion feature information of trajectory points in an abnormal trajectory;

[0034] Training the cascade support vector machine in the initial state based on the sample data to obtain a cascade support vector machine;

[0035] The cascade support vector machine is used to perform association analysis on the motion feature information of the trajectory points in the k-th abnormal trajectory to determine the abnormal type of the k-th abnormal trajectory.

[0036] In some embodiments, determining the abnormal type of the kth abnormal trajectory includes:

[0037] Acquire a target normal trajectory adjacent to the k-th abnormal trajectory; wherein the target normal trajectory includes a normal trajectory adjacent to the k-th abnormal trajectory;

[0038] Determine a first position and a second position; wherein the first position includes the position of the starting trajectory point and the ending trajectory point of the kth abnormal trajectory; the second position includes the position of a normal point in the target normal trajectory that is adjacent to the starting trajectory point and the abnormal trajectory point of the kth abnormal trajectory;

[0039] If the distance between the first position and the second position is greater than a first distance threshold, the abnormality type is determined to be the first type.

[0040] In some embodiments, determining the abnormal type of the kth abnormal trajectory includes:

[0041] Performing clustering processing on the positions of the trajectory points in the k-th abnormal trajectory to obtain a cluster center;

[0042] If the distance between the trajectory point in the kth abnormal trajectory and the cluster center is less than a second distance threshold, the abnormal type is determined to be the second type.

[0043] In some embodiments, determining the abnormal type of the kth abnormal trajectory includes:

[0044] Acquire a target normal trajectory; wherein the target normal trajectory includes a normal trajectory adjacent to the kth abnormal trajectory;

[0045] Determine a trajectory point switching frequency; wherein the trajectory point switching frequency includes a switching frequency between a trajectory point in the kth abnormal trajectory and a trajectory point in the target normal trajectory;

[0046] If the switching frequency of the track points is greater than the switching threshold, it is determined that the abnormal feature is of the third type.

[0047] In some embodiments, determining the kth correction strategy based on the abnormal type of the kth abnormal trajectory includes:

[0048] If the k-th abnormal trajectory is of the first type, the k-th correction strategy is determined as:

[0049] Determine the distance similarity between the first distance and the second distance; wherein the first distance includes the distance between at least two normal points adjacent to the kth abnormal trajectory; and the second distance includes the distance between normal points in the kth abnormal trajectory;

[0050] Determine the angular similarity between the first angle and the second angle; wherein the first angle includes angles associated with at least two normal points adjacent to the kth abnormal trajectory; and the second angle includes angles associated with normal points in the kth abnormal trajectory;

[0051] A fitting parameter is determined based on the distance similarity and the angle similarity, and the kth abnormal trajectory is fitted and corrected based on the fitting parameter.

[0052] In some embodiments, the step of fitting and correcting the kth abnormal trajectory based on the fitting parameters includes:

[0053] If the fitting parameter is greater than a fitting threshold, based on the position of the normal point in the kth abnormal trajectory, fitting and correcting the kth abnormal trajectory;

[0054] If the fitting parameter is less than or equal to the fitting threshold, the kth abnormal trajectory is fitted and corrected based on at least two normal points adjacent to the kth abnormal trajectory.

[0055] In some embodiments, determining the kth correction strategy based on the abnormal type of the kth abnormal trajectory includes:

[0056] If the kth abnormal trajectory is of the second type, the kth correction strategy is determined as: clustering the positions of trajectory points in the kth abnormal trajectory to obtain a cluster center, and correcting the kth abnormal trajectory based on the cluster center.

[0057] In some embodiments, determining the kth correction strategy based on the abnormal type of the kth abnormal trajectory includes:

[0058] If the kth abnormal trajectory is of the third type, the kth correction strategy is determined as:

[0059] The kth abnormal trajectory is corrected based at least on the distribution information of the abnormal points and the normal points in the kth abnormal trajectory; wherein the distribution information includes time information and position information.

[0060] The present application also provides a trajectory correction device, the device comprising:

[0061] An acquisition module, used for acquiring the trajectory to be corrected;

[0062] A determination module, used for determining at least one abnormal trajectory included in the trajectory to be corrected;

[0063] A correction module, used for correcting the at least one abnormal trajectory to obtain at least one corrected trajectory;

[0064] The determination module is further configured to determine a target trajectory based on the at least one corrected trajectory and a normal trajectory included in the trajectory to be corrected.

[0065] An embodiment of the present application further provides a computer-readable storage medium, wherein the storage medium stores a computer program; when the computer program is executed by a processor of an electronic device, the trajectory correction method as described above can be implemented.

[0066] The trajectory correction method provided in the embodiment of the present application realizes the type classification between the normal trajectory and the abnormal trajectory in the trajectory to be corrected by determining at least one abnormal trajectory included in the trajectory to be corrected; and, by correcting at least one abnormal trajectory to obtain at least one corrected trajectory, not only the targeted correction of at least one abnormal trajectory is realized, but also the negative influence on the position of the normal point in the normal trajectory segment caused by repairing the abnormal point in the abnormal trajectory can be weakened; on this basis, by determining the target trajectory based on at least one corrected trajectory and the normal trajectory included in the trajectory to be corrected, the target trajectory contains the original normal trajectory in the trajectory to be corrected and the corrected trajectory obtained after the abnormal trajectory in the trajectory to be corrected is corrected, so that the target trajectory can fully reflect the correction effect of the trajectory to be corrected; at the same time, through the above operation, the negative influence caused by the adjacent correction of trajectory points through the filtering window in the related art can be effectively overcome, so that the trajectory correction needs in practical applications can be met. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A schematic diagram of a flow chart of a trajectory correction method provided in an embodiment of the present application;

[0068] Figure 2 A schematic diagram of a trajectory of a single drift scenario type in motion provided by an embodiment of the present application;

[0069] Figure 3 A schematic diagram of the trajectory of the abnormal type of in-situ flower explosion provided in an embodiment of the present application;

[0070] Figure 4 A schematic diagram of a trajectory with abnormal fan-shaped trajectory drift provided in an embodiment of the present application;

[0071] Figure 5 A schematic diagram of a repair trajectory for an abnormal trajectory of a fan-shaped trajectory drift type provided in an embodiment of the present application;

[0072] Figure 6 A schematic diagram of the effect of correcting the trajectory segment of the abnormal type of in-situ flower explosion provided by an embodiment of the present application;

[0073] Figure 7 Another schematic diagram of a flow chart of a trajectory correction method provided in an embodiment of the present application;

[0074] Figure 8 A schematic diagram of the structure of a trajectory correction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0075] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0076] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0077] In the field of GPS trajectory correction, the method of filtering the entire GPS trajectory has been widely used. In this method, when the filter window length is large and the proportion of abnormal trajectory points in the GPS trajectory is small, the neighboring points of the current trajectory point in the filter window are weighted averaged, and the weighted average result is used to correct the abnormal trajectory point for smoothing, so as to pull the abnormal trajectory point back to a position close to the normal trajectory, or directly pull it back to the normal trajectory.

[0078] However, in the implementation process of the above method, the larger the filter window length means the greater the amount of calculation for correcting each trajectory point, and when the current trajectory point is highly dependent on its neighborhood trajectory points, the current trajectory point will have an impact on the trajectory points in its neighborhood. When the current trajectory point is an abnormal trajectory point, although the above method can reduce the impact of the current trajectory point, it expands the impact range of the current trajectory point; at the same time, for continuous abnormal trajectory points, the correction effect of the above method is directly related to the number of consecutive abnormal trajectory points, that is, the more the number of consecutive abnormal trajectory points, the worse the correction result.

[0079] Therefore, the method of filtering the entire GPS track provided by the related art cannot meet the actual demand for GPS track correction.

[0080] Based on the above problems, the embodiments of the present application provide a trajectory correction method, device and medium.

[0081] It should be noted that the trajectory correction method provided in the embodiment of the present application can be implemented by a processor of an electronic device. The above-mentioned processor can be at least one of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller (MCU), and a microprocessor.

[0082] Exemplarily, the electronic device may be a physical device or a virtual machine device.

[0083] Exemplarily, the electronic device may be a computer device.

[0084] Figure 1 A schematic diagram of a flow chart of a trajectory correction method provided in an embodiment of the present application, such as Figure 1 As shown, the method may include the following steps:

[0085] Step 101: Obtain the trajectory to be corrected.

[0086] In one embodiment, the trajectory to be corrected may include a real-time or historical GPS trajectory.

[0087] In one implementation, the trajectory to be corrected may correspond to the displacement trajectory of the target object; illustratively, the target object may include a vehicle, equipment, a pedestrian, and the like.

[0088] In one implementation, the trajectory to be corrected may not include map data.

[0089] In one implementation, the trajectory to be corrected may cover at least one time period and may include multiple trajectory points.

[0090] Step 102: Determine at least one abnormal trajectory included in the trajectory to be corrected.

[0091] In one embodiment, the abnormal trajectory may be composed of continuous abnormal trajectory points; wherein the abnormal points may include trajectory points whose distance, speed and direction angle do not satisfy the kinematic law.

[0092] In one implementation, the abnormal trajectory may include normal points; wherein the normal points may include trajectory points that satisfy at least one of distance, speed, and direction angle that satisfies kinematic laws.

[0093] In one embodiment, at least one abnormal trajectory may be determined by any of the following methods:

[0094] The position change state between each trajectory point included in the trajectory to be corrected is analyzed, and at least one abnormal trajectory is determined according to the analysis result. For example, if the position change state indicates that the distance between adjacent trajectory points is greater than a first threshold, the trajectory segment where the adjacent trajectory points are located can be determined to be an abnormal trajectory.

[0095] The angle change state of the direction angle of each trajectory point included in the trajectory to be corrected is counted, and at least one abnormal trajectory is determined based on the angle statistics result; illustratively, if the angle change state indicates that the difference between the direction angles of multiple adjacent trajectory points is greater than a second threshold, then the trajectory segment where the multiple adjacent trajectory points are located can be determined to be an abnormal trajectory.

[0096] Step 103: correct at least one abnormal trajectory to obtain at least one corrected trajectory.

[0097] In one implementation, the corrected trajectory may include a trajectory obtained by correcting the abnormal trajectory.

[0098] In one embodiment, at least one corrected trajectory may be obtained by any of the following methods:

[0099] According to the coordinates of the normal points in the normal track adjacent to the abnormal track, the positions of the track points in the abnormal track are smoothed, and the smoothing result is determined as the correction result.

[0100] According to the movement trend of the normal points in the adjacent normal track before the abnormal track, the movement track of the track points within the time range corresponding to the abnormal track is predicted, and the prediction result is determined as the correction result.

[0101] Step 104: Determine a target trajectory based on at least one corrected trajectory and a normal trajectory included in the trajectory to be corrected.

[0102] In one implementation, the trajectory to be corrected may include at least one normal trajectory.

[0103] In one implementation, the target trajectory may include a trajectory obtained after performing global correction optimization on the trajectory to be corrected.

[0104] In one embodiment, the target trajectory can be determined by:

[0105] Based on the timestamps corresponding to the trajectory points contained in the at least one corrected trajectory and the normal trajectory, the at least one corrected trajectory and the normal trajectory are spliced ​​in a time dimension, and a splicing result is determined as a target trajectory.

[0106] From the above, it can be seen that the trajectory correction method provided by the embodiment of the present application realizes the type division between the normal trajectory and the abnormal trajectory in the trajectory to be corrected by determining at least one abnormal trajectory contained in the trajectory to be corrected; and, by correcting at least one abnormal trajectory to obtain at least one corrected trajectory, not only the targeted correction of at least one abnormal trajectory is realized, but also the negative impact of the position of the normal point in the normal trajectory segment caused by repairing the abnormal point in the abnormal trajectory can be weakened; on this basis, by determining the target trajectory based on at least one corrected trajectory and the normal trajectory contained in the trajectory to be corrected, the target trajectory contains the original normal trajectory in the trajectory to be corrected, and the corrected trajectory obtained after the abnormal trajectory in the trajectory to be corrected is corrected, so that the target trajectory can fully reflect the correction effect of the trajectory to be corrected; at the same time, through the above operation, the negative impact caused by the adjacent correction of trajectory points through the filtering window in the related technology can be effectively overcome, so that the trajectory correction needs in practical applications can be met.

[0107] In the related technology, in the field of trajectory correction, there is also a method of topological analysis based on the characteristics of GPS itself, which is more flexible than the filtering algorithm; and, compared with the filtering algorithm which only uses the GPS location information, the topological analysis method uses more features contained in the GPS data, such as speed and direction, etc.

[0108] In practical applications, general topological analysis methods usually use some numerical processing algorithms to perform statistical analysis on GPS information within a local range to obtain a relatively stable reference point or reference trajectory, and then use the reference trajectory to judge the unknown trajectory to extend the topological structure of the trajectory, and process the judged trajectory and add it to the known trajectory set, and update the reference point or reference trajectory based on the updated processed GPS data set.

[0109] The trajectory correction method provided in the embodiment of the present application is also a specific implementation scheme of the topological analysis method.

[0110] Based on the foregoing embodiments, in the trajectory correction method provided in the embodiments of the present application, if the trajectory to be corrected includes multiple abnormal trajectories, the mth normal trajectory is included between the kth abnormal trajectory and the k+1th abnormal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the quantity threshold; k and m are both integers greater than or equal to 1.

[0111] In one implementation, the kth abnormal trajectory and the k+1th abnormal trajectory include the mth normal trajectory, which can indicate that in the trajectory to be corrected, the abnormal trajectory and the normal trajectory appear alternately adjacent to each other, that is, except for the special case that the trajectory to be corrected only includes the normal trajectory or the abnormal trajectory, the starting point of a normal trajectory in the trajectory to be corrected must be before the end point of an abnormal trajectory, or, if the end point of a normal trajectory is not the end point of the trajectory to be corrected, then the trajectory point after the end point must be the starting point of an abnormal trajectory.

[0112] In one implementation, if the number of trajectory points in the mth normal trajectory is greater than or equal to the quantity threshold, it may indicate that the kth abnormal trajectory and the k+1th abnormal trajectory are independent of each other in the trajectory point dimension or the time dimension associated with the trajectory points.

[0113] Accordingly, determining at least one abnormal trajectory included in the trajectory to be corrected can be achieved in the following manner:

[0114] In the process of correcting the kth abnormal trajectory, the trajectory segments except the kth abnormal trajectory in the trajectory to be corrected are analyzed to determine at least one abnormal trajectory.

[0115] In one embodiment, the trajectory segments in the trajectory to be corrected, excluding the kth abnormal trajectory, may include trajectory points for which it is not determined whether they are abnormal points; exemplarily, the above-mentioned trajectory points may include trajectory points after the kth abnormal trajectory; exemplarily, the above-mentioned trajectory points may include trajectory points after the mth normal trajectory or the k+1th abnormal trajectory.

[0116] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, if the trajectory to be corrected includes multiple abnormal trajectories, the mth normal trajectory is included between the kth abnormal trajectory and the k+1th abnormal trajectory, and the number of trajectory points included in the mth normal trajectory is greater than or equal to the quantity threshold. In this way, the independence between the kth abnormal trajectory and the k+1th abnormal trajectory can be improved by including trajectory points greater than or equal to the quantity threshold in the mth normal trajectory; on this basis, in the process of correcting the kth abnormal trajectory, the trajectory segments other than the kth abnormal trajectory in the trajectory to be corrected are analyzed to determine at least one abnormal trajectory, thereby realizing the parallel processing of the correction processing of the kth abnormal trajectory and the operation of dividing whether other trajectory segments in the trajectory to be corrected are abnormal, thereby improving the efficiency of correcting the trajectory to be corrected.

[0117] Based on the foregoing embodiments, in the trajectory correction method provided in the embodiments of the present application, if the trajectory to be corrected includes multiple abnormal trajectories, the mth normal trajectory is included between the kth abnormal trajectory and the k+1th abnormal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the quantity threshold; k and m are both integers greater than or equal to 1.

[0118] Accordingly, determining at least one abnormal trajectory included in the trajectory to be corrected can be achieved in the following manner:

[0119] In the process of correcting the kth abnormal trajectory, the k+mth abnormal trajectory and / or the kxth abnormal trajectory are corrected.

[0120] Wherein, n is an integer greater than or equal to 1; x is a positive integer less than k.

[0121] In one implementation, the kth correction method associated with the kth abnormal trajectory may be the same as or different from the k+mth correction method associated with the k+mth abnormal trajectory and the kxth correction method associated with the kxth abnormal trajectory.

[0122] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, if the trajectory to be corrected includes multiple abnormal trajectories, the kth abnormal trajectory and the k+1th abnormal trajectory contain the mth normal trajectory, and the number of trajectory points in the mth normal trajectory is greater than or equal to the quantity threshold, thus, the independence between the kth abnormal trajectory and the k+1th abnormal trajectory can be improved by including trajectory points greater than or equal to the quantity threshold in the mth normal trajectory; on this basis, in the process of correcting the kth abnormal trajectory, other abnormal trajectory segments in the trajectory to be corrected are synchronously corrected, thereby realizing parallel processing of correction operations of different abnormal trajectory segments in the trajectory to be corrected, thereby improving the efficiency of correcting the trajectory to be corrected.

[0123] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, determining at least one abnormal trajectory included in the trajectory to be corrected can be achieved in the following manner:

[0124] Determine a trajectory point set included in the trajectory to be corrected; perform statistical division on the trajectory points in the trajectory point set based on point types, and determine at least one abnormal trajectory.

[0125] The trajectory points in the trajectory point set include point types of the trajectory points; and the point type is used to characterize whether the trajectory point is an abnormal point.

[0126] In one implementation, the point type may exist as attribute information of a track point in a track point set.

[0127] In one implementation, the trajectory to be corrected may be divided into individual trajectory points based on timestamp information contained in the trajectory to be corrected, thereby obtaining a trajectory point set; illustratively, the timestamp information may include the GPS time in the trajectory to be corrected.

[0128] In one implementation, the point type of the trajectory point may be determined by comparing the position change states and angular velocity change states of a plurality of adjacent trajectory points in the trajectory to be corrected.

[0129] In one implementation, statistical division of trajectory points in a trajectory point set based on point type may be achieved in the following manner:

[0130] The distribution state of the trajectory points in the trajectory point set in the time dimension is counted based on the point type to obtain the statistical results, and then the trajectory points in the trajectory point set are divided according to the continuous state of the abnormal points and the continuous state of the normal points in the statistical results, so as to obtain at least one abnormal trajectory.

[0131] From the above, it can be seen that the trajectory correction method provided in the embodiment of the present application, after determining the trajectory point set contained in the trajectory to be corrected, can accurately and comprehensively characterize the number, proportion and distribution status of abnormal trajectory points in the trajectory to be corrected through the point type of the trajectory points in the trajectory point set; on this basis, the trajectory points in the trajectory point set are statistically divided based on the point type to determine at least one abnormal trajectory, which can improve the division accuracy of the abnormal trajectory.

[0132] Based on the foregoing embodiment, the trajectory correction method provided in the embodiment of the present application may further include the following steps:

[0133] Step A1: Obtain the adjacent trajectory points of the pth trajectory point in the trajectory to be corrected.

[0134] The adjacent trajectory points include P trajectory points before the p-th trajectory point and P trajectory points after the p-th trajectory point; p and P are both integers greater than 1.

[0135] In one embodiment, before obtaining adjacent trajectory points, the credibility of the starting point and the end point of the trajectory to be corrected can be judged first; illustratively, the credibility can be judged by evaluating the positional relationship between the starting point and the end point of the trajectory to be corrected. For example, if the positional relationship between the starting point and the end point conforms to the law of motion, it can be determined that the starting point and the end point are basically credible, and the starting point and the end point can be determined as the valid starting point and the valid end point; otherwise, if the positional relationship between the two does not conform to the law of motion, the valid starting point and the valid end point can be extrapolated from the starting point and the end point to the trajectory to be corrected.

[0136] In one implementation, the pth trajectory point may be a trajectory point between a valid starting point and a valid ending point.

[0137] In one embodiment, the value of P can be determined or adjusted according to actual correction requirements; in the actual trajectory correction process, the larger the value of P, the better the corresponding correction effect, but the resource consumption and time consumption cost corresponding to the correction process will also increase; illustratively, in the trajectory correction process, it can be set according to the hardware computing speed of the electronic device, for example, the value of P can be set to 5.

[0138] Step A2: determining motion feature information of the pth trajectory point and adjacent trajectory points.

[0139] The motion feature information includes at least one of instantaneous speed, average speed, direction angle and static point ratio.

[0140] In one implementation, the trajectory points in the trajectory to be corrected may carry information such as instantaneous speed, average speed, and direction angle.

[0141] In one implementation, the static point ratio may be represented by a percentage between the number of static trajectory points in the trajectory to be corrected and the number of all trajectory points in the trajectory to be corrected; illustratively, the static trajectory point may be one of the attribute parameters of the trajectory in the trajectory to be corrected.

[0142] Step A3: Determine the point type of the pth trajectory point based on the motion feature information.

[0143] In one implementation, the point type of the pth trajectory point can be determined in the following manner:

[0144] According to the proportion of the stationary points, at least one of the instantaneous speed, the average speed and the direction angle in the motion feature information is analyzed, and the point type of the p-th trajectory point is determined according to the analysis result.

[0145] For example, if the proportion of stationary points is greater than the proportion threshold, it can be determined that the trajectory to be corrected is a stationary or low-speed motion trajectory. At this time, if the angular direction of the p-th trajectory point and the adjacent trajectory points changes greatly, and the phenomenon of large angular direction changes occurs more than once, it can be determined that the p-th trajectory point is an abnormal point that drifts in a stationary state; illustratively, if the angular direction of the p-th trajectory point and the adjacent trajectory points changes greatly, but this phenomenon occurs once, and the trajectory point before the p-th trajectory point is determined to be an abnormal point, then the p-th trajectory point can also be determined as an abnormal point, but if the trajectory point before the p-th trajectory point is a normal point, then the p-th trajectory point can be determined as a normal point, and whether the p-th trajectory point is a normal point can be determined based on the state of the trajectory points appearing after the p-th trajectory point.

[0146] For another example, if the stationary point accounts for less than the proportion threshold, but the instantaneous speed is greater than the speed threshold of the target object, the pth trajectory point can be determined as an abnormal point; or, if there is a sudden change in the average speed and the average speed corresponding to the adjacent trajectory points is greater than the speed threshold, the pth trajectory point can be determined as an abnormal point; wherein the speed threshold can be determined according to the type of the target object, for example, the speed threshold of a vehicle can be greater than the speed threshold of a pedestrian.

[0147] From the above, it can be seen that the trajectory correction method provided by the embodiment of the present application obtains the motion feature information of the p-th trajectory point and its adjacent trajectory points in the trajectory to be corrected, and determines the point type of the p-th trajectory point based on the motion feature information. Thus, through the above operation, in the process of determining the point type of the p-th trajectory point, the motion feature information of the p-th trajectory point and its adjacent trajectory points is tracked and sorted along the time extension direction of the trajectory to be corrected, so that the motion state of each trajectory point in the trajectory to be corrected can be accurately displayed; and, by determining the point type of the p-th trajectory point based on the instantaneous speed, average speed, direction angle and proportion of stationary points contained in the motion feature information, it is possible to achieve an accurate evaluation of the motion features of the p-th trajectory point, thereby improving the accuracy of the point type of the p-th trajectory point.

[0148] Based on the foregoing embodiment, in the trajectory correction method provided in the embodiment of the present application, determining the point type of the pth trajectory point based on the motion feature information can be implemented in the following manner:

[0149] If the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory point is greater than the gradient threshold, the p-th trajectory point is determined to be an abnormal point.

[0150] Exemplarily, if the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory point is less than or equal to the gradient threshold, the operation of determining that the p-th trajectory point is an abnormal point may not be performed.

[0151] In one embodiment, if the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory points is greater than the gradient threshold, it can be indicated that there is a mutation in the motion feature information of the p-th trajectory point compared with the statistical average of the motion feature information of the adjacent trajectory points, or there is a mutation in the motion feature information of the p-th trajectory point compared with the motion feature information of a few directly adjacent trajectory points among the adjacent trajectory points.

[0152] In one implementation, a gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory point is greater than a gradient threshold, which may indicate that the instantaneous speed of the p-th trajectory point is greater than the average speed or instantaneous speed of the adjacent trajectory points, or may indicate that the difference between the direction angle of the p-th trajectory point and the direction angle of the adjacent trajectory point is greater than an angle threshold.

[0153] In the process of implementing the trajectory correction method in the related art, the point type of the current trajectory point is determined by relying on the historical trajectory points or historical trajectories of the current trajectory point. Obviously, in the above scheme, the historical trajectory points or historical trajectories only contain the information of the trajectory points before the current trajectory point, so they are not comprehensive enough, resulting in insufficient accuracy of the point type.

[0154] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, if the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory point is greater than the angle threshold, the p-th trajectory point is determined to be an abnormal point. In this way, in the above process, not only the motion state information of the historical trajectory points adjacent to the p-th trajectory point is referenced, but also the motion state information of the future trajectory points adjacent to the p-th trajectory point is referenced, so that the point type of the p-th trajectory point can be associated with the motion feature information of the adjacent trajectory points associated with it, thereby improving the accuracy of the point type of the p-th trajectory point.

[0155] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, the trajectory points in the trajectory point set are statistically divided based on point types to determine at least one abnormal trajectory, which can be achieved in the following manner:

[0156] If the N consecutive trajectory points adjacent to the p-th trajectory point are all normal points, the p-1-th trajectory point is an abnormal point, and the p+N+1-th trajectory point is an abnormal point, the trajectory formed by the trajectory points between the p-th trajectory point and the p+N-th trajectory point is classified as a normal trajectory; and the trajectory segment excluding the normal trajectory in the trajectory to be corrected is determined as an abnormal trajectory.

[0157] Wherein, p is an integer greater than 1; N is greater than or equal to the quantity threshold.

[0158] Exemplarily, if the N trajectory points adjacent to the p-th trajectory point are not normal points, or the p-1-th trajectory point is an abnormal point, or the p+N+1-th trajectory point is not an abnormal point, then the trajectory interval formed by the trajectory points between the p-th trajectory point and the p+N-th trajectory point may not be classified as a normal trajectory.

[0159] In one embodiment, a normal trajectory can be determined based on a counting result obtained by performing a counting operation through a normal point counter. For example, the trajectory points in the trajectory to be corrected can be traversed according to the time dimension. If the first abnormal point is detected, the abnormal point can be determined as the starting point of the abnormal trajectory; if the first normal function point is detected, the counting result of the normal point counter is set to 1; and when the counting result of the normal point counter is less than N, if an abnormal point is detected, the normal point counter is cleared; illustratively, if normal points appear continuously, the counting result of the normal point counter is continuously incremented. If the counting result of the normal point counter is N, the trajectory point before the trajectory point where the normal point counter starts counting is taken as the end point of the abnormal trajectory, and the trajectory between the trajectory point where the normal point counter starts counting and the corresponding trajectory point when the counting result is N is determined as the normal trajectory. At this time, the normal point counter can be cleared, and the section of the trajectory from the first abnormal point to the end point of the abnormal trajectory is determined as the abnormal trajectory.

[0160] Exemplarily, the value of N can be determined based on the stability of the timestamps of the trajectory points. For example, if the time intervals between the timestamps of the trajectory points in the trajectory to be corrected are stable, the value of N can be determined by using the number of trajectory points. For another example, if the time intervals between the timestamps of the trajectory points in the trajectory to be corrected are unstable, the distance of the normal trajectory can be accumulated by adding the distances between the trajectory points while the normal point counter accumulates the normal points, and the value of N can be determined based on the above distance.

[0161] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, if the N consecutive trajectory points adjacent to the p-th trajectory point are all normal points, and the p-1-th trajectory point is an abnormal point, and the p+1+N-th trajectory point is an abnormal point, then the trajectory interval composed of the trajectory points between the p-th trajectory point and the p+N-th trajectory point is divided into a normal trajectory. In this way, the accurate division of the normal trajectory in the trajectory to be corrected is achieved by setting the conditions of the continuity of the N trajectory points and that they are all normal points; and, determining that the trajectory segments other than the normal trajectory in the trajectory to be corrected are abnormal trajectories can reduce the probability that the trajectory segments cannot be classified due to the alternating appearance of normal points and abnormal points in the trajectory to be corrected, thereby improving the accuracy of the abnormal trajectory.

[0162] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, correcting at least one abnormal trajectory can be achieved by the following steps:

[0163] Step B1, determining the abnormal type of the kth abnormal trajectory segment.

[0164] Here, k is an integer greater than or equal to 1.

[0165] In one implementation, the abnormality type may include the type of distribution state of trajectory points in the kth abnormal trajectory.

[0166] In one implementation, the exception type may be determined by:

[0167] The type of the distribution state of the abnormal points contained in the k-th abnormal trajectory is determined as the abnormal type of the k-th abnormal trajectory.

[0168] Step B2: Determine the kth correction strategy based on the abnormal type of the kth abnormal trajectory.

[0169] In one embodiment, the kth correction strategy may include methods and conditions for correcting at least one of the position, angular direction, and speed of at least part of the abnormal points in the kth abnormal trajectory.

[0170] In one implementation, the correction strategies corresponding to abnormal trajectories of different abnormal types may be different.

[0171] In one implementation, the kth revision strategy may be determined by:

[0172] An association relationship may be established between the abnormality type and the correction strategy. After determining the abnormality type of the kth abnormal trajectory, the kth correction strategy may be determined from the above association relationship based on the degree of matching between the abnormality type of the kth abnormal trajectory and the abnormality type in the above association relationship.

[0173] Step B3: correct the kth abnormal trajectory based on the kth correction strategy.

[0174] From the above, it can be seen that the trajectory correction method provided in the embodiment of the present application can classify the types of multiple different abnormal trajectories by determining the abnormal type of the kth abnormal trajectory; on this basis, determining the kth correction strategy based on the abnormal type of the kth abnormal trajectory can not only reduce the number of kth correction strategies, but also improve the pertinence of the kth correction strategy; in this way, the amount of calculation for correcting the kth abnormal trajectory based on the kth correction strategy can be reduced, thereby improving the efficiency of correcting the trajectory to be corrected.

[0175] Based on the above embodiments, the trajectory correction method provided in the embodiments of the present application can determine the abnormal type of the kth abnormal trajectory by the following steps:

[0176] The decision tree model is used to perform correlation analysis on the motion feature information of the trajectory points in the kth abnormal trajectory to determine the abnormal type of the kth abnormal trajectory.

[0177] In one embodiment, abnormal trajectories composed of various abnormal point sets and normal trajectories composed of various normal point sets can be obtained as decision sample data, and an initial decision tree model is trained based on the above decision sample data to obtain the above decision tree model; exemplarily, in the above training process, the decision tree model can perform statistical analysis on the features of the abnormal trajectories and the normal trajectories in each feature dimension, so as to construct a decision tree model in each feature dimension or in multiple feature joint dimensions, thereby achieving the distinction between various types of trajectory point sets.

[0178] In one embodiment, the decision tree model can extract motion features of the kth abnormal trajectory in each feature dimension or in multiple feature joint dimensions, and determine the abnormal type of the kth abnormal trajectory based on the feature extraction result.

[0179] As can be seen from the above, in the trajectory correction method provided by the embodiment of the present application, the motion feature information of the trajectory points in the kth abnormal trajectory is correlated and analyzed through the decision tree model to determine the abnormal type of the kth abnormal trajectory. In this way, with the advantage of the decision tree model that can be used repeatedly after one training and has high classification efficiency, the classification efficiency of the abnormal trajectory in the trajectory to be corrected can be improved.

[0180] Based on the foregoing embodiment, the trajectory correction method provided in the embodiment of the present application determines the abnormal type of the kth abnormal trajectory, and may further include the following steps:

[0181] Step C1: Obtain sample data.

[0182] The sample data includes motion feature information of trajectory points in a normal trajectory and motion feature information of trajectory points in an abnormal trajectory.

[0183] In one embodiment, the sample data may include training data and verification data; wherein, the training data and the verification data.

[0184] Exemplarily, the motion feature information of the i-th track point in the sample data can be expressed as (X j ,δ) i Indicates that, X j It indicates that it belongs to the jth category and is used to characterize whether the i-th trajectory point is a normal point or an abnormal point. δ indicates the angle change value of the i-th trajectory point; wherein i is an integer greater than or equal to 1.

[0185] Step C2: training the cascade support vector machine in the initial state based on the sample data to obtain a cascade support vector machine.

[0186] In one embodiment, the trajectory point sequences corresponding to various types of trajectories in the sample data can be input into the cascade support vector machine in the initial state. During the process of the cascade support vector machine in the initial state processing the above trajectory point sequences, each cascade support vector machine can respectively capture and track the motion feature information embodied by the abnormal point sets in different types of abnormal tracks, thereby having the ability to recognize and distinguish abnormal trajectories of corresponding categories.

[0187] It should be noted that steps C1-C2 may be pre-executed before determining the abnormal type of the kth abnormal trajectory.

[0188] Step C3: performing association analysis on the motion state information of the trajectory points in the kth abnormal trajectory by using a cascade support vector machine to determine the abnormal type of the kth abnormal trajectory.

[0189] In one embodiment, multiple abnormal trajectories can be input into a cascade support vector machine. After being processed by each level of the cascade support vector machine, the abnormal types of the multiple abnormal trajectories can be obtained respectively; for example, if the k-th level vector machine of the cascade support vector machine determines that the k-th abnormal trajectory is of the first type, other abnormal trajectories can be input into other level vector machines to obtain the abnormal types of other abnormal trajectories.

[0190] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, after obtaining sample information including motion feature information of trajectory points in a normal trajectory and motion feature information of trajectory points in an abnormal trajectory, in the process of training the cascade support vector machine in the initial state based on the sample data, the cascade support vector machine's ability to recognize and extract motion feature information of different types of trajectory segments can be improved; on this basis, by performing association analysis on the motion feature information of the trajectory points in the kth abnormal trajectory segment through the cascade support vector machine, the efficiency of determining the abnormal type of the kth abnormal trajectory segment can be improved; and, with the aid of the cascade structure of the cascade support vector machine, the efficiency of determining the abnormal type can be improved.

[0191] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, determining the abnormal type of the kth abnormal trajectory can also be achieved by the following steps:

[0192] Step D1, obtaining a target normal trajectory adjacent to the kth abnormal trajectory.

[0193] The target normal trajectory includes the normal trajectory adjacent to the kth abnormal trajectory.

[0194] In one implementation, the number of trajectory segments in the target normal trajectory may be at least one.

[0195] Step D2: determine the first position and the second position.

[0196] The first position includes the positions of the starting trajectory point and the ending trajectory point of the kth abnormal trajectory; the second position includes the position of the normal point in the target normal trajectory that is adjacent to the starting trajectory point and the abnormal trajectory point of the kth abnormal trajectory.

[0197] In one implementation, the first location and the second location may be embodied in the form of longitude and latitude.

[0198] Step D3: If the distance between the first position and the second position is greater than a first distance threshold, determine that the abnormality type is the first type.

[0199] Exemplarily, if the distance between the first position and the second position is less than or equal to a first distance threshold, the operation of determining that the abnormality type is the first type may not be performed.

[0200] In one embodiment, the distance between the first position and the second position is greater than the first distance threshold, which may indicate that the distance between the starting trajectory point of the kth abnormal trajectory and the ending trajectory point of the historical normal trajectory of the kth abnormal trajectory in the target normal trajectory is greater than the first distance threshold, and may also indicate that the distance between the ending trajectory point of the kth abnormal trajectory and the starting trajectory point of the future normal trajectory of the kth abnormal trajectory in the target normal trajectory is greater than the first distance threshold.

[0201] In one embodiment, the first type may be a single drift type in motion; illustratively, the single drift type in motion may include a scene containing at least two position mutation points; wherein the first scene may include a trajectory point suddenly jumping from a normal trajectory to an abnormal trajectory, and the other scene may include a trajectory point suddenly jumping from an abnormal trajectory to a normal trajectory.

[0202] Exemplarily, in the above two scenarios, the distance between the trajectory points where the position mutation occurs may be greater than the first distance threshold; exemplarily, the above first distance threshold may be a distance that the target object cannot reach within the time period of the difference between the timestamps corresponding to adjacent trajectory points.

[0203] In one embodiment, in the single drift type during motion, the number of trajectory points associated with the position mutation can be multiple. For example, the GPS data continues to drift within a certain period of time, resulting in a position mutation of the trajectory point along one direction within a first period of time. At a certain moment, after the GPS signal quality improves, the trajectory point suddenly mutates to a normal trajectory at one time.

[0204] For example, in a single drifting scenario in motion, all trajectory points that deviate from a normal trajectory may be abnormal points.

[0205] Figure 2This is a schematic diagram of the trajectory of a single drift scenario type in motion provided by an embodiment of the present application. Figure 2 As shown, the target normal trajectory adjacent to the first abnormal trajectory 201 may include the first normal trajectory 202 and the second normal trajectory 203; wherein, the distance between the end point of the first normal trajectory 202 and the start and end points of the first abnormal trajectory 201 is greater than the first distance threshold, and the distance between the start point of the second normal trajectory 203 and the end point of the first abnormal trajectory 201 is also greater than the first distance threshold, so that two position mutations occur between the first normal trajectory 202 and the first abnormal trajectory 201 and the second normal trajectory 203.

[0206] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, after obtaining the normal trajectory adjacent to the k-th abnormal trajectory, that is, the target normal trajectory, the position of the starting trajectory point and the ending trajectory point of the k-th abnormal trajectory, that is, the first position, and the position of the normal trajectory point adjacent to the starting trajectory point and the abnormal trajectory point of the k-th abnormal trajectory in the target normal trajectory, that is, the second position, are determined. In this way, through the first position and the second position, the position change state between the k-th abnormal trajectory and the normal trajectory adjacent to it can be clearly characterized; on this basis, if the distance between the first position and the second position is greater than the first distance threshold, the abnormal type is determined to be the first type, which not only realizes the accurate judgment of the first type, but also can finely reflect the position mutation characteristics of the first type.

[0207] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, determining the abnormal type of the kth abnormal trajectory can also be achieved in the following manner:

[0208] Clustering is performed on the positions of the trajectory points in the kth abnormal trajectory to obtain a cluster center; if the distance between the trajectory points in the kth abnormal trajectory and the cluster center is less than a second distance threshold, the abnormal type is determined to be the second type.

[0209] Exemplarily, if the distance between the trajectory point in the kth abnormal trajectory and the cluster center is greater than or equal to the second distance threshold, the operation of determining that the abnormality type is the second type may not be performed.

[0210] In one embodiment, the second type may include an in-situ explosion anomaly type; illustratively, the in-situ explosion anomaly type may include that the target object stays in the same place, but the position represented by the GPS data drifts slightly within a range near the original place, and the long-term small drift will result in a very messy abnormal trajectory.

[0211] In practical applications, if the GPS signal quality at the target object's location is poor, it may cause abnormal trajectories of the in-situ explosion type to appear.

[0212] Figure 3 The trajectory diagram of the abnormal type of in-situ flower explosion provided in the embodiment of the present application is as follows: Figure 3 As shown, the third position 301 of the target object has not changed, but due to the poor quality of the GPS signal, the position represented by the GPS data repeatedly jumps within the grid-filled area centered on the third position 301, thereby forming an abnormal trajectory type of exploding in situ.

[0213] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, the positions of the trajectory points in the kth abnormal trajectory are clustered, and after the cluster center is obtained, if the distance between the trajectory points in the kth abnormal trajectory and the cluster center is less than the second distance threshold, the abnormal type is determined to be the second type. In this way, through the above clustering process, the cluster center can comprehensively and accurately characterize the central position of the trajectory points in the kth abnormal trajectory; and based on the relationship between the distance between the trajectory points in the kth abnormal trajectory and the cluster center and the second distance threshold, the type of the kth abnormal trajectory is determined, which can not only reflect the abnormal characteristics of the second type, but also improve the accuracy of the abnormal type of the kth abnormal trajectory.

[0214] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, determining the abnormal type of the kth abnormal trajectory can also be achieved by the following steps:

[0215] Step E1, obtaining the normal trajectory of the target.

[0216] The target normal trajectory includes the normal trajectory adjacent to the kth abnormal trajectory.

[0217] Step E2: Determine the track point switching frequency.

[0218] The trajectory point switching frequency is the switching frequency between the trajectory points in the kth abnormal trajectory and the trajectory points in the target normal trajectory.

[0219] In one embodiment, the trajectory point switching frequency may include the frequency at which a trajectory point in the kth abnormal trajectory jumps to any trajectory point in the future trajectory in the target normal trajectory.

[0220] In one implementation, the track switching frequency may include a frequency at which a track point in a historical abnormal track of the kth abnormal track is switched to an arbitrary track point in the kth abnormal track.

[0221] In one implementation, the trajectory point switching frequency may also include the frequency of switching between abnormal points and normal points in the kth abnormal trajectory.

[0222] Step E3: If the switching frequency of the track point is greater than the switching threshold, determine that the abnormal feature is the second type.

[0223] Exemplarily, if the trajectory point switching frequency is less than or equal to the switching threshold, it can be determined that the abnormal feature is not of the second type.

[0224] In one implementation, if the switching frequency of the trajectory point is greater than the switching threshold, it may indicate that frequent switching occurs between normal points and abnormal points within a unit time.

[0225] In one embodiment, the second type may include sector trajectory drift anomalies; exemplarily, sector trajectory drift anomalies may include repeated occurrence of trajectory point drift within a period of time, and the trajectory points frequently jump between the drift trajectory and the target normal trajectory; exemplarily, the drift trajectory may be the kth abnormal trajectory, and may also include other segments of normal trajectories.

[0226] Figure 4 This is a schematic diagram of a trajectory with abnormal fan-shaped trajectory drift provided by an embodiment of the present application. Figure 4 As shown, the target normal trajectory 401 is adjacent to the fan-shaped area formed by the k-th abnormal trajectory 402 , and the trajectory points repeatedly jump between the target normal trajectory 401 and the k-th abnormal trajectory 402 .

[0227] From the above, it can be seen that in the trajectory correction method provided by the embodiment of the present application, after obtaining the target normal trajectory adjacent to the k-th abnormal trajectory, the switching frequency between the trajectory points in the k-th abnormal trajectory and the trajectory points in the target normal trajectory, that is, the trajectory point switching frequency, is determined. In this way, through the trajectory point switching frequency, the switching state of the trajectory points between the k-th abnormal trajectory and the target normal trajectory can be comprehensively and meticulously reflected; and if the trajectory point switching frequency is greater than the switching threshold, the abnormal type is determined to be the third type. In this way, not only the accurate judgment of the third type is achieved, but also the switching characteristics of the trajectory points embodied by the third type can be reflected.

[0228] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, the kth correction strategy is determined based on the abnormal type of the kth abnormal trajectory, which can be implemented in the following manner:

[0229] If the kth abnormal trajectory is of the first type, the kth correction strategy is determined as:

[0230] Determine the distance similarity between the first distance and the second distance; determine the angle similarity between the first angle and the second angle; determine fitting parameters based on the distance similarity and the angle similarity, and fit and correct the kth abnormal trajectory based on the fitting parameters.

[0231] Among them, the first distance includes the distance between at least two normal points adjacent to the kth abnormal trajectory; the second distance includes the distance between normal points in the kth abnormal trajectory; the first angle includes the angle associated with at least two normal points adjacent to the kth abnormal trajectory; the second angle includes the angle associated with the normal point in the kth abnormal trajectory.

[0232] In one implementation, the first distance may include a distance between an end point of a historical normal trajectory of the kth abnormal trajectory and a start point of a future normal trajectory of the kth abnormal trajectory.

[0233] In one implementation, the first angle may include an angle between a line connecting an end point of a historical normal trajectory of the kth abnormal trajectory and a start point of a future normal trajectory of the kth abnormal trajectory.

[0234] In one implementation, a set of normal points in the kth abnormal trajectory may be extracted, and a normal trajectory segment formed by the set of normal points may be determined. Then, the distance between the starting trajectory point and the ending trajectory point of the normal trajectory segment may be determined as the second distance, and the angle of the line between the starting trajectory point and the ending trajectory point of the normal trajectory segment may be determined as the second angle. Exemplarily, the second distance and the second angle may be calculated by equations (1) to (2), respectively:

[0235] d=dis(p start ,p end ) (1)

[0236] a=angle(p start ,p end ) (2)

[0237] Among them, d and a are the second distance and the second angle respectively, p start and p end They are respectively the positions of the starting trajectory point and the ending trajectory point of the above normal trajectory segment.

[0238] In one implementation, the similarity between the first angle and the second angle may be calculated using a cosine similarity method.

[0239] In a real-time manner, the distance similarity may include at least one of the trajectories, directions, and times associated with the first distance and the second distance, respectively, and the degree of matching in the dimension of motion physics; illustratively, the distance between at least two normal points included in the first distance may be predicted in the time dimension based on the motion speed and / or motion direction to obtain a prediction result; if the prediction result is consistent with the second distance, the distance similarity may be the first similarity; if the prediction result is different from the second distance, the distance similarity may be the second similarity, and the first similarity may be greater than the second similarity.

[0240] In one embodiment, the fitting parameters may include a weighted result obtained by weighting the distance similarity and the angle similarity; exemplarily, in the above weighting process, the sum of the distance weight associated with the distance similarity and the angle weight associated with the angle similarity may be 1, and the distance weight may be smaller than the direction weight to strengthen the direction factor in the trajectory correction process, for example, the distance weight may be 0.35 and the direction weight may be 0.65.

[0241] In one implementation, fitting and correcting the kth abnormal trajectory can be achieved by:

[0242] The fitting parameters are determined as fitting weights, and the normal points in the kth abnormal trajectory and at least two normal points adjacent to the kth abnormal trajectory are fitted based on the fitting weights to obtain a fitting curve, and then the kth abnormal trajectory is replaced by the fitting curve to fit and correct the kth abnormal trajectory.

[0243] As can be seen from the above, in the trajectory correction method provided by the embodiment of the present application, if the k-th abnormal trajectory is of the first type, the distance similarity between the first distance and the second distance is determined. Since the first distance includes the distance between at least two normal points adjacent to the k-th abnormal trajectory, and the second distance includes the distance between the normal points in the k-th abnormal trajectory, the distance change state between the normal points around and contained in the k-th abnormal trajectory can be fully reflected through the distance similarity; and since the first angle is the angle associated with at least two normal points adjacent to the k-th abnormal trajectory, and the second angle includes the angle associated with the normal points in the k-th abnormal trajectory, by determining the angle similarity between the first angle and the second angle, the angle change state before and after the k-th abnormal trajectory and between each point in the k-th abnormal trajectory can be reflected; on this basis, by determining the fitting parameters based on the distance similarity and the angle similarity, the fitting parameters can reflect the trajectory extension state of the k-th abnormal trajectory from the distance dimension and the trajectory extension direction dimension; on the other hand, by fitting and correcting the k-th abnormal trajectory based on the fitting parameters, the k-th abnormal trajectory can be corrected in a targeted manner.

[0244] Based on the above embodiments, in the trajectory correction method provided in the embodiment of the present application, based on the fitting parameters, the fitting correction of the kth abnormal trajectory can be achieved in the following manner:

[0245] If the fitting parameter is greater than the fitting threshold, the kth abnormal trajectory is fitted and corrected based on the position of the normal point in the kth abnormal trajectory; if the fitting parameter is less than or equal to the fitting threshold, the kth abnormal trajectory is fitted and corrected based on at least two normal points adjacent to the kth abnormal trajectory.

[0246] In one implementation, if the fitting parameter is greater than the fitting threshold, it may indicate that the trajectory change state represented by the normal points included in the k-th abnormal trajectory is consistent with the trajectory change state represented by the target normal trajectory adjacent to the k-th abnormal trajectory. In this case, the k-th abnormal trajectory may be replaced by the normal trajectory segment composed of the normal points in the k-th abnormal trajectory, thereby completing the fitting correction of the k-th abnormal trajectory.

[0247] In one embodiment, if the fitting parameter is less than or equal to the fitting threshold, it can be indicated that the difference between the trajectory change state represented by the normal points included in the k-th abnormal trajectory and the trajectory change state represented by the target normal trajectory adjacent to the k-th abnormal trajectory is large. In this case, the position of the normal points included in the target normal trajectory of the k-th abnormal trajectory can be used to fit, correct and replace the k-th abnormal trajectory.

[0248] In one implementation, the fitting threshold may be a number greater than or equal to 0 and less than or equal to 1. In practical applications, the fitting threshold may be 0.7.

[0249] like Figure 2 As shown, the first corrected trajectory 204 may be a corrected trajectory obtained after fitting and correcting the first abnormal trajectory 201. In the first corrected trajectory 204, each trajectory point is smoothly associated in the time dimension, the distance dimension, and the motion direction dimension.

[0250] From the above, it can be seen that the trajectory correction method provided in the embodiment of the present application fits the kth abnormal trajectory in different ways according to the size relationship between the fitting parameters and the fitting threshold. In this way, it can meet the diverse fitting needs, thereby improving the fitting accuracy and improving the fitting effect.

[0251] Based on the above embodiments, in the path correction method provided in the embodiments of the present application, the kth correction strategy is determined based on the abnormal type of the kth abnormal trajectory, which can also be implemented in the following manner:

[0252] If the kth abnormal trajectory is of the second type, the kth correction strategy is determined as:

[0253] The positions of the trajectory points in the k-th abnormal trajectory are clustered to obtain the cluster center, and the k-th abnormal trajectory is corrected based on the cluster center.

[0254] In one embodiment, the correction requirement of the second type, i.e., the in-situ flower explosion abnormal type, can be determined in advance, and based on the correction requirement, it is determined whether to correct the kth abnormal trajectory based on the cluster center; exemplarily, if the correction requirement includes determining the position of the first trajectory point of the kth abnormal trajectory as the center point of the kth abnormal trajectory, then the above-mentioned first trajectory point can be determined as the center point of the kth abnormal trajectory; if the correction requirement does not include determining the position of the first trajectory point of the kth abnormal trajectory as the center point of the kth abnormal trajectory, then the kth abnormal trajectory can be corrected based on the cluster center.

[0255] In one implementation, the kth abnormal trajectory is corrected based on the cluster center, and the cluster center may be set as the trajectory point of the kth abnormal trajectory.

[0256] Figure 6 The schematic diagram of the effect of correcting the trajectory segment of the abnormal type of in-situ flower explosion provided by the embodiment of the present application is as follows: Figure 6 As shown, the first trajectory description 601 corresponding to the kth abnormal trajectory shows an abnormal trajectory of 9.39 km in the period from 2021-5-31 02:06:41 to 2021-5-31 15:59:44, which is consistent with Figure 6 The repeated jumps in the grid-filled area centered on the third position 301 shown in the figure correspond to; after being processed by the trajectory correction method provided in the embodiment of the present application, the second trajectory description 602 corresponding to the corrected trajectory shows that the length of the path trajectory during the above period is 0.00 km, and the determined corrected trajectory can be the cluster center matched by the third position 301.

[0257] As can be seen from the above, in the trajectory correction method provided by the embodiment of the present application, if the kth abnormal trajectory is of the second type, the kth correction strategy is determined to cluster the positions of the trajectory points in the kth abnormal trajectory to obtain the cluster center, and correct the kth abnormal trajectory based on the cluster center. In this way, the distribution state of the trajectory points in the kth abnormal trajectory can be accurately characterized by the cluster center, and the position of the cluster center can represent the actual trajectory points in the in-situ flower explosion abnormal type, thereby meeting the actual abnormal trajectory correction needs.

[0258] Based on the above embodiments, in the trajectory correction method provided in the embodiments of the present application, the kth correction strategy is determined based on the abnormal type of the kth abnormal trajectory, which can also be implemented in the following manner:

[0259] If the kth abnormal trajectory is of the third type, the kth correction strategy is determined as:

[0260] The kth abnormal trajectory is corrected based on at least the distribution information of the abnormal points and the normal points in the kth abnormal trajectory.

[0261] The distribution information includes time information and location information.

[0262] In one embodiment, if the kth abnormal trajectory is of the third type, i.e., the fan-shaped trajectory drift type, and the drift point in the kth abnormal trajectory is an abnormal point, the time information of the abnormal point can be obtained, while the normal point in the kth abnormal trajectory is retained, and the position information and time information of the above normal point are obtained, and then the position information and time information of the normal point adjacent to the abnormal point in the kth abnormal trajectory are determined according to the time information of the abnormal point, and each position information is weighted based on the above time information to correct the position of the abnormal point.

[0263] For example, the position information and time information of the two normal points with the smallest distance from the abnormal point in the kth abnormal trajectory can be determined according to the time information of the abnormal point in the kth abnormal trajectory, and the position information of the abnormal point can be corrected, as shown in formula (3):

[0264]

[0265] Among them, Location correct It can represent the corrected location information of the above outliers. back and Location front is the location information of the two normal points adjacent to the above abnormal point, timeDiff back and timeDiff front It can represent the difference in time information between two normal points and abnormal points respectively.

[0266] In one embodiment, if the drift point in the kth abnormal trajectory repeatedly jumps between two normal trajectories, the trajectory points in the kth abnormal trajectory can be first divided according to the point type to obtain a first point set and a second point set; wherein the first point set can be a set consisting of abnormal points in the kth abnormal trajectory, and the second point set can be a set consisting of normal points in the kth abnormal trajectory.

[0267] For example, starting from the end point P of the normal trajectory segment, the nearest neighboring points p of the point can be extracted from the first point set. nearest and the next nearest neighbor p subnearest , the nearest neighbor point P of the point in the second point set nearest , and determine the correction method based on the positions of the above four points.

[0268] For example, if only p nearest Located between P and P nearest Then we can put p nearest and P nearestAfter the correction time is added to the corrected trajectory, the two trajectory points are deleted from the first point set and the second point set respectively, and P is used nearest Update P and repeat the above calculation steps.

[0269] For example, if p nearest and p subnearest Both located in P and P nearest Between, take at least one normal point before P, and then compare it with P nearest Fit a quadratic curve and calculate p nearest and p subnearest The degree of deviation from the above quadratic curve, if the degree of deviation is less than the deviation threshold, then add the point to the correction trajectory after the correction time, delete the point from the first point set, and use the last point in the correction trajectory to update P, and repeat the previous calculation step.

[0270] For example, if p nearest and p subnearest All located in P nearest After that, we can directly correct P nearest After the time P nearest Add to the corrected trajectory and delete P from the second point set nearest , and use P nearest Update P and repeat the previous calculation step.

[0271] Exemplarily, the above calculation steps may be iterated repeatedly until all points in the first point set and the second point set are added to the corrected trajectory and the two sets are empty.

[0272] Figure 5 A schematic diagram of a repair trajectory for an abnormal trajectory of a fan-shaped trajectory drift type provided in an embodiment of the present application. Figure 5 As shown, the second corrected trajectory 501 is distributed at the edge of the k-th abnormal trajectory 402 , and is consistent with the trajectory extension direction of the target normal trajectory 401 .

[0273] As can be seen from the above, in the trajectory correction method provided by the embodiment of the present application, if the kth abnormal trajectory is of the third type, the kth abnormal trajectory is corrected based on at least the time information and position information of the abnormal points and normal points in the kth abnormal trajectory. In this way, through the above operation, in the process of correcting the kth abnormal trajectory, not only the distribution information of the normal points in the kth abnormal trajectory is referred to, but also the distribution information of the abnormal points in the kth abnormal trajectory is fully considered, so that the pertinence of the correction of the kth abnormal trajectory can be improved.

[0274] Figure 7 Another schematic diagram of the process flow of the trajectory correction method provided in the embodiment of the present application is as follows: Figure 7 As shown, the process may include the following steps:

[0275] Step 701: Obtain the trajectory to be corrected.

[0276] Exemplarily, the trajectory to be corrected may be a historical trajectory.

[0277] Step 702: classify the trajectory points of the trajectory to be corrected.

[0278] Exemplarily, the method provided in the aforementioned embodiment can be used to determine whether a trajectory point in the trajectory to be corrected is a normal point or an abnormal point, thereby completing the classification operation of the trajectory points included in the trajectory to be corrected.

[0279] Step 703: Classify abnormal trajectory segments.

[0280] Exemplarily, the method provided in the aforementioned embodiment can be used to analyze the trajectory points in any abnormal trajectory in the trajectory to be corrected, such as the kth abnormal trajectory, so as to determine the abnormal type of the kth abnormal trajectory.

[0281] Step 704: Correct the abnormal trajectory segment.

[0282] Exemplarily, a correction strategy corresponding to any abnormal trajectory may be determined, and the abnormal trajectory segment may be corrected based on the correction strategy.

[0283] Step 705: Output the corrected trajectory.

[0284] Exemplarily, the corrected trajectory may include a trajectory segment obtained by correcting the abnormal trajectory segment.

[0285] From the above, it can be seen that the trajectory correction method provided in the embodiment of the present application, by classifying the trajectory points in the trajectory to be corrected, divides the trajectory to be corrected into abnormal trajectory segments and normal trajectory segments, and performs targeted correction on the abnormal trajectory segments, thereby reducing the negative impact on the normal trajectory end caused by correcting the abnormal trajectory segments and improving the trajectory correction effect.

[0286] Based on the above embodiments, the present application also provides a trajectory correction device. Figure 8 A schematic diagram of the structure of the trajectory correction device provided in the embodiment of the present application is shown in FIG. Figure 8 As shown, the device 8 may include:

[0287] An acquisition module 801 is used to acquire the trajectory to be corrected;

[0288] A determination module 802 is used to determine at least one abnormal trajectory included in the trajectory to be corrected;

[0289] A correction module 803 is used to correct at least one abnormal trajectory to obtain at least one corrected trajectory;

[0290] The determination module 802 is further configured to determine a target trajectory based on at least one corrected trajectory and a normal trajectory included in the trajectory to be corrected.

[0291] In some embodiments, if the trajectory to be corrected includes multiple abnormal trajectories, the mth normal trajectory is included between the kth abnormal trajectory and the k+1th abnormal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the number threshold; k and m are both integers greater than or equal to 1;

[0292] The determination module 802 is used to analyze the trajectory segments except the kth abnormal trajectory in the trajectory to be corrected during the process of correcting the kth abnormal trajectory, so as to determine at least one abnormal trajectory.

[0293] In some embodiments, if the trajectory to be corrected includes multiple abnormal trajectories, the mth normal trajectory is included between the kth abnormal trajectory and the k+1th abnormal trajectory; the number of trajectory points in the mth normal trajectory is greater than or equal to the number threshold; k and m are both integers greater than or equal to 1;

[0294] The determination module 802 is used to correct the k+nth abnormal trajectory and / or the kxth abnormal trajectory during the process of correcting the kth abnormal trajectory; wherein n is an integer greater than or equal to 1; and x is a positive integer less than k.

[0295] In some embodiments, the determination module 802 is used to determine a set of trajectory points included in the trajectory to be corrected; wherein the trajectory points in the set of trajectory points include point types of the trajectory points; and the point type is used to characterize whether the trajectory point is an abnormal point;

[0296] The determination module 802 is further configured to perform statistical division on the trajectory points in the trajectory point set based on point types, and determine at least one abnormal trajectory.

[0297] In some embodiments, the acquisition module 801 is used to acquire adjacent trajectory points of the p-th trajectory point in the trajectory to be corrected; wherein the adjacent trajectory points include the P trajectory points before the p-th trajectory point and the P trajectory points after the p-th trajectory point; p and P are both integers greater than 1;

[0298] A determination module 802 is used to determine motion feature information of the p-th trajectory point and adjacent trajectory points; wherein the motion feature information includes at least one of instantaneous speed, average speed, direction angle, and static point ratio;

[0299] The determination module 802 is further configured to determine the point type of the p-th trajectory point based on the motion feature information.

[0300] In some embodiments, the determination module 802 is configured to determine that the p-th trajectory point is an abnormal point if a gradient between the motion feature information of the p-th trajectory point and the motion feature information of an adjacent trajectory point is greater than a gradient threshold.

[0301] In some embodiments, the determination module 802 is used to classify the trajectory interval formed by the trajectory points between the pth trajectory point and the p+Nth trajectory point as a normal trajectory if the N consecutive trajectory points adjacent to the pth trajectory point are all normal points, and the p-1th trajectory point is an abnormal point and the p+N+1th trajectory point is an abnormal point; wherein p is an integer greater than 1; and N is greater than or equal to a quantity threshold;

[0302] The determination module 802 is further configured to determine that the trajectory segments other than the normal trajectory in the trajectory to be corrected are abnormal trajectories.

[0303] In some embodiments, the determination module 802 is used to determine the abnormal type of the kth abnormal trajectory; wherein k is an integer greater than or equal to 1;

[0304] The determination module 802 is further used to determine a kth correction strategy based on the abnormal type of the kth abnormal trajectory;

[0305] The correction module 803 is used to correct the kth abnormal trajectory based on the kth correction strategy.

[0306] In some embodiments, the determination module 802 is used to perform association analysis on the motion feature information of the trajectory points in the kth abnormal trajectory segment through a decision tree model to determine the abnormal type of the kth abnormal trajectory segment.

[0307] In some embodiments, the acquisition module 801 is used to acquire sample data; wherein the sample data includes motion feature information of trajectory points in a normal trajectory and motion feature information of trajectory points in an abnormal trajectory;

[0308] The determination module 802 is used to train the cascade support vector machine in the initial state based on the sample data to obtain the cascade support vector machine; perform association analysis on the motion feature information of the trajectory points in the kth abnormal trajectory through the cascade support vector machine to determine the abnormal type of the kth abnormal trajectory.

[0309] In some embodiments, the acquisition module 801 is used to acquire a target normal trajectory adjacent to the kth abnormal trajectory; wherein the target normal trajectory includes a normal trajectory adjacent to the kth abnormal trajectory;

[0310] A determination module 802 is used to determine a first position and a second position; wherein the first position includes the position of the starting trajectory point and the ending trajectory point of the kth abnormal trajectory; and the second position includes the position of a normal point in the target normal trajectory that is adjacent to the starting trajectory point and the abnormal trajectory point of the kth abnormal trajectory;

[0311] The determination module 802 is further configured to determine that the abnormality type is the first type if the distance between the first position and the second position is greater than a first distance threshold.

[0312] In some embodiments, the determination module 802 is used to cluster the positions of the trajectory points in the kth abnormal trajectory to obtain the cluster center; if the distance between the trajectory points in the kth abnormal trajectory and the cluster center is less than the second distance threshold, the abnormal type is determined to be the second type.

[0313] In some embodiments, the determination module 802 is used to obtain a target normal trajectory; wherein the target normal trajectory includes a normal trajectory adjacent to the kth abnormal trajectory;

[0314] A determination module 802 is used to determine a trajectory point switching frequency; wherein the trajectory point switching frequency includes a switching frequency between a trajectory point in the kth abnormal trajectory and a trajectory point in the target normal trajectory;

[0315] The determination module 802 is used to determine that the abnormal feature is of the third type if the switching frequency of the trajectory point is greater than the switching threshold.

[0316] In some embodiments, the determination module 802 is used to determine the kth correction strategy as follows if the kth abnormal trajectory is of the first type:

[0317] Determine the distance similarity between the first distance and the second distance; wherein the first distance includes the distance between at least two normal points adjacent to the kth abnormal trajectory; and the second distance includes the distance between normal points in the kth abnormal trajectory;

[0318] Determine the angular similarity between the first angle and the second angle; wherein the first angle includes the angle associated with at least two normal points adjacent to the kth abnormal trajectory; the second angle includes the angle associated with the normal point in the kth abnormal trajectory;

[0319] The fitting parameters are determined based on the distance similarity and the angle similarity, and the kth abnormal trajectory is fitted and corrected based on the fitting parameters.

[0320] In some embodiments, the determination module 802 is configured to fit and correct the kth abnormal trajectory based on the position of the normal point in the kth abnormal trajectory if the fitting parameter is greater than the fitting threshold;

[0321] If the fitting parameter is less than or equal to the fitting threshold, the kth abnormal trajectory is corrected by fitting based on at least two normal points adjacent to the kth abnormal trajectory.

[0322] In some embodiments, the determination module 802 is used to determine the kth correction strategy if the kth abnormal trajectory is of the second type: clustering the positions of the trajectory points in the kth abnormal trajectory to obtain the cluster center, and correcting the kth abnormal trajectory based on the cluster center.

[0323] In some embodiments, the determination module 802 is used to determine the kth correction strategy as follows if the kth abnormal trajectory is of the third type:

[0324] The kth abnormal trajectory is corrected based on at least the distribution information of the abnormal points and the normal points in the kth abnormal trajectory, wherein the distribution information includes time information and position information.

[0325] Based on the foregoing embodiments, an embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored; when the computer program is executed by a processor of an electronic device, the trajectory correction method provided in any of the previous embodiments can be implemented.

[0326] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0327] The methods disclosed in the various method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0328] The features disclosed in the various product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0329] The features disclosed in the various method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

[0330] It should be noted that the above-mentioned computer-readable storage medium can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM) and other memories; it can also be various electronic devices including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0331] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0332] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0333] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus necessary general hardware nodes, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0334] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0335] 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 an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0336] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0337] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A trajectory correction method, characterized in that, the method includes: obtaining a trajectory to be corrected; determining at least one abnormal trajectory included in the trajectory to be corrected; correcting the at least one abnormal trajectory to obtain at least one corrected trajectory; determining a target trajectory based on the at least one corrected trajectory and the normal trajectory included in the trajectory to be corrected.

2. The method according to claim 1, characterized in that, if the trajectory to be corrected includes multiple abnormal trajectories, there is a m-th normal trajectory between the k-th abnormal trajectory and the (k + 1)-th abnormal trajectory; the number of trajectory points in the m-th normal trajectory is greater than or equal to a quantity threshold; both k and m are integers greater than or equal to 1; the determining of at least one abnormal trajectory included in the trajectory to be corrected includes: during the process of correcting the k-th abnormal trajectory, analyzing the trajectory segments of the trajectory to be corrected except the k-th abnormal trajectory to determine the at least one abnormal trajectory.

3. The method according to claim 1, characterized in that, if the trajectory to be corrected includes multiple abnormal trajectories, there is a m-th normal trajectory between the k-th abnormal trajectory and the (k + 1)-th abnormal trajectory; the number of trajectory points in the m-th normal trajectory is greater than or equal to a quantity threshold; both k and m are integers greater than or equal to 1; the correcting of the at least one abnormal trajectory to obtain at least one corrected trajectory includes: during the process of correcting the k-th abnormal trajectory, correcting the (k + n)-th abnormal trajectory and / or the (k - x)-th abnormal trajectory; where n is an integer greater than or equal to 1; x is a positive integer less than k.

4. The method according to claim 1, characterized in that, the determining of at least one abnormal trajectory included in the trajectory to be corrected includes: determining a set of trajectory points included in the trajectory to be corrected; wherein, the trajectory points in the set of trajectory points include the point types of the trajectory points; the point types are used to characterize whether the trajectory points are abnormal points; based on the point types, statistically partitioning the trajectory points in the set of trajectory points to determine the at least one abnormal trajectory.

5. The method according to claim 4, characterized in that, the method further includes: obtaining adjacent trajectory points of the p-th trajectory point in the trajectory to be corrected; wherein, the adjacent trajectory points include P trajectory points before the p-th trajectory point and P trajectory points after the p-th trajectory point; both p and P are integers greater than 1; determining the motion feature information of the p-th trajectory point and the adjacent trajectory points; wherein, the motion feature information includes at least one of instantaneous velocity, average velocity, direction angle, and the proportion of stationary points; determining the point type of the p-th trajectory point based on the motion feature information.

6. The method according to claim 5, characterized in that, the determining of the point type of the p-th trajectory point based on the motion feature information includes: if the gradient between the motion feature information of the p-th trajectory point and the motion feature information of the adjacent trajectory points is greater than a gradient threshold, determining the p-th trajectory point as an abnormal point.

7. The method according to claim 4, It is characterized in that statistically dividing the trajectory points in the set of trajectory points based on the point type to determine the at least one abnormal trajectory, including: If the consecutive N trajectory points adjacent to the p-th trajectory point are all normal points, and the (p - 1)-th trajectory point is an abnormal point and the (p + N + 1)-th trajectory point is the abnormal point, dividing the trajectory interval formed by the trajectory points between the p-th trajectory point and the (p + N)-th trajectory point into a normal trajectory; where p is an integer greater than 1; N is greater than or equal to the quantity threshold; Determining the trajectory segment other than the normal trajectory in the to-be-corrected trajectory as the abnormal trajectory.

8. The method according to claim 1, It is characterized in that correcting the at least one abnormal trajectory, including: determining the abnormal type of the k-th abnormal trajectory; where k is an integer greater than or equal to 1; determining the k-th correction strategy based on the abnormal type of the k-th abnormal trajectory; correcting the k-th abnormal trajectory based on the k-th correction strategy.

9. The method according to claim 8, It is characterized in that determining the abnormal type of the k-th abnormal trajectory includes: performing correlation analysis on the motion feature information of the trajectory points in the k-th abnormal trajectory through a decision tree model to determine the abnormal type of the k-th abnormal trajectory.

10. The method according to claim 8, It is characterized in that determining the abnormal type of the k-th abnormal trajectory includes: obtaining sample data; where the sample data includes the motion feature information of the trajectory points in the normal trajectory and the motion feature information of the trajectory points in the abnormal trajectory; training a cascaded support vector machine in an initial state based on the sample data to obtain a cascaded support vector machine; performing correlation analysis on the motion feature information of the trajectory points in the k-th abnormal trajectory through the cascaded support vector machine to determine the abnormal type of the k-th abnormal trajectory.

11. The method according to claim 8, It is characterized in that determining the abnormal type of the k-th abnormal trajectory includes: obtaining a target normal trajectory adjacent to the k-th abnormal trajectory; where the target normal trajectory includes the normal trajectory adjacent to the k-th abnormal trajectory; determining a first position and a second position; where the first position includes the positions of the starting trajectory point and the ending trajectory point of the k-th abnormal trajectory; the second position includes the positions of the normal points adjacent to the starting trajectory point and the abnormal trajectory point in the target normal trajectory; if the distance between the first position and the second position is greater than a first distance threshold, determining the abnormal type as the first type.

12. The method according to claim 8, It is characterized in that determining the abnormal type of the k-th abnormal trajectory includes: performing clustering processing on the positions of the trajectory points in the k-th abnormal trajectory to obtain a clustering center; if the distance between the trajectory points in the k-th abnormal trajectory and the clustering center is less than a second distance threshold, determining the abnormal type as the second type.

13. The method according to claim 8, It is characterized in that determining the abnormal type of the k-th abnormal trajectory includes: Obtain the target normal trajectory; wherein, the target normal trajectory includes the normal trajectory adjacent to the k-th abnormal trajectory; Determine the trajectory point switching frequency; wherein, the trajectory point switching frequency includes the switching frequency between the trajectory points in the k-th abnormal trajectory and the trajectory points in the target normal trajectory; If the trajectory point switching frequency is greater than the switching threshold, determine that the abnormal feature is of the third type.

14. The method according to claim 8, wherein, The determining the k-th correction strategy based on the abnormal type of the k-th abnormal trajectory includes: If the k-th abnormal trajectory is of the first type, determine that the k-th correction strategy is: Determine the distance similarity between a first distance and a second distance; wherein, the first distance includes the distance between at least two normal points adjacent to the k-th abnormal trajectory; the second distance includes the distance between the normal points in the k-th abnormal trajectory; Determine the angle similarity between a first angle and a second angle; wherein, the first angle includes the angles associated with at least two normal points adjacent to the k-th abnormal trajectory; the second angle includes the angles associated with the normal points in the k-th abnormal trajectory; Determine the fitting parameter based on the distance similarity and the angle similarity, and fit and correct the k-th abnormal trajectory based on the fitting parameter.

15. The method according to claim 14, wherein, The fitting and correcting the k-th abnormal trajectory based on the fitting parameter includes: If the fitting parameter is greater than the fitting threshold, fit and correct the k-th abnormal trajectory based on the positions of the normal points in the k-th abnormal trajectory; If the fitting parameter is less than or equal to the fitting threshold, fit and correct the k-th abnormal trajectory based on at least two normal points adjacent to the k-th abnormal trajectory.

16. The method according to claim 8, wherein, The determining the k-th correction strategy based on the abnormal type of the k-th abnormal trajectory includes: If the k-th abnormal trajectory is of the second type, determine that the k-th correction strategy is: perform clustering on the positions of the trajectory points in the k-th abnormal trajectory to obtain a clustering center, and correct the k-th abnormal trajectory based on the clustering center.

17. The method according to claim 8, wherein, The determining the k-th correction strategy based on the abnormal type of the k-th abnormal trajectory includes: If the k-th abnormal trajectory is of the third type, determine that the k-th correction strategy is: Correct the k-th abnormal trajectory based on at least the distribution information of the abnormal points and normal points in the k-th abnormal trajectory; wherein, the distribution information includes time information and position information.

18. A trajectory correction device, wherein, The device includes: An acquisition module, configured to acquire the trajectory to be corrected; A determination module, configured to determine at least one abnormal trajectory included in the trajectory to be corrected; A correction module, configured to correct the at least one abnormal trajectory to obtain at least one corrected trajectory; The determining module is further configured to determine a target trajectory based on the at least one corrected trajectory and the normal trajectory included in the trajectory to be corrected.

19. A computer-readable storage medium, characterized in that, a computer program is stored in the storage medium; when the computer program is executed by a processor of an electronic device, the trajectory correction method according to any one of claims 1 to 17 can be implemented.

Images