Motion track processing method and device, equipment and storage medium

Abstract

The embodiment of the invention discloses a motion trail processing method and device, equipment and a storage medium. The method comprises the following steps: acquiring a target motion track corresponding to a target object; the target motion trail is input into a target circulation network model to carry out mounting prediction of staying information points, the target circulation network model is obtained through training based on sample object data in advance, and the sample object data comprises sample motion trails corresponding to sample objects and sample staying information point data; and obtaining target stay information point data corresponding to the target object based on the output of the target loop network model. Through the technical scheme of the embodiment of the invention, automatic mounting between the motion trail and the staying information point can be realized, and the staying information point of the object can be obtained from the redundant motion trail.

Description

Technical Field

[0001] The embodiments of the present invention relate to computer technology, and in particular to a motion trajectory processing method, device, equipment and storage medium. Background Art

[0002] With the rapid development of computer technology, the motion trajectory of objects such as cars or people is the most direct characterization of the physical activities of the object. Therefore, the motion trajectory can reflect the properties and behaviors of the object, which is of great significance in applications such as product recommendation, navigation, and taxi-hailing.

[0003] However, in the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] Since the trajectory points in the motion trajectory are location points in the geographical sense, they cannot directly reflect the object's stop information points (Point of Information, POI) in the physical world, such as stopping at stores, bars, gas stations, etc., so there is an urgent need for a way to automatically mount the motion trajectory with the stop information points, so as to link the geometric motion trajectory with the commercial activities in the physical world. Summary of the invention

[0005] The embodiments of the present invention provide a motion trajectory processing method, device, equipment and storage medium to realize automatic mounting between motion trajectories and stop information points, and obtain object stop information points from redundant motion trajectories.

[0006] In a first aspect, an embodiment of the present invention provides a motion trajectory processing method, comprising:

[0007] Obtain the target motion trajectory corresponding to the target object;

[0008] Inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object;

[0009] Based on the output of the target recurrent network model, target stay information point data corresponding to the target object is obtained.

[0010] In a second aspect, an embodiment of the present invention further provides a motion trajectory processing device, comprising:

[0011] A motion trajectory acquisition module is used to acquire a target motion trajectory corresponding to a target object;

[0012] A mounting prediction module, used for inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object;

[0013] The stay information point determination module is used to obtain the target stay information point data corresponding to the target object based on the output of the target cyclic network model.

[0014] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0015] one or more processors;

[0016] A memory for storing one or more programs;

[0017] When the one or more programs are executed by the one or more processors, the one or more processors implement the motion trajectory processing method provided by any embodiment of the present invention.

[0018] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a motion trajectory processing method as provided in any embodiment of the present invention.

[0019] One embodiment of the above invention has the following advantages or beneficial effects:

[0020] The target recurrent network model is obtained by pre-training based on the sample motion trajectory and sample stop information point data corresponding to the sample object. The target recurrent network model can automatically fit the features with long-term dependence in the motion trajectory. Therefore, the target recurrent network model can be used to predict the mounting of the stop information point of the target motion trajectory corresponding to the target object, and the target stop information point data corresponding to the target object can be quickly and accurately obtained from the redundant target motion trajectories, thereby realizing the automatic mounting between the motion trajectory and the stop information point. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 is a flow chart of a motion trajectory processing method provided by an embodiment of the present invention;

[0023] Figure 2 is an example of a target motion trajectory involved in an embodiment of the present invention;

[0024] Figure 3 is a flow chart of another motion trajectory processing method provided by an embodiment of the present invention;

[0025] Figure 4 is an example diagram of the architecture of a target recurrent network model involved in an embodiment of the present invention;

[0026] Figure 5 is a structural schematic diagram of a motion trajectory processing device provided by an embodiment of the present invention;

[0027] Figure 6 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0029] It should be noted that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of user personal information involved in the technical solution of this disclosure are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information, network security and national security.

[0030] Figure 1 This is a flow chart of a motion trajectory processing method provided by an embodiment of the present invention. This embodiment can be applied to the case of determining the stop information points mounted on the motion trajectory. The method can be executed by a motion trajectory processing device, which can be implemented by software and / or hardware and integrated into an electronic device. Figure 1 As shown, the method specifically comprises the following steps:

[0031] S110: Obtain a target motion trajectory corresponding to the target object.

[0032] The target object may be any object that can move. For example, the target object may be, but is not limited to, a person or a vehicle. The target motion trajectory may refer to the motion trajectory generated by the target object. The target motion trajectory is essentially time series data. For example, the target motion trajectory may include multiple trajectory point data, i.e., Traj=(Traj 1,Traj 2 …Traj n ). Each track point data may include: the longitude and latitude coordinates of the track point and a movement timestamp. The track point may refer to a location point with geographical significance on a map. The longitude and latitude coordinates of the track point may refer to the longitude and latitude coordinates of the current location of the target object, which may be obtained by satellite positioning of the target object. The movement timestamp may be used to characterize the time when the target object moves to the track point.

[0033] Specifically, the target motion trajectory generated by the target object that has been collected in advance can be obtained, or the target motion trajectory generated by the target object can be obtained in real time. Figure 2 An example of a target motion trajectory is given. Figure 2 As shown, the target motion trajectory is the trajectory of the target object moving from point A to point B.

[0034] S120, inputting the target motion trajectory into the target cyclic network model to perform mounting prediction of the stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object.

[0035] Among them, the stop information point may refer to an information point POI where the target object has stopped. An information point may refer to a location point on a map that has no geographical significance but has business significance. For example, an information point may refer to a location in the physical world where business activities can be generated, such as a store, a bar, a gas station, a hospital, a bus station, etc. An information point database may be established in advance, which contains various types of information point POIs. Mounting prediction is to associate the track points in the motion trajectory with the information points to determine the information point POI where the target object has stopped during the movement. It should be noted that the track point only refers to the latitude and longitude coordinates of the current location of the target object, which cannot directly reflect the current activity place of the target object, so it is necessary to mount the track point and the information point. The correspondence between the track point and the information point can be many-to-one, that is, multiple track points are mounted to the same information point, so that the track point is redundant with respect to the information point.

[0036] Among them, the target recurrent network model can be a recurrent neural network model RNN (Recurrent Neural Network) used to determine the stop information points mounted on the motion trajectory. Since there are long-term dependency features in the motion trajectory, the recurrent neural network model can better fit the long-term dependency features of the time series and accurately mine the stop information points in the motion trajectory. The target recurrent network model can be obtained by pre-training based on sample object data to ensure the accuracy of the mounting prediction of the target recurrent network model. The sample object can be an object used for model training. The sample motion trajectory can refer to the historical motion trajectory generated by the sample object. The sample stop information point data can refer to the information point data where the sample object has stopped during the movement. The sample stop information point data can be used as label data for model training.

[0037] Specifically, the target recurrent network model can be trained in advance based on the sample motion trajectory and sample stop information point data corresponding to the sample object, so that the target recurrent network model can autonomously learn the stop time and distance thresholds, thereby accurately determining the mounting relationship between the trajectory point and the stop information point. The target motion trajectory is input into the trained target recurrent network model, and the target recurrent network model can automatically predict the mounting of the stop information point of the input target motion trajectory, determine the target stop information point data mounted with the target motion trajectory and output it.

[0038] S130. Based on the output of the target recurrent network model, obtain target stay information point data corresponding to the target object.

[0039] Among them, the target stop information point data may include: the longitude and latitude coordinates of the target stop information point. The target stop information point may refer to the information point where the target object has stopped during the movement. The number of target stop information points may be one or more. The target stop information point has a certain area range. The longitude and latitude coordinates of the target stop information point may refer to the longitude and latitude coordinates at the center of the target stop information point, that is, the longitude and latitude coordinates of the center of the POI. For example, if the stop information point is a school, the longitude and latitude coordinates at the center of the geographical area occupied by the school can be used as the longitude and latitude coordinates of the stop information point. In addition to the longitude and latitude coordinates of the target stop information point, the target stop information point data may also include: at least one of the stop information point radius, stop duration, stop start time and stop end time corresponding to the target stop information point. Among them, the stop information point radius can be used to characterize the trajectory range where the target object stops, which can show the distribution of the stop trajectory points to a certain extent, such as whether it is relatively concentrated or relatively dispersed, and can also reflect the credibility of the mounting prediction from the side, such as the credibility is relatively high in the case of concentration. The stop duration may refer to the length of time the target object stays at the information point. The stop start time may refer to the time when the target object arrives at the information point. The stop end time may refer to the time when the target object leaves the information point.

[0040] Specifically, the target stop information point data output by the target loop network model can be quickly obtained, thereby realizing automatic mounting between the motion trajectory and the stop information point. The output target stop information point data can be characterized in the form of a target stop information point sequence. For example, the target stop information point sequence may include one or more target stop information point data. By utilizing the target loop network model, the motion trajectory sequence can be translated into a stop information point sequence, thereby accurately mining the activity places where the object has stayed in the redundant motion trajectory. Exemplarily, after determining the longitude and latitude coordinates of the target stop information point, the target place name corresponding to each target stop information point can be determined based on the correspondence between the longitude and latitude coordinates of the information point and the place name, and the target place name is displayed, so that the activity places where the target object has stayed during the movement can be more clearly known.

[0041] For example, Figure 2The target motion trajectory in is input into the target cyclic network model for mounting prediction of the stop information point, and the target stop information point data mounted with the target motion trajectory are determined as follows: {service area a, stop 17.01mins, stop start time 2022-08-31 14:24:09, stop end time 2022-08-3114:41:10; checkpoint b, stop 15.86mins, stop start time 2022-08-31 21:16:10, stop end time 2022-08-31 21:32:02; service area c, stop 19.66mins, stop start time 2022-08-31 10:08:48, stop end time 2022-08-3110:28:28}.

[0042] The technical solution of this embodiment is to obtain a target recurrent network model by pre-training based on the sample motion trajectory and sample stop information point data corresponding to the sample object. The target recurrent network model can automatically fit the features with long-term dependence in the motion trajectory, so that the target recurrent network model can be used to predict the mounting of the stop information point of the target motion trajectory corresponding to the target object, and the target stop information point data corresponding to the target object can be quickly and accurately obtained from the redundant target motion trajectory, thereby realizing the automatic mounting between the motion trajectory and the stop information point.

[0043] Based on the above technical solution, the training process of the target recurrent network model may include the following steps S210-S240:

[0044] S210: Obtain sample motion trajectory and sample stay information point data corresponding to the sample object.

[0045] Among them, the sample object can be but is not limited to logistics vehicles in logistics scenarios. The number of sample objects is multiple to ensure the model training effect. The sample motion trajectory can be a cross-provincial trajectory of more than 100 kilometers or a city trajectory of less than 100 kilometers. The sample motion trajectory may include multiple trajectory point data, and each trajectory point data may include: the latitude and longitude coordinates of the trajectory point and the motion timestamp. The data dimension of the sample motion trajectory is N*3, where N is the number of trajectory points and 3 is the number of features of each trajectory point. The sample stop information point data may be obtained by manual marking and is used for training model label data. The sample stop information point data may include the latitude and longitude coordinates of the sample stop information point. In addition to the latitude and longitude coordinates of the sample stop information point, the sample stop information point data may also include: at least one of the stop information point radius, stop duration, stop start time and stop end time corresponding to the sample stop information point. The data types contained in the sample stop information point data can be set based on business needs.

[0046] S220: Input the sample motion trajectory into a preset cyclic network model to perform mounting prediction of the stop information point, and obtain predicted stop information point data corresponding to the sample object.

[0047] The preset recurrent network model may be a recurrent neural network model that is created first and has not been trained. Specifically, the sample motion trajectory is input into the preset recurrent network model, and the preset recurrent network model predicts the mounting of the stop information point of the sample motion trajectory based on the current model parameters, and obtains the prediction result output by the preset recurrent network model, that is, the predicted stop information point data corresponding to the sample object.

[0048] S230: Determine a target training error corresponding to the sample object based on the predicted stay information point data and the sample stay information point data.

[0049] Specifically, the predicted stay information point data currently output by the preset recurrent network model can be compared with the sample stay information point data that should be output, so as to determine the target training error corresponding to the sample object.

[0050] Exemplarily, S230 may include: determining a first training error based on a first loss function, the longitude and latitude coordinates of the predicted stay information point, and the longitude and latitude coordinates of the sample stay information point; determining a second training error based on a second loss function, the radius of the predicted stay information point, and the radius of the sample stay information point; adding the first training error to the second training error to obtain a target training error corresponding to the sample object.

[0051] The first loss function may be a function for calculating the error of the longitude and latitude coordinates. For example, the first loss function may be, but is not limited to, a function that uses log to amplify the error value. The second loss function may be a function for calculating the error of the radius of the stay information point. The second loss function may be, but is not limited to, a mean square error function.

[0052] Specifically, the sample stop information point data includes: the latitude and longitude coordinates of the sample stop information point i (i.e. and ) and the radius of the sample stop information point The predicted stop information point data includes: the longitude and latitude coordinates of the predicted stop information point i (i.e., center X i and center Y i ) and the radius of the predicted stop information point centerR i When the center X i and The longitude difference between the center Y i and The latitude difference between the two values ​​is calculated by using logarithm to amplify the longitude difference and latitude difference to obtain the first training error. i and The radius difference between the two points is calculated, and the radius difference is processed by the mean square to obtain the second training error. The first training error and the second training error are added to obtain the final target training error of the sample object. By integrating the errors of the latitude and longitude coordinates and the radius of the information point, the accuracy of the model training can be further guaranteed and the model training effect can be improved.

[0053] S240, back-propagating the target training error to the preset recurrent network model, adjusting the model parameters in the preset recurrent network model, and determining that the training of the preset recurrent network model is completed until a preset convergence condition is reached, and obtaining the target recurrent network model.

[0054] Specifically, the target training error of the current time is back-propagated to the preset recurrent network model, and the model parameters in the preset recurrent network model are adjusted until the preset convergence condition is reached, such as when the number of iterations is equal to the preset number, or when the target training error changes tend to be stable, the preset recurrent network model is determined to be finished, and the preset recurrent network model after training can be used as the target recurrent network model. Through model training, the target recurrent network model can have the function of automatically mounting the motion trajectory and the stop information point, thereby realizing end-to-end information point mounting.

[0055] Figure 3 A flowchart of another motion trajectory processing method provided by an embodiment of the present invention. In this embodiment, based on the above embodiments, the target recurrent network model may include an encoder, a decoder, and a fully connected layer, and on this basis, the mounting prediction process of the target recurrent network model is described in detail. The explanations of the terms that are the same or corresponding to the above embodiments are not repeated here.

[0056] See also Figure 3 Another motion trajectory processing method provided in this embodiment specifically includes the following steps:

[0057] S310: Obtain a target motion trajectory corresponding to the target object.

[0058] S320, inputting the target motion trajectory into the encoder to extract hidden features and obtain a target hidden feature vector.

[0059] Among them, the encoder can be used to extract time-related hidden feature information in the input information. The encoder may include multiple encoding layers. Each encoding layer may be a long short-term memory neural network (LSTM). The cell state in the LSTM is equivalent to the path of information transmission, allowing information to be passed down in the sequence, and even information from earlier time steps can be carried to cells in later time steps, thereby overcoming the influence of short-term memory. The addition and removal of information in the LSTM is achieved through a gate structure, which can learn which information to save or forget during the training process. The target hidden feature information in the target hidden feature vector may include, but is not limited to: the movement speed, acceleration, starting position, end position, stop position and stop duration corresponding to the target object.

[0060] Specifically, Figure 4 A schematic diagram of the architecture of a target recurrent network model is given. Figure 4 As shown, the N*3-dimensional target motion trajectory can be input into the encoder. Each encoding layer in the encoder can embed the input information into a 32-dimensional vector space, convert it into a spatial feature vector that can be accepted by the network, and obtain the target hidden feature vector output by the last encoding layer. For example, the t-th encoding layer in the encoder can be expressed as follows:

[0061]

[0062]

[0063] in, is the longitude and latitude coordinates of trajectory point i at time t. FC1 is the embedding layer in the encoding layer. FC1 are the parameters of the embedding layer. i (t) is the output of the encoding layer. W is the hidden feature vector output by the previous encoding layer t-1. encoder Parameters of the encoding function LSTM. is the hidden feature vector at time t. The entire motion trajectory can be encoded into a target hidden feature vector containing hidden feature information through the encoding layer.

[0064] S330: Input the target hidden feature vector into the decoder to perform hidden feature decoding to obtain a target decoded feature vector.

[0065] The decoder may be used to decode the encoded features accordingly. The decoder may also include multiple decoding layers. Each decoding layer may also be a long short-term memory neural network LSTM.

[0066] Specifically, Figure 4 As shown, the target hidden feature vector output by the last encoding layer can be input into the decoder for hidden feature decoding to obtain the target decoded feature vector output by the last decoding layer. For example, the tth encoding layer in the encoder can be expressed as follows:

[0067]

[0068] in, is the output of the encoding layer LSTM at time t, is the hidden vector at time t-1, W decoder Parameters of the encoding function LSTM.

[0069] S340, inputting the target decoded feature vector into the fully connected layer to perform mounting prediction of the stop information point, and obtaining predicted target stop information point data.

[0070] The fully connected layer may be used to map the output result of the decoder with the information points in the POI database, so as to predict the stop information points mounted on the track points. For example, the fully connected layer may be, but is not limited to, a multilayer perceptron (MLP).

[0071] Specifically, Figure 4 As shown, the target decoded feature vector output by the decoder is input into the fully connected layer, and the splicing mapping is performed in the fully connected layer to determine the target stop information point data mounted on the motion trajectory and output it. For example, the fully connected layer can be expressed as follows:

[0072]

[0073] Among them, W MLP are the parameters of the fully connected layer MLP. The predicted mounting result.

[0074] S350: Based on the output of the fully connected layer, obtain the target stop information point data corresponding to the target object.

[0075] The technical solution of this embodiment is to obtain a target hidden feature vector by inputting the target motion trajectory into the encoder for hidden feature extraction; inputting the target hidden feature vector into the decoder for hidden feature decoding to obtain a target decoded feature vector; inputting the target decoded feature vector into the fully connected layer for mounting prediction of the stop information point to obtain predicted target stop information point data, thereby extracting hidden features to more accurately mine the stop information points in the motion trajectory, further improving the accuracy of mounting prediction.

[0076] The following is an embodiment of a motion trajectory processing device provided by an embodiment of the present invention. The device and the motion trajectory processing methods of the above embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the motion trajectory processing device, reference can be made to the embodiment of the above motion trajectory processing method.

[0077] Figure 5 Schematic diagram of the structure of a motion trajectory processing device provided by an embodiment of the present invention. This embodiment is applicable to the case of determining the stop information point mounted on the motion trajectory. Figure 5 As shown, the device specifically includes: a motion trajectory acquisition module 510, a mounting prediction module 520 and a stop information point determination module 530.

[0078] Among them, the motion trajectory acquisition module 510 is used to obtain the target motion trajectory corresponding to the target object; the mounting prediction module 520 is used to input the target motion trajectory into the target cyclic network model to perform mounting prediction of the stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes sample motion trajectories and sample stop information point data corresponding to the sample object; the stop information point determination module 530 is used to obtain the target stop information point data corresponding to the target object based on the output of the target cyclic network model.

[0079] The technical solution of this embodiment is to obtain a target recurrent network model by pre-training based on the sample motion trajectory and sample stop information point data corresponding to the sample object. The target recurrent network model can automatically fit the features with long-term dependence in the motion trajectory, so that the target recurrent network model can be used to predict the mounting of the stop information point of the target motion trajectory corresponding to the target object, and the target stop information point data corresponding to the target object can be quickly and accurately obtained from the redundant target motion trajectory, thereby realizing the automatic mounting between the motion trajectory and the stop information point.

[0080] Optionally, the target motion trajectory includes a plurality of trajectory point data, each trajectory point data includes: latitude and longitude coordinates of the trajectory point and a motion timestamp; wherein the trajectory point refers to a location point with geographical significance;

[0081] The target stop information point data includes: the latitude and longitude coordinates of the target stop information point; wherein the information point refers to a location point that has no geographical significance but has business significance.

[0082] Optionally, the target stop information point data further includes: at least one of a stop information point radius, a stop duration, a stop start time and a stop end time corresponding to the target stop information point.

[0083] Optionally, the target recurrent network model includes: an encoder, a decoder, and a fully connected layer;

[0084] The mount prediction module 520 is specifically used for:

[0085] The target motion trajectory is input into the encoder for hidden feature extraction to obtain a target hidden feature vector; the target hidden feature vector is input into the decoder for hidden feature decoding to obtain a target decoded feature vector; the target decoded feature vector is input into the fully connected layer for mounting prediction of the stop information point to obtain predicted target stop information point data.

[0086] Optionally, the target hidden feature information in the target hidden feature vector includes: movement speed, acceleration, starting position, end position, stop position and stop duration corresponding to the target object;

[0087] The encoder includes multiple encoding layers, and the decoder includes multiple decoding layers; wherein both the encoding layer and the decoding layer are long short-term memory neural networks.

[0088] Optionally, the device further includes: a target cyclic network model training module, including:

[0089] A sample object data acquisition unit, used to acquire sample motion trajectory and sample stop information point data corresponding to the sample object;

[0090] A sample motion trajectory input unit, used to input the sample motion trajectory into a preset cyclic network model to perform mounting prediction of a stop information point, and obtain predicted stop information point data corresponding to the sample object;

[0091] a target training error determining unit, configured to determine a target training error corresponding to the sample object based on the predicted stay information point data and the sample stay information point data;

[0092] The model parameter adjustment unit is used to back-propagate the target training error to the preset recurrent network model, adjust the model parameters in the preset recurrent network model, and determine that the training of the preset recurrent network model is completed when the preset convergence condition is reached to obtain the target recurrent network model.

[0093] Optionally, the target training error determination unit is specifically used to:

[0094] Based on the first loss function, the longitude and latitude coordinates of the predicted stay information point, and the longitude and latitude coordinates of the sample stay information point, a first training error is determined; based on the second loss function, the radius of the predicted stay information point, and the radius of the sample stay information point, a second training error is determined; the first training error is added to the second training error to obtain the target training error corresponding to the sample object.

[0095] The motion trajectory processing device provided in the embodiment of the present invention can execute the motion trajectory processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the motion trajectory processing method.

[0096] It is worth noting that in the embodiment of the above-mentioned motion trajectory processing device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0097] Figure 6 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 6 A block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 6 The electronic device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0098] like Figure 6 As shown, the electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0099] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0100] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0101] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 6 not shown, usually called a "hard drive"). Although Figure 6 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The system memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0102] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0103] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 12, and / or may communicate with any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the electronic device 12 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via a bus 18. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0104] The processing unit 16 executes various functional applications and data processing by running the program stored in the system memory 28, for example, implementing a motion trajectory processing method provided by the embodiment of the present invention, the method comprising:

[0105] Obtain the target motion trajectory corresponding to the target object;

[0106] Inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object;

[0107] Based on the output of the target recurrent network model, target stay information point data corresponding to the target object is obtained.

[0108] Of course, those skilled in the art can understand that the processor can also implement the technical solution of the motion trajectory processing method provided by any embodiment of the present invention.

[0109] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the motion trajectory processing method provided in any embodiment of the present invention are implemented. The method includes:

[0110] Obtain the target motion trajectory corresponding to the target object;

[0111] Inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object;

[0112] Based on the output of the target recurrent network model, target stay information point data corresponding to the target object is obtained.

[0113] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0114] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0115] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0116] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0117] It should be understood by those skilled in the art that the modules or steps of the present invention described above can be implemented by a general-purpose computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, optionally, they can be implemented by a program code executable by a computer device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0118] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A motion trajectory processing method, characterized in that: include: Obtain the target motion trajectory corresponding to the target object; Inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object; Based on the output of the target recurrent network model, target stay information point data corresponding to the target object is obtained.

2. The method according to claim 1, characterized in that The target motion trajectory includes a plurality of trajectory point data, each of which includes: the latitude and longitude coordinates of the trajectory point and a motion timestamp; wherein the trajectory point refers to a location point with geographical significance; The target stop information point data includes: the latitude and longitude coordinates of the target stop information point; wherein the information point refers to a location point that has no geographical significance but has business significance.

3. The method according to claim 2, characterized in that The target stop information point data also includes at least one of a stop information point radius, a stop duration, a stop start time, and a stop end time corresponding to the target stop information point.

4. The method according to claim 1, characterized in that The target recurrent network model includes: an encoder, a decoder and a fully connected layer; The step of inputting the target motion trajectory into the target cyclic network model to perform mounting prediction of the stop information point includes: Inputting the target motion trajectory into the encoder to extract hidden features and obtain a target hidden feature vector; Inputting the target hidden feature vector into the decoder to perform hidden feature decoding to obtain a target decoded feature vector; The target decoded feature vector is input into the fully connected layer to perform mounting prediction of the stop information point, and the predicted target stop information point data is obtained.

5. The method according to claim 4, characterized in that The target hidden feature information in the target hidden feature vector includes: the movement speed, acceleration, starting position, end position, stop position and stop duration corresponding to the target object; The encoder includes multiple encoding layers, and the decoder includes multiple decoding layers; wherein both the encoding layer and the decoding layer are long short-term memory neural networks.

6. The method according to claim 1, characterized in that The training process of the target recurrent network model includes: Obtain the sample motion trajectory and sample stay information point data corresponding to the sample object; Inputting the sample motion trajectory into a preset cyclic network model to perform mounting prediction of the stop information point, and obtaining the predicted stop information point data corresponding to the sample object; Determine a target training error corresponding to the sample object based on the predicted stay information point data and the sample stay information point data; The target training error is back-propagated to the preset recurrent network model, and the model parameters in the preset recurrent network model are adjusted until a preset convergence condition is reached, and it is determined that the training of the preset recurrent network model is completed to obtain the target recurrent network model.

7. The method according to claim 6, characterized in that The determining the target training error corresponding to the sample object based on the predicted stay information point data and the sample stay information point data includes: Determining a first training error based on the first loss function, the longitude and latitude coordinates of the predicted stay information point, and the longitude and latitude coordinates of the sample stay information point; Determining a second training error based on the second loss function, the predicted stop information point radius, and the sample stop information point radius; The first training error is added to the second training error to obtain a target training error corresponding to the sample object.

8. A motion trajectory processing device, characterized in that: include: A motion trajectory acquisition module is used to acquire a target motion trajectory corresponding to a target object; A mounting prediction module, used for inputting the target motion trajectory into a target cyclic network model to perform mounting prediction of a stop information point, wherein the target cyclic network model is pre-trained based on sample object data, and the sample object data includes a sample motion trajectory and sample stop information point data corresponding to the sample object; The stay information point determination module is used to obtain the target stay information point data corresponding to the target object based on the output of the target cyclic network model.

9. An electronic device, characterized in that: The electronic device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the motion trajectory processing method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the motion trajectory processing method as described in any one of claims 1 to 7 is implemented.

Images