CN117194391A

CN117194391A - Data processing method, device, equipment and computer readable storage medium

Info

Publication number: CN117194391A
Application number: CN202311480246.9A
Authority: CN
Inventors: 杨远航; 张洋; 冯晨昊
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2023-11-08
Filing date: 2023-11-08
Publication date: 2023-12-08
Anticipated expiration: 2043-11-08

Abstract

The application provides a data processing method, a device, equipment and a computer readable storage medium; the method is applied to the map field, and comprises the following steps: acquiring road data to be processed, and determining a plurality of initial boarding points in the road data; acquiring historical riding data, and determining a plurality of target boarding points based on the historical riding data and a plurality of initial boarding points; acquiring first position information of each target get-on point, and determining a get-on point name of each target get-on point; first location information and get-on point names of a plurality of target get-on points are stored. The method and the device can improve the accuracy and completeness of the target get-on point in the road data.

Description

Data processing method, device, equipment and computer readable storage medium

Technical Field

The present application relates to the field of computer networks, and in particular, to a data processing method, apparatus, device, and computer readable storage medium.

Background

People travel through internet vehicle reduction has become a common phenomenon, and with the rapid development of intelligent services, people travel more and more conveniently and efficiently. In order to enable the internet about motor vehicle driver to connect to the user in the shortest time, the positioning of the point of the user's get-on is a key technology.

The applicant finds in the specific implementation process that: in the actual process, the positioning can drift, so that the finally generated boarding point may not be on the road, the boarding point which is not on the road can cause misunderstanding of a user, and a driver does not know the specific position of the receiving passenger. In addition, for cities or areas not covered by the user orders or the taxi stopping points, no suitable boarding points are recommended to the user for selection, so that the meeting cost of the driver and the passenger is increased, and the number of cancelled orders caused by the fact that the user cannot find the taxi is increased.

Disclosure of Invention

The embodiment of the application provides a data processing method, a data processing device and a computer readable storage medium, which can improve the accuracy and completeness of a target get-on point in road data.

The technical scheme of the embodiment of the application is realized as follows:

the embodiment of the application provides a data processing method, which comprises the following steps:

acquiring road data to be processed, and determining a plurality of initial boarding points in the road data;

acquiring historical riding data, and determining a plurality of target boarding points based on the historical riding data and the plurality of initial boarding points;

acquiring first position information of each target get-on point, and determining a get-on point name of each target get-on point;

And storing the first position information and the get-on point names of the multiple target get-on points.

An embodiment of the present application provides a data processing apparatus, including:

the first determining module is used for acquiring road data to be processed and determining a plurality of initial boarding points in the road data;

the second determining module is used for acquiring historical riding data and determining a plurality of target boarding points based on the historical riding data and the plurality of initial boarding points;

the third determining module is used for obtaining the first position information of each target get-on point and determining the get-on point name of each target get-on point;

and the data storage module is used for storing the first position information and the get-on point names of the multiple target get-on points.

An embodiment of the present application provides an electronic device, including:

a memory for storing computer executable instructions;

and the processor is used for realizing the method provided by the embodiment of the application when executing the computer executable instructions stored in the memory.

The embodiment of the application provides a computer readable storage medium, which stores a computer program or computer executable instructions for implementing the data processing method provided by the embodiment of the application when being executed by a processor.

The embodiment of the application provides a computer program product, which comprises a computer program or a computer executable instruction, and the computer program or the computer executable instruction realize the data processing method provided by the embodiment of the application when being executed by a processor.

The embodiment of the application has the following beneficial effects:

in the embodiment of the application, after the road data to be processed is acquired, firstly, uniform dotting is carried out on the road data so as to obtain a plurality of initial boarding points, thus, complete boarding points can be obtained on each road, then, density adjustment is carried out on the plurality of initial boarding points based on historical boarding data, a plurality of target boarding points are determined, and the first position information and boarding names of each target boarding point are acquired, so that a final complete target boarding point set is obtained and stored. Because the target boarding point is determined from a plurality of initial boarding points through the historical boarding data, the accuracy of the determined target boarding point can be ensured, the determined target boarding point can be ensured to conform to the riding habit of a user, and the experience and effect of network taxi service can be improved, so that the viscosity of the user is improved.

Drawings

FIG. 1 is a schematic diagram of a data processing system according to an embodiment of the present application;

fig. 2 is a schematic structural diagram of a server according to an embodiment of the present application;

FIG. 3A is a schematic flow chart of an implementation of a data processing method according to an embodiment of the present application;

FIG. 3B is a schematic diagram of an implementation flow for determining a plurality of target boarding points according to an embodiment of the present application;

FIG. 3C is a schematic diagram of another implementation flow for determining multiple target get-on points according to an embodiment of the present application;

fig. 3D is a schematic implementation flow chart of determining a get-on roll name of each target get-on roll point according to an embodiment of the present application;

fig. 3E is a schematic flowchart of an implementation process for obtaining a plurality of visible interest points in road data according to an embodiment of the present application;

FIG. 3F is a flowchart illustrating another implementation of the data processing method according to an embodiment of the present application;

FIG. 3G is a schematic diagram of an implementation flow for determining a recommended get-on point from a plurality of target get-on points according to an embodiment of the present application;

fig. 3H is a schematic implementation flow diagram of a correspondence between a generation area and a get-on point according to an embodiment of the present application;

FIG. 4 is an interface schematic diagram of a network vehicle application according to an embodiment of the present application;

FIG. 5 is a schematic flow chart of another implementation of the data processing method according to the embodiment of the present application;

FIG. 6 is a diagram illustrating the generation of target get-on points using the number of times the net-jockey trajectory starts;

FIG. 7 is a textCNN network architecture diagram;

FIG. 8 is a crowd-sourced picture of a quick hotel and home dish;

FIG. 9 is a schematic view of an interface when driving in a building;

fig. 10 is a schematic diagram of a terminal interface when a terminal is driving in a T3 terminal building at an airport.

Detailed Description

The present application will be further described in detail with reference to the accompanying drawings, for the purpose of making the objects, technical solutions and advantages of the present application more apparent, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.

In the following description, the terms "first", "second", "third" and the like are merely used to distinguish similar objects and do not represent a specific ordering of the objects, it being understood that the "first", "second", "third" may be interchanged with a specific order or sequence, as permitted, to enable embodiments of the application described herein to be practiced otherwise than as illustrated or described herein.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the embodiments of the application is for the purpose of describing embodiments of the application only and is not intended to be limiting of the application.

Before describing embodiments of the present application in further detail, the terms and terminology involved in the embodiments of the present application will be described, and the terms and terminology involved in the embodiments of the present application will be used in the following explanation.

1) And (5) loading: when passengers use riding service, the riding-appointed boarding sites are generally recommended to the passengers by a taxi-taking platform for selection. The boarding point can be a fixed place such as an airport, a railway station, etc., or a specific position on the street. Determination of the point of entry generally requires consideration of traffic convenience and safety.

2) Driving in a street scene: refers to the act of passengers looking for taxis on the street and taking them. In a street scene taxi taking, passengers can actively indicate a taxi driver to stop and take a load, and temporary traffic conventions are carried out in a gesture or hand lifting mode. The user can drive on the road side, and the recommendation of the easy-to-drive on point which is slightly closer to the user is more suitable

3) Regional scene taxi taking: in a certain area, a passenger requests a taxi taking service through a mobile phone application software or a call center and the like. In regional scene taxi taking, the passenger can select a destination and pick up the passenger by the system by distributing the appropriate vehicle. For example, the user is in a district, office building or mall etc. area. If the user is in a residential community, it is appropriate to recommend pick-up points associated with north-south doors.

4) Special station scene driving: meaning that at a particular station or stop, passengers can conveniently find a taxi or share a vehicle. Special stations are airport railway stations, and have fixed network bus boarding and disembarking points, which are defined by relevant scenes and are generally not changed at will.

5) Point of interest (Point of Interest, POI): commonly referred to as points of interest, refer to geographic locations identified in a map or navigation system that have a particular meaning. POIs generally refer to point-like data in an Internet electronic map, such as shops, restaurants and the like, and are convenient for users to quickly locate and search.

6) Optical character recognition technology (Optical Character Recognition, OCR): the optical character recognition technology is a technology for recognizing characters in an image and outputting recognized text information. OCR technology recognizes and extracts text information by scanning and analyzing characters in an image and converts it into a computer-processable text format for subsequent data analysis, storage, and application. OCR technology can be used for identifying information such as surrounding contents of a get-on point in a taxi taking scene.

7) Text convolutional neural network (Text Convolutional Neural Network, textCNN): a deep learning model is commonly used for natural language processing tasks such as text classification, emotion analysis and the like. TextCNN networks consist mainly of a convolutional layer and a pooling layer. The convolution layer carries out convolution operation on the text through the sliding window, extracts features with different lengths, and is similar to carrying out feature extraction on images in image convolution. In TextCNN networks, the convolution layer typically convolves with a plurality of convolution kernels of different sizes to obtain feature information of different sizes. And the pooling layer reduces the dimension of the output of the convolution layer, and the maximum value or the average value of the output of the convolution kernel is generally obtained by adopting modes such as maximum pooling or average pooling, so that the characteristic representation of the output of the convolution layer is more stable, and the calculated amount is reduced. Furthermore, the TextCNN network also contains a full connectivity layer and a softmax layer for classification or prediction tasks.

In the prior art, the generation of the on-board point data is typically based on analysis of user order logs and taxi stops. However, this method has some technical problems, resulting in some adverse effects. For example, the positioning device may drift for a variety of reasons (e.g., signal interference, building blockage, etc.), resulting in a potential inaccuracy in the location of the last generated pick-up point. If the boarding point is not on the road, misunderstanding is caused to the user, and the driver cannot accurately find the specific position of the passenger. This may lead to an increase in passenger waiting time, increasing service dissatisfaction. In addition, for some cities or areas, there is an inability to generate pick-up point data for recommendation to the user's choice due to a lack of sufficient user order or taxi-stop data. The cold start problem can affect user experience, and the user cannot enjoy personalized get-on point recommendation service, and the user needs to select the get-on point by himself, so that the operation complexity and uncertainty of the user are increased. In addition, the user may encounter situations where the point of the pick-up is wrong, resulting in waiting for a vehicle, getting lost, or getting on the wrong location. This can be inconvenient and unsatisfactory for the user, affecting the travel experience. Meanwhile, drivers can suffer from the trouble that accurate boarding points cannot be found, and the time and cost for receiving passengers are increased.

In order to solve the problems, the embodiment of the application provides a data processing method. The method comprises the steps of determining initial getting-on points by utilizing road data and track heat, performing density adjustment by combining network vehicle track data, generating a target getting-on point set, and assigning names of visible interest points to the target getting-on points by utilizing visible interest points in a search range generated by dotting of the road data. Compared with the prior art, the method realizes high-precision generation and intelligent recommendation of the target boarding point by utilizing the technical means of the road data, the track heat and the density optimization algorithm, brings more convenient and satisfactory boarding experience to users, and reduces the time and cost of taxis and network taxi-booking industries.

The embodiment of the application provides a data processing method, a device, equipment, a computer readable storage medium and a computer program product, which can improve the accuracy of recommended get-on position information.

The following describes exemplary applications of the electronic device provided by the embodiments of the present application, where the device provided by the embodiments of the present application may be implemented as a notebook computer, a tablet computer, a desktop computer, a set-top box, a mobile device (for example, a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, a portable game device), a smart phone, a smart speaker, a smart watch, a smart television, a vehicle-mounted terminal, and other various types of user terminals, and may also be implemented as a server. In the following, an exemplary application when the device is implemented as a server will be described.

With reference to fig. 1, fig. 1 is a schematic architecture diagram of a data processing system 100 according to an embodiment of the present application, in order to support an exemplary application, a terminal (a terminal 200-1 and a terminal 200-2 are shown as an example) are connected to a server 400 through a network 300, where the network 300 may be a wide area network or a local area network, or a combination of the two. The data processing system 100 may further include a database 500 for storing road data, crowd-sourced pictures, historical driving data, first location of a target driving point, driving point name, and the like, where the database 500 may be independent of the server 400 or may be located inside the server 400, and in fig. 1, the database 500 is illustrated as being independent of the server 400.

The server 400 receives a riding order transmitted from a plurality of terminals (the terminals 200-1 and 200-2 are exemplarily shown) and acquires riding data including information of a boarding point, a alighting point, and a trip cost, and stores the riding data in the database 500. When it is required to determine a target boarding point in the road network, the server 400 acquires the road data to be processed from the database 500, and determines a plurality of initial boarding points in the road data, wherein the distances between two adjacent initial boarding points are the same. Then, the server 400 acquires the historical boarding data and determines a plurality of target boarding points based on the historical boarding data and the plurality of initial boarding points. Next, the server 400 acquires first location information of each target get-on spot, and determines a get-on spot name of each target get-on spot. Finally, the server 400 stores the first location information of the plurality of target get-on points and the get-on roll names to the database 500.

Illustratively, when the terminal 200-1 starts the network taxi-taking application or applet and determines a taxi-taking start point, the server 400 acquires the taxi-taking start point, recommends a suitable taxi-taking point for the user according to the taxi-taking start point and the determined target taxi-taking point, and feeds back the result to the terminal 200-1. Therefore, the communication cost and the collision cost between passengers and drivers can be reduced, the order cancellation rate can be reduced, the viscosity of a user can be improved, and the quality of travel service can be improved.

In some embodiments, the server 400 may be a stand-alone physical server, a server cluster or a distributed system formed by a plurality of physical servers, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDNs), and basic cloud computing services such as big data and artificial intelligence platforms. The terminals 200-1 and 200-2 may be, but are not limited to, smart phones, tablet computers, notebook computers, desktop computers, smart speakers, smart watches, car terminals, etc. The terminal and the server may be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiment of the present application.

Referring to fig. 2, fig. 2 is a schematic structural diagram of a server 400 according to an embodiment of the present application, and the server 400 shown in fig. 2 includes: at least one processor 410, a memory 450, at least one network interface 420, and a user interface 430. The various components in server 400 are coupled together by bus system 440. It is understood that the bus system 440 is used to enable connected communication between these components. The bus system 440 includes a power bus, a control bus, and a status signal bus in addition to the data bus. But for clarity of illustration the various buses are labeled in fig. 2 as bus system 440.

The processor 410 may be an integrated circuit chip having signal processing capabilities such as a general purpose processor, such as a microprocessor or any conventional processor, a digital signal processor (Digital Signal Processor, DSP), or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like.

The user interface 430 includes one or more output devices 431, including one or more speakers and/or one or more visual displays, that enable presentation of the media content. The user interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

Memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard drives, optical drives, and the like. Memory 450 optionally includes one or more storage devices physically remote from processor 410.

Memory 450 includes volatile memory or nonvolatile memory, and may also include both volatile and nonvolatile memory. The non-volatile Memory may be a Read Only Memory (ROM), and the volatile Memory may be a random access Memory (Random Access Memory, RAM). The memory 450 described in embodiments of the present application is intended to comprise any suitable type of memory.

In some embodiments, memory 450 is capable of storing data to support various operations, examples of which include programs, modules and data structures, or subsets or supersets thereof, as exemplified below.

An operating system 451 including system programs, e.g., framework layer, core library layer, driver layer, etc., for handling various basic system services and performing hardware-related tasks, for implementing various basic services and handling hardware-based tasks;

a network communication module 452 for accessing other electronic devices via one or more (wired or wireless) network interfaces 420, the exemplary network interface 420 comprising: bluetooth, wireless compatibility authentication (WiFi), and universal serial bus (Universal Serial Bus, USB), etc.;

A presentation module 453 for enabling presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 (e.g., a display screen, speakers, etc.) associated with the user interface 430;

an input processing module 454 for detecting one or more user inputs or interactions from one of the one or more input devices 432 and translating the detected inputs or interactions.

In some embodiments, the apparatus provided by the embodiments of the present application may be implemented in software, and fig. 2 shows a data processing apparatus 455 stored in a memory 450, which may be in the form of a program, a plug-in, or the like, including the following software modules: the first determining module 4551, the second determining module 4552, the third determining module 4553 and the data storage module 4554 are logical, and thus may be arbitrarily combined or further split according to the functions implemented. The functions of the respective modules will be described hereinafter.

In other embodiments, the apparatus provided by the embodiments of the present application may be implemented in hardware, and by way of example, the apparatus provided by the embodiments of the present application may be a processor in the form of a hardware decoding processor that is programmed to perform the data processing method provided by the embodiments of the present application, e.g., the processor in the form of a hardware decoding processor may employ one or more application specific integrated circuits (Application Specific Integrated Circuit, ASIC), digital signal processors (Digital Signal Processor, DSP), programmable logic devices (ProgrammableLogic Device, PLD), complex programmable logic devices (Complex Programmable Logic Device, CPLD), field programmable gate arrays (Field-Programmable Gate Array, FPGA), or other electronic components.

Next, the data processing method provided by the embodiment of the present application is described, and as mentioned above, the electronic device implementing the data processing method of the embodiment of the present application may be a server. The execution subject of the respective steps will not be repeated hereinafter. Referring to fig. 3A, fig. 3A is a schematic flow chart of an implementation of a data processing method according to an embodiment of the present application, which will be described with reference to the steps shown in fig. 3A, where the main body of the steps in fig. 3A is a server.

In step 101, road data to be processed is acquired, and a plurality of initial pick-up points in the road data are determined.

Wherein the distance between two adjacent initial boarding points may be the same. In some embodiments, the road data is obtained from a road network, and the specific implementation may obtain the road data based on a network map service, where the network map service provides a complete map and includes various information, such as roads, points of interest, buildings, traffic restrictions, and the like. Therefore, through these services, road data including road names, start and end positions, lengths, link topology relations, and the like are acquired using their API interfaces. Road data may also be obtained based on open source road network data, using open source road network data, either public or personally developed, which data sets primarily include map data, which data is accessed through associated tools or APIs, including road attributes, route planning, and the like.

The determining of the plurality of initial getting-on points in the road data may be obtained by uniformly dotting each road edge included in the road data, that is, sampling is performed on each road edge at intervals of a preset distance, where the preset distance may be 5 meters, 8 meters, 10 meters, and the like, to obtain one initial getting-on point. The distance between two adjacent initial boarding points is the same.

In step 102, historical ride data is obtained and a plurality of target ride points are determined based on the historical ride data and the plurality of initial ride points.

In some embodiments, the historical ride data includes user taxi taking behavior data, net taxi or taxi track start point data. The travel information of the user can be obtained from the record of the specific mobile application program, wherein the travel information comprises the coordinates of the boarding location for driving, and the travel information of the user under different traffic modes can also be obtained from a traffic system, and mainly comprises track starting point data.

In some embodiments, referring to fig. 3B, fig. 3B is a schematic flow chart of one implementation of determining a plurality of on-target points according to an embodiment of the present application, and step 102 shown in fig. 3A may be implemented through steps 1021A to 10210A shown in fig. 3B, which is described in detail below in connection with fig. 3B.

Step 1021A: and acquiring a historical riding starting point from the historical riding data.

In some embodiments, starting point location information for a user's past drive is extracted by parsing historical drive data. And storing the historical riding data of the user in the cloud under the condition that the user allows. When the historical riding starting point needs to be acquired, the corresponding riding data storage position is accessed, and relevant information is read. Then, the read historical riding data is analyzed to obtain the starting point position information of each riding record. The analyzed data can be used for extracting the starting point position of each riding, and the starting point position can comprise longitude and latitude coordinates, geographical position description and the like.

Step 1022A: and determining a boarding area of the initial boarding point.

In some embodiments, when determining the boarding area of the initial boarding point, first acquiring fourth position information of the initial boarding point, and then determining the boarding area of the initial boarding point by taking the initial boarding point as an area center according to the fourth position information of the initial boarding point, wherein the boarding area can be a circular area taking the initial boarding point as a center, a preset distance as a radius, a square area taking the initial boarding point as the center, a preset distance as a side length, or a rectangular area taking the initial boarding point as the center. The boarding areas of the different initial boarding points may be free of overlapping areas.

Step 1023A: and judging whether the boarding area of the initial boarding point is a parking area or not.

In some embodiments, after determining the boarding area to which each initial boarding point belongs, it is further necessary to determine whether the area is a no-parking or limited-parking area, and if no parking is possible in the area, the location is not suitable as the boarding point. And determining whether the boarding area is a parking-restricted area or not according to the attribute information of the road section where the boarding area is located, wherein the boarding area comprises whether the boarding area is defined as a parking-forbidden area or not, whether a taxi or a taxi in a network is allowed to enter or not, and the like. And matching the information of the boarding area with map data, and judging whether the boarding area is a limited parking area according to the data matching result. If it is a restricted parking area, then other suitable initial pick-up points need to be found.

When the boarding area of the initial boarding point is not the restricted parking area, step 1024A is entered, and when the boarding area of the initial boarding point is the restricted parking area, step 10210A is entered.

Step 1024A: and judging whether the boarding area comprises a historical boarding start point or not.

In some embodiments, when the boarding area of the initial boarding point is not limited to the parking area, it is indicated that the boarding area where the initial boarding point is located may be parked, and thus may be used as a boarding start point or a riding end point. In order to determine whether the initial boarding point can be determined as the target boarding point at this time, it is necessary to further determine whether the boarding area contains the starting point of the history of boarding, that is, whether the user is boarding in the area. When judging whether the history riding starting point is included in the boarding area, the position information of each history riding starting point can be obtained from the history riding data, whether the history riding starting point is in the boarding area or not is determined based on the position information of each history riding starting point, namely whether the history riding starting point is included in the boarding area is determined, if the history riding starting point is included in the boarding area, the boarding point which is used by a user when traveling is described, and the initial boarding point can be considered as a target boarding point.

Step 1025A is entered when the history starting point is included in the boarding area, and step 1026A is entered when the history starting point is not included in the boarding area.

Step 1025A: and determining the initial boarding point as a target boarding point.

In some embodiments, when the boarding area of the initial boarding point is not limited to the parking area, it is further determined whether the area includes a historical boarding start point. If the area includes a historical departure point, the initial pick-up point is determined to be the target pick-up point.

Step 1026A: a minimum distance between the initial pick-up point and the determined target pick-up point is determined.

In some embodiments, if there is no historical driving point in the driving area corresponding to the initial driving point, which indicates that no user drives in the driving area, it is necessary to determine whether to reserve the initial driving point according to the minimum distance between the initial driving point and the determined target driving point. To obtain the distance between the initial get-on point and the target get-on point, this may be achieved by calculating a straight line distance between them or a distance of an actual travelable path. By comparing all calculated distances, the smallest of all distances can be automatically selected as the smallest distance between the initial pick-up point and the target pick-up point.

Step 1027A: it is determined whether the minimum distance is less than a first distance threshold.

In some embodiments, after determining the minimum distance between the initial pick-up point and the target pick-up point, it is determined whether the minimum distance is less than a set first distance threshold, wherein step 1028A is entered when the minimum distance is less than the first distance threshold and step 1029A is entered when the minimum distance is greater than or equal to the first distance threshold.

Illustratively, the first distance threshold may be 50 meters, 60 meters, 100 meters, etc.

Step 1028A: and deleting the initial boarding point.

In some embodiments, when the minimum distance is less than the first distance threshold, it is indicated that the distance between the initial entry point and the determined target entry point is sufficiently close, and the initial entry point is deleted because there is no historical entry point in the entry region corresponding to the initial entry point, and therefore the initial entry point is not required to be added as the target entry point.

Step 1029A: and determining the initial boarding point as a target boarding point.

In some embodiments, when the minimum distance between the initial pick-up point and the determined target pick-up point is greater than or equal to the set first distance threshold, it means that the distance between them has exceeded the preset threshold. In this case, the initial entry point will be determined as the target entry point. For example, if the distance between two target pick-up points does not exceed 50m and the actual distance exceeds 50m, then the initial pick-up point 50m from the first target pick-up point would be selected as the second target pick-up point. Even if there is no historical driving point in the area, if the distance between the initial driving point and the determined target driving point reaches the set distance threshold, the initial driving point is used as the target driving point, so that the future user is prevented from being recommended to a distant driving point when driving in the area. Through the processing mode, more reasonable selection of the get-on points can be ensured, and user experience and service accuracy are improved.

Step 10210A: and deleting the initial boarding point.

In some embodiments, when the boarding area of the initial boarding point is determined to be a parking-restricted area, that means an area where parking is not allowed or access to a net taxi or a taxi is not allowed. In this case, a deletion operation is performed, that is, an initial boarding point located in a limited parking area is to be deleted, to avoid that a service request transmitted to a network-bound vehicle is an impossible parking area. For example, in the context of airports, train stations, etc., there are often officially designated boarding points that have been approved and that meet the regulations of traffic management. Therefore, in scenes such as airports, train stations, etc., a boarding point designated by a official is determined as a target boarding point to ensure compliance and smoothness of service.

In the steps 1021A to 10210A, the density adjustment is performed by calculating specific distance and area information according to the history data, so that a large amount of data generated by the equal density dotting can be reduced, and the storage cost of the data can be reduced. By reserving high-density data of the area where the user frequently drives and performing density reduction processing on the area where the user does not frequently drive or does not drive, the storage requirement can be effectively reduced under the condition that the completeness of the target get-on point is met. In addition, the boarding points can be distributed more reasonably in space through density adjustment. For example, the density of points on roads near commercial areas is higher and the density of points on roads in remote mountainous areas is lower. The density adjustment is performed by counting the network vehicle track starting point data, so that the distribution of the vehicle loading points can be more in line with the actual demands and the habits of users.

In some embodiments, step 102 may also be implemented through steps 1021B to 1028B shown in fig. 3C, and fig. 3C is a schematic flow chart of another implementation of determining a plurality of target boarding points according to an embodiment of the present application, which is specifically described below with reference to fig. 3C.

Step 1021B: and acquiring the historical riding starting points included in each road section from the historical riding data.

In some embodiments, a plurality of road segments are first divided based on the road data, for example, one road segment may be divided every 100 meters, and a segment start point and a segment end point of the road segment are determined. And then acquiring a plurality of historical riding starting points from the historical riding data, and determining the historical riding starting points included in each road section based on the starting point position information (which can be longitude and latitude) of each historical riding starting point, the section starting point and the section ending point of each road section, and providing necessary data information for the subsequent determination of the on-destination riding point.

Step 1022B: it is determined whether at least one historical departure point is included in the road segment.

When the road section does not include the historical riding starting point, indicating that no user gets on the road section, and then entering step 1023B; when at least one historical departure point is included in the road segment, step 1024B is entered.

In step 1023B, a preset number of initial pick-up points are selected from a plurality of initial pick-up points included in the road segment as target pick-up points.

In some embodiments, if the historical departure point is not included in the road segment, it is indicated that no user is getting on the road segment. At this time, a preset number of initial boarding points need to be selected as target boarding points from a plurality of initial boarding points included in the road section. The preset number is an integer greater than or equal to 1, for example, the preset number may be 1, or may be 2 or 3. When a preset number of initial boarding points are selected from a plurality of initial boarding points included in the road section to serve as target boarding points, the preset number of initial boarding points can be determined randomly from the plurality of initial boarding points to serve as target boarding points, the preset number of initial boarding points can be determined from the plurality of initial boarding points according to preset rules, and the determined preset number of initial boarding points can be determined to serve as target boarding points.

For example, the preset number is 1, and one road section includes 10 initial boarding points, then one initial boarding point is randomly selected from the 10 initial boarding points to serve as a target boarding point, or the 5 th initial boarding point in the 10 initial boarding points serves as a target boarding point.

In step 1024B, it is determined whether the number of occurrences of the historical departure point included in the road segment is greater than a preset number threshold.

In some embodiments, when determining a history starting point included in a road section, a history starting point set is created for each road section, the history starting point set having an initial value of null, when determining that a history starting point a is located in the road section, it is first determined whether the history starting point a exists in the history starting point set, if the history starting point a exists in the history starting point set, the number of occurrences corresponding to the history starting point a is added by 1, if the history starting point a does not exist in the history starting point set, the history starting point a is added to the history starting point set, and the number of occurrences corresponding to the history starting point a is 1, so that the history starting point a included in each road section and the number of occurrences corresponding to the history starting point a are obtained by traversing the history starting data, and in this step, it is necessary to sum the number of occurrences corresponding to each of the history starting points included in the road section, and then determine whether the number of occurrences obtained by summation is greater than a preset number of threshold.

Here, the number of times threshold may be 100, 150, 200, etc., when the road section includes at least one historical driving start point and the total number of occurrences of the historical driving start point is greater than the preset number of times threshold, it indicates that the number of times of calling in the road section is greater, that is, the heat of calling in the road section is higher, and step 1025B is performed at this time; when at least one historical driving source is included in the road section and the number of occurrences of the historical driving source is less than or equal to the threshold number of occurrences, it is indicated that the user is calling the road section but the number of occurrences is relatively small, and step 1028B is entered.

In step 1025B, the historical driving start points included in the road segments are clustered to obtain at least one cluster point.

Clustering is an unsupervised learning method that groups data points by calculating the similarity between the data points, and the clustering result is the center point of each cluster and the data points that the cluster contains. In some embodiments, a preset clustering algorithm may be used to perform clustering on historical car starts to obtain a plurality of clusters, where each cluster represents a possible target get-on point, and a center point of the cluster is used as a cluster point. For example, a historical ride starting point included within a road segment may be clustered using a K-1 Means (K-Means) algorithm, a Fuzzy C-Means clustering (FCM) algorithm.

Step 1026B: and judging whether a target cluster point exists in the plurality of cluster points.

The nearest distance between the target cluster point and the initial get-on point in the road section is greater than a third distance threshold. When there is a target cluster point in the plurality of cluster points, it is indicated that there are a plurality of starting points gathered together at non-initial boarding points in the historical boarding data in the road section, step 1027B is entered at this time, and when there is no target cluster point in the plurality of cluster points, it is indicated that there is no apparent starting point dense region in the historical boarding data in the road section, and there is no specific hot boarding point on the section, step 1028B is entered at this time.

In step 1027B, the initial get-on point and the target cluster point within the road segment are determined as target get-on points.

In the clustering process, a cluster having a higher density, i.e., a target cluster point, is found, which can be determined by calculating the number or density of starting points in each cluster. And then, matching the initial get-on point in the road section with the target clustering point to determine the target get-on point. For each initial get-on point, calculating the distance between the initial get-on point and the target clustering point, and then selecting the target clustering point with the distance larger than a third distance threshold as the target get-on point of the get-on point.

In step 1028B, the initial pick-up point within the road segment is determined as the target pick-up point.

When at least one historical driving starting point is included in the road section and the occurrence number of the historical driving starting point is less than or equal to the threshold number, the user is stated to call the vehicle on the road section, but the number of times is relatively small, and the initial driving point is selected from a plurality of initial driving points included in the road section to serve as the target driving point. These initial pick-up points are determined in advance according to other rules or algorithms. When the target cluster point does not exist in the plurality of cluster points, the fact that a starting point dense area is not obvious in the historical driving data in the road section is indicated, no specific hot driving point exists in the road section, and the initial driving point is selected from the plurality of initial driving points included in the road section to serve as the target driving point.

In some embodiments, it is determined whether a historical departure point is included in the road segment based on the historical departure data. If the historical riding start point is not included, a preset number of initial boarding points are selected as target boarding points. If at least one historical riding starting point is included, whether the occurrence number of the historical riding starting point in the road section is larger than a preset number threshold is further judged. If the total occurrence times of the historical riding start points are larger than a preset time threshold, indicating that the calling heat of the road section is higher, and performing clustering processing to obtain at least one clustering point. If the occurrence number of the historical riding start point is smaller than or equal to the threshold number of times, the condition that the user calls the car on the road section is indicated, but the number of times is smaller, and the initial boarding point is determined to be the target boarding point. And then judging whether a target cluster point exists in the plurality of cluster points, namely whether the nearest distance between the target cluster point and the initial get-on point in the road section is larger than a third distance threshold value. If the target clustering point exists, indicating that a plurality of aggregated starting points exist in the historical driving data in the road section, and determining the initial driving point and the target clustering point as the target driving point. And if the target cluster point does not exist, determining the initial get-on point as the target get-on point. The target clustering points are determined by clustering historical riding starting point data to obtain a plurality of clustering points, and then determining the target clustering points by judging whether the target clustering points with the nearest distance larger than a third distance threshold exist. The third distance threshold is used for judging whether the nearest distance between the target clustering point and the initial get-on point in the road section is larger than the threshold, so that whether the target clustering point exists is determined. The preset frequency threshold is used for judging whether the occurrence frequency of the historical riding start point is larger than the threshold, so as to determine the heat of calling the road section and whether clustering is carried out.

In the above steps 1021B to 1023B, the distribution of the user traveling can be better understood by acquiring the historical riding start point in the road section from the historical riding data. And secondly, setting a threshold according to the total occurrence times of the historical riding starting points, screening out road sections with higher calling frequency, and preferentially considering the road sections as candidate areas of target riding points to improve the matching accuracy. Then, the historical riding starting points are clustered, so that the starting points with a relatively close distance can be classified into the same clustering point, and the selection process of the target riding points is further simplified. And finally, by judging the distance between the clustering point and the initial boarding point, boarding places which are far away from the initial boarding point and have more historical boarding points are increased, and the convenience of service and the satisfaction of users are improved. For example, for a road segment of 100 meters, there will be 10 initial pick-up points for the segment, with the initial pick-up points divided at 10 meter intervals. Assuming that a plurality of office buildings are arranged around the road section, the demand for calling the bus is very high, and the number of historical bus taking starting points is also large. Wherein, a certain historical riding starting point appears 1000 times, which indicates that the calling frequency of the road section is very high. When clustering the 10 historical riding starts, it was found that 3, 3 and 4 historical riding starts are very close to each other and are clustered into three points. However, the closest distance between these three cluster points and the initial pick-up point exceeds 5 meters. Therefore, the three clustering points are used as additional target get-on points, so that the three target get-on points are added on the road section besides the preset initial get-on points.

In step 103, first location information of each target get-on point is obtained, and a get-on point name of each target get-on point is determined.

In some embodiments, since each target get-on point is selected from the initial get-on points, when the road data is uniformly dotted to obtain the initial get-on points, that is, the position information of each initial get-on point is determined, and after the target get-on point is determined, the first position information of the target get-on point can be directly obtained. The name of the target pick-up point may be the name of a shop, a building, a district, a school, etc. located at the target pick-up point.

In some embodiments, referring to fig. 3D, fig. 3D is a schematic flowchart of implementation of determining a get-on point name of each target get-on point according to an embodiment of the present application, and step 103 shown in fig. 3A may be implemented through steps 1031 to 1036 shown in fig. 3D, which is specifically described below in connection with fig. 3D.

Step 1031: and acquiring the interest point name of the visible interest point and the second position information of the visible interest point in the road data.

In some embodiments, a dataset containing road information is obtained, the dataset containing geometric shapes of individual roads and point of interest data associated therewith. And screening out the interest points visible in the road data according to a certain rule or algorithm. These visible points of interest refer to points of interest that have a certain degree of recognition on the road and that can be captured in time by the user, such as malls, gas stations, etc. In addition to visible points of interest, collectively referred to as invisible points of interest, an invisible point of interest refers to a point of interest that a user cannot see or capture on a road or map, such as a user at a floor in a building, etc. And extracting the interest point names of the visible interest points from the road data. These point-of-interest names are predefined or based on an identification derived by identifying crowd-sourced pictures, which are used to describe points of interest at a particular location. For each visible point of interest, its second location information is determined. The second location information includes latitude and longitude coordinates of the visible interest point, and the second location information may also include a description of the visible interest point relative to a certain reference point, where the reference point may be a building with a relatively obvious surrounding of the visible interest point, for example, location information used for supplementing the interest point before the exit of a building B.

Step 1032: and determining the nearest visible interest point corresponding to the on-target point based on the first position information of the on-target point and the second position information of the visible interest point.

In some embodiments, determining the nearest visible point of interest corresponding to the target on-board point based on the first location information of the target on-board point and the second location information of the visible point of interest may be by comparing the first location information of the target on-board point with the second location information of the visible point of interest using a location calculation algorithm to find the visible point of interest nearest to the target on-board point. And determining the nearest visible interest point corresponding to the on-target point, namely the visible interest point nearest to the on-target point. This target on-board point has a name or other identifying information associated with the nearest visible point of interest.

Step 1033: and judging whether the first distance between the nearest visible interest point and the on-board target point is smaller than a second distance threshold value.

In some embodiments, determining whether the first distance between the nearest visible point of interest and the target on-board point is less than the second distance threshold may be by calculating the first distance between the nearest visible point of interest and the target on-board point, which may be a straight line distance or other distance metric. The first distance represents the actual physical distance between two points, i.e. their straight-line distance on the map. During this time, a second distance threshold is set, which is a predefined value. The second distance threshold represents a maximum tolerated distance between the nearest visible point of interest and the point on the target.

Wherein step 1034 is entered when the first distance between the nearest visible point of interest and the target pick-up point is less than the second distance threshold, and step 1035 is entered when the first distance between the nearest visible point of interest and the target pick-up point is greater than or equal to the second distance threshold.

Step 1034: and determining the point-of-interest name of the nearest visible point-of-interest as the point-of-boarding name of the target point-of-boarding.

In some embodiments, when it is determined that the first distance between the nearest visible point of interest and the target on-board point is less than the second distance threshold, the point of interest name of the nearest visible point of interest is considered as the on-board point name of the target on-board point. Firstly, determining the nearest visible interest point corresponding to the on-target point by acquiring the interest point name of the visible interest point and the second position information of the visible interest point in the road data. And then, on the basis of judging whether the first distance between the nearest visible interest point and the target get-on point is smaller than a second distance threshold value, regarding the interest point name of the nearest visible interest point as the get-on point name of the target get-on point. A nearest visible point of interest is considered to be the same location when a first distance between it and a target pick-up point is less than a second distance threshold. Therefore, the point-of-interest name of the nearest visible point of interest is considered as the point-of-entry name of the target point-of-entry. And if the distances between the plurality of visible interest points and the target get-on point are the same and the first distance between the plurality of visible interest points and the target get-on point is smaller than the second distance threshold value, randomly regarding the interest point name of the nearest visible interest point as the get-on point name of the target get-on point.

Step 1035: and acquiring a default get-on point name.

In some embodiments, when it is determined that the first distance between the nearest visible interest point and the target get-on point is greater than or equal to the second distance threshold, it is indicated that the visible interest point is not around the target get-on point, and exceeds the range in which the target get-on point is located, and the default get-on point name is obtained as the get-on point name of the target get-on point. Illustratively, the second distance threshold may be 20 meters. In the embodiment of the application, a default get-on point name is preset for the case when the first distance between the nearest visible interest point and the target get-on point does not meet the requirement. For example, the default get-on point name may be a net-bound get-on point.

Step 1036: and determining the default get-on point name as the get-on point name of the target get-on point.

In some embodiments, when the first distance between the nearest visible point of interest and the target pick-up point is greater than or equal to the second distance threshold, a pick-up roll name that defaults to the target pick-up point is determined. And comparing the values of the nearest visible interest point and the target on-board point when judging whether the first distance between the two distances meets the requirement. If the first distance between the nearest visible interest point and the target get-on point is greater than or equal to the second distance threshold, the current nearest visible interest point is determined to be unsatisfactory, and cannot be the get-on point of the target get-on point. In this case, the default get-on roll name is determined as the get-on roll name of the target get-on roll. And associating the default get-on point name with the target get-on point by using a preset rule or algorithm, so that the default get-on point name becomes the mark of the target get-on point. The default roll-up name is determined when the first distance between the nearest visible interest point and the target roll-up point is greater than or equal to the second distance threshold, i.e., the default roll-up name is used when the nearest visible interest point meeting the requirements cannot be found, such as a net-bound roll-up point.

Referring to fig. 3E, fig. 3E is a schematic flowchart of an implementation of obtaining a plurality of visible interest points in road data according to an embodiment of the present application, and before step 1032 shown in fig. 3D, steps 1037 to 10310 may be further performed, which is specifically described below.

In step 1037, a trained visibility prediction model is obtained.

In some embodiments, a trained visibility prediction model needs to be obtained before determining the nearest visible point of interest corresponding to the target on-board point. The trained visibility prediction model is a neural network model, and can be a convolutional neural network model, a cyclic neural network model and a deep learning neural network model. The trained visibility prediction model is obtained by training a preset visibility prediction model by using training data and is used for predicting whether an interest roll name is visible or not. The training data includes a plurality of training point-of-interest names and labels for the respective training point-of-interest names, which may be visible or invisible. The names of the training interest points can be obtained from road data, can be obtained by carrying out image recognition on interest point images, can be obtained from a crowdsourcing image set of a map, and can be obtained by crawling from a network. After the training data is obtained, model training is carried out by using the training data and using a machine learning algorithm (such as TextCNN) to obtain a trained visibility prediction model.

In step 1038, a crowd-sourced picture set and gate address information for each point of interest included in the road data is obtained.

In some embodiments, it is desirable to obtain a road dataset containing crowdsourcing picture sets and individual point-of-interest gate information. The crowdsourcing picture set refers to a series of pictures containing sceneries in a target area, which are obtained in a crowdsourcing mode; the gate address information of the point of interest refers to a data set containing information such as the name, address, longitude and latitude of the point of interest, and the like, and is used for determining the nearest visible point of interest corresponding to the target on-board point. During training and testing, OCR is performed on the crowdsourcing picture set, text information in the picture is extracted and combined with geographical position information of the picture to form training data, and gate address information of the interest point can be obtained, so that the gate address information of the interest point is predicted by using a trained visibility prediction model in a subsequent step, and whether the interest point is visible or not is determined.

In step 1039, image recognition is performed on each crowd-sourced picture in the crowd-sourced picture set, so as to obtain a point of interest name included in each crowd-sourced picture.

In some embodiments, crowd-sourced pictures are identified by image recognition techniques to extract names of points of interest contained in the pictures. For example, image recognition may be performed using an image recognition model such as a convolutional neural network in a deep learning algorithm. When the trained image recognition model is used for carrying out image recognition on the crowdsourcing pictures, firstly, preprocessing is carried out on the crowdsourcing pictures to be recognized, wherein the operations comprise size normalization, graying, denoising, binarization and the like of the images. Then, by using image processing technology, such as edge detection, contour extraction and other methods, text areas in the image are found, so that the identification range is defined, and interference is reduced. And then dividing the positioned text area, and separating each character or text line so as to enable the subsequent recognition to be more accurate. And extracting the characteristics of each character or text line, helping the model understand the shape, texture and structure of the characters, and further carrying out correct classification. In the image recognition model application stage, the extracted characters or text lines are classified and recognized by using an image recognition model such as a convolutional neural network. The models are trained through a large number of character samples, and different characters can be accurately classified. And then, performing post-correction on the recognition result, including operations such as eliminating recognition errors, repairing broken words, merging text lines and the like. Finally, the identified text is converted into an editable text format, such as a text file, document, etc., for subsequent use and editing.

In step 10310, the gate address information and the interest point names of the respective interest points are predicted by using the trained visibility prediction model, so as to obtain a plurality of visible interest points in the road data.

In some embodiments, the predicting the gate address information and the names of the points of interest by using the trained visibility prediction model may be to predict whether each point of interest is visible or not and the gate address information and the names corresponding to the point of interest by using the trained visibility prediction model. For example, the trained visibility prediction model is a TextCNN network model that can be trained using artificially marked visible and invisible data. During prediction, the information and the name of the point of interest gate address are input into a trained visibility prediction model, so that the trained visibility prediction model is utilized to conduct visibility classification prediction on the point of interest, a visibility score is obtained, and whether the point of interest is visible or not is judged according to the visibility score. If the score is above a certain set threshold, the point of interest may be considered visible.

After each target get-on point is obtained in the steps 1031 to 10310, the door address information, the interest point name and the crowdsourcing picture identification result are put into the TextCNN network model to perform visibility classification prediction, so that whether the interest point is visible or not can be accurately judged, the accuracy and the reliability of the interest point in road data are improved, more accurate visible interest point information is provided for the subsequent get-on scene, and the accuracy of the user navigation to the target get-on point is improved.

Step 104: first location information and get-on point names of a plurality of target get-on points are stored.

In some embodiments, storing the first location information and the get-on point names of the plurality of target get-on points may be storing the location information and the corresponding get-on point names of the plurality of target get-on points to a local storage space of the server for subsequent use. In order to facilitate subsequent calling, the first position information and the boarding point names of the plurality of target boarding points are stored. In this way, in practical application, when the position information or the get-on point name of a specific target get-on point is required to be used, the position information or the get-on point name can be directly obtained from the stored data without repeated calculation or query.

In the data processing method provided by the embodiment of the application, after the road data to be processed are obtained, firstly, uniform dotting is carried out on the road data so as to obtain a plurality of initial boarding points, thus, complete boarding points can be obtained on each road, then, density adjustment is carried out on the plurality of initial boarding points based on historical boarding data, if a boarding area corresponding to the initial boarding points comprises a historical boarding start point, the initial boarding points are determined to be target boarding points, if the uploading area corresponding to the initial boarding points does not comprise historical boarding points, and the distance between the initial boarding points and the determined target boarding points is smaller than a first distance threshold, the initial boarding points are deleted, and if the distance is larger than or equal to the first distance threshold, the initial boarding points are also determined to be target boarding points, so that future users are prevented from being recommended to distant boarding points when the area is driven. Through the processing mode, more reasonable selection of the get-on points can be ensured, and user experience and service accuracy are improved. In addition, after a plurality of target get-on points are determined, the visible interest point which is closest to the target get-on point and has a distance smaller than a second distance threshold is determined as the name of the target get-on point, so that a user and a driver can timely determine the position of the target get-on point according to the name of the target get-on point, communication cost is reduced, and further experience and effect of network taxi service can be improved.

The data processing method of the get-on point provided by the embodiment of the application will be described in connection with the exemplary application and implementation of the combination of the server and the terminal provided by the embodiment of the application. Referring to fig. 3F, fig. 3F is a schematic flow chart illustrating still another implementation of the data processing method according to the embodiment of the present application, and will be described with reference to steps 201 to 205 shown in fig. 3F, where the server 400 receives a riding order sent by a plurality of terminals (the terminal 200-1 and the terminal 200-2 are exemplarily shown), and the terminal 200-1 is exemplified.

In step 201, the terminal 200-1 starts a network taxi taking application or applet.

In some embodiments, the terminal 200-1 initiates the network vehicle restraint application in response to a touch operation for the network vehicle restraint application icon. In addition, the terminal 200-1 may also start the network contract applet in response to a start operation for the network contract applet provided through an instant communication application, an online payment application, or the like.

In step 202, the terminal 200-1 selects a ride starting point.

In some embodiments, after the network taxi-taking application or applet is started, the terminal 200-1 presents a taxi-taking interface, presents a taxi-taking start point selection control and a taxi-taking end point selection control in the main page, acquires the current position of the terminal 200-1, determines the current position of the terminal 200-1 as a taxi-taking start point, and displays the current position in a display area corresponding to the taxi-taking start point selection control. In some embodiments, the user may be calling the other person, so the user may enter the ride start point information through the ride start point selection control and the terminal 200-1 displays the new ride start point.

In step 203, the server 400 receives third position information of the riding-start point transmitted from the terminal.

In some embodiments, the terminal 200-1 may transmit the third location information of the selected ride share to the server 400 through a notification message. The server 400 analyzes the notification message to obtain the third position information of the riding start point transmitted by the terminal. When receiving the third position information of the riding-start point transmitted from the terminal, the server 400 may receive the data transmitted from the terminal through a wireless communication technology such as a network or bluetooth. The specific implementation may vary depending on the communication protocols of the mobile phone and the terminal, and may be resolved in the application or system of the mobile phone. Regardless of the particular implementation, it is desirable to ensure that the server 400 is able to properly receive and parse the ride-on origin location information transmitted from the terminal 200-1.

In step 204, the server 400 determines a recommended get-on point from the plurality of target get-on points based on the third location information.

In some embodiments, referring to fig. 3G, fig. 3G is a schematic flowchart of an implementation process for determining a recommended get-on point from a plurality of target get-on points according to an embodiment of the present application, and step 204 shown in fig. 3F may be implemented through steps 2041 to 2044 shown in fig. 3G, which is described in detail below in connection with fig. 3G.

Step 2041: and judging whether the riding starting point is positioned in the road area or not.

In some embodiments, whether the ride starting point is located in the road area is determined by the third position information of the ride starting point, wherein when the ride starting point is determined to be located in the road area based on the third position information, it is indicated that the ride starting point is located at the roadside, step 2042 is performed at this time, and when it is determined that the ride starting point is not located in the road area but is located in the building area based on the third position information, step 2043 is performed.

Step 2042: and determining the target get-on point closest to the riding start point from the plurality of target get-on points as the recommended get-on point.

In some embodiments, the determining of the target get-on point closest to the departure point among the plurality of target get-on points as the recommended get-on point may be determining the closest recommended get-on point by calculating a distance between the departure point and each of the target get-on points among the plurality of target get-on points. The distance here may be a straight line distance or an actual distance of a walking path from a start point of riding to a point of getting on a target, because it may be necessary to walk on a road or walk to a road junction, or the like. Upon receiving third party location information for the ride start point, the distance between the ride start point and each target pick-up point is calculated using the information. By comparing the distances, a target boarding point closest to the riding start point is determined, and the target boarding point is determined as the recommended boarding point.

Step 2043: and acquiring a corresponding relation between the pre-generated area and the boarding point.

In some embodiments, the correspondence includes area identifiers of the areas and boarding points corresponding to the area identifiers, where the correspondence between the area identifiers and the boarding points may be one-to-one or one-to-many, that is, one area identifier may correspond to one boarding point or may correspond to multiple boarding points. Illustratively, the area identifier may be a cell, and the entry point corresponding to the area identifier may include a cell east gate, a cell west gate, a cell south gate, and the like. The area identifier may be a B building, and the entry point corresponding to the area identifier may be a B building exit.

In some embodiments, referring to fig. 3H, fig. 3H is a schematic flow chart illustrating the implementation of the correspondence between the generation area and the get-on point according to the embodiment of the present application, before step 2043, the server may generate the correspondence between the area and the get-on point through steps 401 to 402 shown in fig. 3H, and the following is specifically described with reference to fig. 3H.

Step 401: and acquiring riding order data sent by a plurality of terminals, and extracting the area and the boarding point of the terminal from the riding order data.

In some embodiments, the server 400 obtains the riding order data sent by the plurality of terminals, where the riding order data may be understood as historical riding order data, and the riding order data includes the location of the terminal when sending the riding order and information such as a boarding point, a alighting point, a vehicle identification, a riding cost and the like in the riding order. The server 400 analyzes the plurality of riding order data to obtain the position and the boarding point when the terminal sends the riding order, and if the position of the terminal when sending the riding order is determined to be in a building area (such as a district, a mall and an office building) at the moment, the position and the boarding point when the terminal sends the riding order are reserved; if the terminal is located in the road area according to the position of the terminal when the terminal sends the riding order, the area and the boarding point of the terminal when the terminal sends the riding order are not reserved.

Step 402: and clustering the area where the terminal is located and the corresponding get-on point to obtain the corresponding relation between the area and the get-on point.

In some embodiments, the area where the terminal is located and the corresponding boarding point are classified and grouped to form a corresponding relationship between the area and the boarding point. Firstly, obtaining riding order data sent by a plurality of terminals, and extracting information of an area where the terminals are located and a boarding point from the order data. These data may include terminal location coordinates, get-on point names, etc. And then classifying and grouping the area where the terminal is located and the corresponding boarding points by using a clustering algorithm. The clustering algorithm can divide the terminal into different clusters or groups according to the characteristics of the distance, the density and the like between the area where the terminal is located and the boarding point. After clustering, a plurality of clusters or groups are obtained, each cluster or group representing an area. In each region, a corresponding boarding point is included. And establishing and recording the corresponding relation between the area and the boarding point. By associating the pick-up point with the corresponding zone or by assigning an identifier to each zone. Through the steps, clustering is carried out on the area where the terminal is located and the corresponding get-on point, and the corresponding relation between the area and the get-on point is obtained.

Step 2044: based on the corresponding relation, getting on the get-on points corresponding to the building areas, and determining the get-on points corresponding to the building areas as recommended get-on points.

In some embodiments, when it is determined that the boarding start point of the passenger is located in the building area, an area identifier of the building area in which the boarding start point is located is obtained, then, according to the area identifier, a boarding point corresponding to the building area is obtained from a corresponding relationship between the area and the boarding point, and the boarding point corresponding to the building area is determined as the recommended boarding point. In addition, since one area identifier may correspond to a plurality of boarding points, in practical application, if a building area where the terminal is located corresponds to a plurality of boarding points, a boarding point closest to the terminal among the plurality of boarding points may be determined as a recommended boarding point, and a boarding point with the highest heat degree among the plurality of boarding points may be determined as a recommended boarding point, where the heat degree is the highest, that is, the number of times that the boarding point is selected as an actual boarding start point is the highest. If the user previously sent a ride order within the building area, the ride start point that was last selected by the user when sending a ride vertex within the building area may also be determined to be the recommended ride point.

Through the steps 2041 to 2044, when the server receives the third position information of the boarding start point sent by the terminal and determines that the boarding start point is located in the building area, the server queries according to the corresponding relationship between the pre-generated area and the boarding point. The pre-generated correspondence between the area and the boarding point already contains the boarding point corresponding to the building area, so that the boarding point corresponding to the building area can be obtained by inquiring the correspondence. For example, an identifier of a building area or other specific identifier may be used to find a record of a get-on point corresponding to the area in a corresponding relationship, and obtain first location information of the get-on point and a get-on point name from the record. Once the first position information and the get-on roll name of the get-on point corresponding to the building area are acquired, the get-on point can be used as a recommended get-on point and provided for a user to select a riding vehicle. In the process, the corresponding relation between the pre-generated area and the boarding point is required to be followed, and the boarding point corresponding to the building area is acquired according to the query result.

In step 205, the recommended get-on point is sent to the terminal to present the recommended get-on point in a display interface of the terminal.

In some embodiments, after determining the recommended get-on point, the server 400 sends the recommended get-on point to the terminal through the network, and presents the recommended get-on point on a display interface of the terminal, that is, displays the name of the recommended get-on point in a display area corresponding to the get-on start point selection control in the network owner interface of the terminal, and marks the recommended get-on point in a map display area, so that a user can more intuitively know the position of the recommended get-on point.

In the embodiment of the application, after the terminal starts the network taxi taking application or applet and selects the taxi taking starting point, the server can determine the recommended taxi taking point of the taxi taking time from the determined multiple target taxi taking points or the corresponding relation between the pre-established area and the taxi taking point based on the third position information after acquiring the third position information of the taxi taking starting point, and feeds back the recommended taxi taking point to the terminal, so that the server can recommend the recommended taxi taking point which is close to and easier to find for the user no matter the user is positioned in a road section, a building area or a special place (such as an airport, a railway station and the like), thereby ensuring that the taxi taking order can be successfully and efficiently completed, improving the success rate of taxi taking transactions, and improving the use convenience and experience of both sides of the taxi taking.

In the following, an exemplary application of the embodiment of the present application in a practical application scenario will be described.

The data processing method provided by the embodiment of the application can be applied to the online riding service in the map application and can also be applied to the network taxi-taking application program or the applet.

Referring to fig. 4, fig. 4 is an interface schematic diagram of a network vehicle application according to an embodiment of the present application. Interface diagram of taxi service provided in left map application, left side as in fig. 4Is shown in the interface diagram of (1) when the riding start point is selectedThe headquarter displays the recommended get-on point 401 and the names of the get-on point 401 in the left interface diagram, and displays the interface diagram on the right after determining the get-on start point and the get-on end point, and displays not only the get-on point but also the recommended travel path from the get-on start point to the get-on end point and the estimated get-on cost in the interface diagram.

Referring to fig. 5, fig. 5 is a schematic flow chart of another implementation of the data processing method according to the embodiment of the present application, and the steps shown in fig. 5 will be specifically described below.

Step 301: and obtaining road data, carrying out equal-density dotting on the road data, and determining an initial on-vehicle point set.

Here, an initial set of on-coming point positions is determined by acquiring road data and performing an equal-density break. However, there are some problems with such get-on points generated based on the clustering method, for example, the clustering position may not be on the road, or there are no get-on points in an area where the data is not covered. In order to solve the problems, the application adopts a way of dotting the road network, generates a get-on point set containing all areas, and ensures that the get-on points are all located on the road. By dotting the road network, the boarding points are all positioned on the road, and the nearby boarding points can be recommended to users in all areas.

However, merely performing the isodensity dotting results in a huge amount of data, thereby increasing the storage cost of the data. To address this problem, embodiments of the present application contemplate using a more intelligent method to generate the get-on point. For example, the road data is dotted by combining the track heat, and the get-on points can be selectively generated according to the actual travel demands and road conditions, so that unnecessary data redundancy is reduced. In addition, on the basis of user order data mining, the get-on point data can be further optimized and perfected. By analyzing the travel characteristics and preferences of the user and combining the regional conditions, the recommendation strategy of the boarding points can be further refined. For example, in a urban central area, it may be desirable to densely arrange boarding points, while in suburban areas or places with less traffic, it may be moderately sparsely arranged. Therefore, the actual requirements of users can be better met, and the service quality and the user satisfaction are improved.

Step 302: and acquiring network vehicle track starting point position data.

Step 303: and (5) performing density optimization on the initial on-board point set by utilizing the network vehicle track starting point position data.

Here, the starting point position data of the network vehicle track is acquired, and density adjustment is performed to generate a target vehicle-on point position set. However, only the equal density dotting results in huge data volume and high storage cost. Therefore, the embodiment of the application needs to use other data for tuning, such as behavior data and positioning data of the driving of the user.

Firstly, dotting is carried out by using a 5m interval, and point location reservation is carried out according to the condition that a user gets on a car within the 5m range. If 5m or the section of road has no user to get on the vehicle, the midpoint of the section of road is selected as a point location for reservation.

During the period, the embodiment of the application can carry out density optimization according to the historical behavior data and track data of the driving of the user, so as to reduce the data quantity and the storage cost. For the behavior that a user frequently drives in a certain area, denser point location dotting is carried out on the area, and for the area that the user rarely or infrequently drives, only one point location is needed. And using starting point information in track data of the taxis and the network taxi as position information of the user on the taxi to perform density optimization. From the thermodynamic diagram of the track start times, it can be seen that the user's driving behavior is more concentrated in areas that require denser dotting, as these points can play an important role in the recommended service.

Finally, the threshold value between the getting-on points is set to be 50m, and 10 points are hit on each 50m road. If the on-road point is sparse, selecting one point on 50m, namely the midpoint. Meanwhile, the farthest distance between two adjacent boarding points is 50m, and the nearest distance is 5m. Therefore, the generated target on-board point location set can be ensured to meet the requirements of users, and the data storage cost is not too high.

As shown in fig. 6, fig. 6 is a diagram illustrating the generation of a target entry point using the number of times of network traffic track start points. The area 601 shown in the left graph in fig. 6 is higher, which indicates that the behavior of the user frequently getting on the road is comparatively more, the right graph in fig. 6 is not thermodynamic, which indicates that the user hardly gets on the area, so that the area frequently getting on by the user is more dense in the left graph in fig. 6, and the right graph in fig. 6 is that the user does not frequently get on or does not get on, so that one getting on point is enough. In the figure, the on-board point is not a very accurate point, is a range of a point, the coverage of the range is about 5m, and is a point in practical application, but is a range in determination, and the center point of the range is taken as the most accurate point.

Step 304: and acquiring the point of interest data and the crowdsourcing picture, classifying the POI data by using a textCNN text, performing OCR (optical character recognition) on the crowdsourcing picture, and generating visible POI data.

Here, point of interest data and crowd sourced pictures need to be acquired and identified to generate a visible get-on point name. First, names in the point of interest data are classified using TextCNN text classification, as shown in fig. 7, fig. 7 is a TextCNN network structure diagram. TextCNN is a convolutional neural network structure that converts each word into an embedded vector of fixed dimensions, uses convolution operations to generate intermediate hidden vectors, and finally outputs predictive probabilities through a multi-layer perceptron. Therefore, the visibility prediction can be performed on the names in the interest point data according to the model, namely, the names which can be seen by the user are judged, and misleading to the user is avoided.

In addition, for crowd-sourced pictures, recognition is required to be performed by using an OCR technology to extract POI name information in the crowd-sourced pictures. For example, as shown in fig. 8, fig. 8 is a crowdsourcing picture of a quick hotel and a home dish, in which a corresponding POI name is identified. Thus, when the user sees the get-on point name, the user can quickly find the get-on point according to the name.

However, not all roads have crowd-sourced pictures available for OCR recognition. Therefore, it is also possible to roughly judge whether the POI is visible using the gate address information and the name of the POI. For example, if the door address information is "building a base 1810 room," it may be simply determined that the POI is located at 18 floors of the building, and is likely to be invisible. Whereas for a chain of stores like "a house" there is typically a corresponding brand visible.

For POI visibility prediction, in the embodiment of the present application, manually marked training data is used, including visible and invisible point of interest data, which can be randomly extracted from the whole network, and by using the TextCNN method, a classification prediction model (corresponding to the visibility prediction model in other embodiments) is obtained by combining POI name and gate address information. From the classification prediction model, it can be determined whether a POI is visible. And inputting the gate address information, the interest point names and the OCR recognition results of the crowdsourcing pictures into a classification prediction model to perform classification prediction, so that the data of the full-network visible interest point names can be obtained.

Step 305: and generating a complete target get-on point set.

Here, the complete set of target boarding points includes the location and name of each target boarding point. Firstly, dotting is carried out by utilizing road data, and density adjustment is carried out on the whole road network so as to obtain an optimized get-on point set. This means that some of the most representative and suitable locations in the road grid will be selected as candidate locations for the pick-up point.

And then, the data of the starting point of the network vehicle track is utilized, and the selection of the on-vehicle point positions is further optimized by combining the on-vehicle point position set with the density adjusted. The data of the network taxi track starting point can reflect the taxi demand distribution situation of the user, and the analysis of the data can determine which taxi-taking points are more suitable for meeting the demands of the user.

Once multiple target get-on points are obtained, a search may be performed for visible points of interest around each target get-on point. And determining the nearest visible interest point to each target get-on point, if a plurality of visible interest points are the same as the target get-on point in distance and are all the nearest distances, and the nearest distances are smaller than a preset distance threshold (for example, 20 meters), randomly selecting one of the visible interest points, determining the name of the selected visible interest point as the name of the target get-on point, and if only one visible interest point is nearest to the target get-on point and the nearest distances are smaller than the distance threshold, determining the name of the nearest visible interest point as the name of the target get-on point. If the distance between the visible interest point closest to the target get-on point and the target get-on point is greater than a preset distance threshold, the default get-on point name is determined to be the name of the target get-on point, and the default get-on point name may be "network about get-on point" for example, so that the user can accurately find the positions.

Furthermore, in other scenarios, such as cells or office building areas, the traditional recommended get-on point approach is not applicable. Because many cells or office building areas limit network access, recommending pick-up points inside a cell is not appropriate. Therefore, it is necessary to establish a relationship between the boarding point and the area by means of data mining. To establish the relationship of the area and the pick-up point, privacy-desensitized user order data is first required. And extracting the point-to-point relation between the area where the user is and the boarding position of the user according to the data. Then, the relation between the area and the get-on point is established in a clustering mode. For example, as shown in fig. 9, fig. 9 is a schematic view of an interface when a building is driven, and a user driving a vehicle in a building often gets on a south door, a west door, an east door, and a north door of the building. Thus, when other users drive a building, the east, west, south and north doors are preferably recommended for selection by the user.

For a common road side taxi taking, the position is mainly used as a basis for selecting a taxi taking point. However, in the case of driving in a district, office building, market, or the like, it is necessary to consider the concept of area, that is, the data distribution in a certain plane. The map data is provided with some area data, for example, when a cell is driven, most people tend to get on the north door and the south door through clustering, and the relationship between the cell and the north door and the south door can be established. When a user drives in the district, the north door and the south door are recommended to serve as boarding points preferentially.

On the other hand, the target boarding point data of the special station is manufactured by means of manual operation.

Here, in the network taxi-taking platform, in order to better provide services for users, it is necessary to establish a correspondence between an area and a target taxi-taking point, and provide an optimal taxi-taking point selection according to the needs of the users. However, for some special stations, such as airports, train stations, etc., it is necessary to create target boarding point data of the special station by means of manual operation because the corresponding data is specified by an official unit.

For scenes such as airports and railway stations, it is necessary to acquire first the boarding point data prescribed by the authorities of the corresponding stations. Since these data are specified by the authorities, it is necessary to establish correspondence between stations and boarding points by means of manual operation. For example, as shown in fig. 10, fig. 10 is a schematic diagram of a terminal interface when a terminal is driven on a T3 terminal of an airport, and as shown in fig. 10, a loading point 1001 related to the T3 terminal may be selected to be loaded when the terminal is driven on the T3 terminal of the airport. Similarly, the train station and the bus station also have fixed network bus-stop on-bus points, and the corresponding relation between the bus-stop on-bus points and the bus-stop is uniformly planned and managed by airport authorities, so that the target bus-stop data of the special bus-stop can be manufactured in a manual operation mode. Therefore, when the user takes the net to get on the bus at an airport, a railway station or a bus station, the user can select the designated get-on point and take the bus at the designated get-on point, the user can be ensured to take the bus normally, and the situation that the driver waits for passengers at other positions except the designated get-on point for too long time to cause violations can be avoided.

When the corresponding relation between the area and the target get-on point is established, personal information security of the user needs to be protected, and the privacy protection principle is followed. Meanwhile, the recommended get-on point data needs to be updated in time in the operation process so as to ensure the accuracy and instantaneity of the data.

In the embodiment of the application, related data such as user information, historical riding data of a user and the like are related, when the embodiment of the application is applied to specific products or technologies, user permission or consent is required to be obtained, and the collection and use of the related data are required to comply with related laws and regulations and standards of related countries and regions.

According to the data processing method based on the get-on points, the road network is dotted, and the density adjustment is performed by using the network vehicle track starting point data, so that an optimized get-on point set is obtained. Then, a nearby visible POI is searched for each get-on point and assigned a name. And finally, combining the two attributes of the position and the name to generate a complete target get-on point set. The process can enable the user to quickly find the get-on point meeting the requirements, improves the accuracy and efficiency of network taxi-taking matching, recommends the get-on point to the user for use, and can reduce the meeting cost of both drivers and passengers, thereby reducing the quantity of cancelled orders caused by the fact that the user cannot find the taxi, namely reducing the order cancellation rate and improving the profit of network taxi-taking application.

Continuing with the description below of an exemplary architecture of the data processing device 455 implemented as a software module provided by embodiments of the present application, in some embodiments, as shown in FIG. 2, the software modules stored in the data processing device 455 of the memory 450 may include:

a first determining module 4551, configured to obtain road data to be processed, and determine a plurality of initial boarding points in the road data;

a second determining module 4552, configured to obtain historical boarding data, and determine a plurality of target boarding points based on the historical boarding data and the plurality of initial boarding points;

a third determining module 4553, configured to obtain first location information of each target get-on point, and determine a get-on roll name of each target get-on point;

the data storage module 4554 is configured to store first location information and get-on point names of the multiple target get-on points.

In some embodiments, the second determining module 4552 is further configured to obtain a historical ride starting point from the historical ride data; and determining a get-on area of an initial get-on point, and determining the initial get-on point as a target get-on point when the get-on area of the initial get-on point is not a limited parking area and the get-on area comprises a historical get-on starting point.

In some embodiments, the second determining module 4552 is further configured to determine a minimum distance between the initial boarding point and the determined target boarding point when the boarding area of the initial boarding point is not a parking area, and the boarding area does not include a historical boarding start point; deleting the initial get-on point when the minimum distance is smaller than a first distance threshold; and when the minimum distance is greater than or equal to a first distance threshold, determining the initial boarding point as a target boarding point.

In some embodiments, the second determining module 4552 is further configured to delete the initial boarding point when the boarding area of the initial boarding point is a parking area.

In some embodiments, the third determining module 4553 is further configured to obtain a point of interest name of the visible point of interest in the road data and second location information of the visible point of interest; determining the nearest visible interest point corresponding to the target on-board point based on the first position information of the target on-board point and the second position information of the visible interest point; and when the first distance between the nearest visible interest point and the target get-on point is smaller than a second distance threshold value, determining the interest point name of the nearest visible interest point as the get-on point name of the target get-on point.

In some embodiments, the third determining module 4553 is further configured to obtain a default get-on point name when the first distance between the nearest visible interest point and the target get-on point is greater than or equal to the second distance threshold; and determining the default get-on point name as the get-on point name of the target get-on point.

In some embodiments, the third determining module 4553 is further configured to obtain a trained visibility prediction model; acquiring a crowdsourcing picture set and gate address information of each interest point included in the road data; performing image recognition on each crowdsourcing picture in the crowdsourcing picture set to obtain the names of the interest points included in each crowdsourcing picture; and predicting the gate address information and the interest point names of all the interest points by using the trained visibility prediction model to obtain a plurality of visible interest points in the road data.

In some embodiments, the data processing apparatus further comprises: the receiving module is used for receiving third position information of the riding starting point sent by the terminal; the fourth determining module is used for determining a recommended get-on point from the target get-on points based on the third position information; and the sending module is used for sending the recommended get-on point to the terminal so as to present the recommended get-on point in a display interface of the terminal.

In some embodiments, the data processing apparatus further comprises: a fifth determining module, configured to determine, based on the third location information, a target get-on point closest to the get-on start point from among the plurality of target get-on points as a recommended get-on point when determining that the get-on start point is located in a road area; a sixth determining module, configured to determine, based on the third position information, that the boarding start point is located in a building area, and obtain a correspondence between a pre-generated area and a boarding point; and a seventh determining module, configured to obtain a get-on point corresponding to the building area based on the correspondence, and determine the get-on point corresponding to the building area as a recommended get-on point.

In some embodiments, the data processing apparatus further comprises: the extraction module is used for acquiring riding order data sent by a plurality of terminals and extracting the area and the boarding point of the terminal from the riding order data; the clustering module is used for carrying out clustering processing on the area where the terminal is located and the corresponding get-on point to obtain the corresponding relation between the area and the get-on point.

a memory for storing executable instructions;

And the processor is used for realizing the file comparison method provided by the embodiment of the application when executing the executable instructions stored in the memory.

Embodiments of the present application provide a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. The processor of the electronic device reads the computer-executable instructions from the computer-readable storage medium, and executes the computer-executable instructions, so that the electronic device performs the data processing method according to the embodiment of the present application.

Embodiments of the present application provide a computer-readable storage medium storing computer-executable instructions or a computer program stored therein, which when executed by a processor, cause the processor to perform a data processing method provided by embodiments of the present application, for example, the data processing method as shown in fig. 3A, 5.

In some embodiments, the computer readable storage medium may be RAM, ROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; but may be a variety of devices including one or any combination of the above memories.

In some embodiments, computer-executable instructions may be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, in the form of programs, software modules, scripts, or code, and they may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

As an example, computer-executable instructions may, but need not, correspond to files in a file system, may be stored as part of a file that holds other programs or data, such as in one or more scripts in a hypertext markup language (Hyper Text Markup Language, HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).

As an example, computer-executable instructions may be deployed to be executed on one electronic device or on multiple electronic devices located at one site or, alternatively, on multiple electronic devices distributed across multiple sites and interconnected by a communication network.

In summary, by the way of dotting the road network in the embodiment of the application, all sets of initial boarding points are generated, and the density adjustment is performed by using the data of the track starting points such as the network taxi, and the like, so as to obtain the adjusted target boarding point set. Next, searching nearby visible interest points by using the get-on point generated by the road data dotting, and assigning names to the target get-on point. In addition, through user order data mining, the relation between the area and the target boarding point is established, the driving requirements of the user in different scenes are met, and target boarding point data of a special station is manufactured through a manual operation means. The mode adopted by the embodiment of the application can greatly improve the efficiency and convenience of the network taxi service. By means of the road network dotting mode, an initial get-on point set is generated, all roads can be covered more comprehensively, and the target get-on point density is adjusted according to the heat difference of different roads, so that simplified and accurate target get-on point selection is provided. Meanwhile, by searching the nearby interest points and assigning names to the on-target point positions, the user can more conveniently select the on-target point positions. By combining the relation between the area and the get-on point, the get-on requirements of the user in different scenes can be better met, and the service efficiency and the user experience are improved. Finally, the point data of the boarding point of the special station is manufactured in a manual operation mode, so that the driving requirement of a user in a special scene can be met, and the service quality is improved.

The foregoing is merely exemplary embodiments of the present application and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. A method of data processing, the method comprising:

2. The method of claim 1, wherein the determining a plurality of target boarding points based on the historical boarding data and the plurality of initial boarding points comprises:

acquiring a historical riding starting point from the historical riding data;

and determining a get-on area of an initial get-on point, and determining the initial get-on point as a target get-on point when the get-on area of the initial get-on point is not a limited parking area and the get-on area comprises a historical get-on starting point.

3. The method of claim 2, wherein the determining a plurality of target boarding points based on the historical boarding data and the plurality of initial boarding points comprises:

when the boarding area of the initial boarding point is not limited to a parking area and the boarding area does not comprise a historical boarding start point, determining the minimum distance between the initial boarding point and the determined target boarding point;

deleting the initial get-on point when the minimum distance is smaller than a first distance threshold;

and when the minimum distance is greater than or equal to a first distance threshold, determining the initial boarding point as a target boarding point.

4. The method of claim 1, wherein the determining a plurality of target boarding points based on the historical boarding data and the plurality of initial boarding points comprises:

acquiring historical riding starting points included in each road section from the historical riding data;

when at least one historical riding starting point is included in a road section and the total occurrence times of the historical riding starting points are larger than a preset time threshold, clustering the historical riding starting points included in the road section to obtain at least one clustering point;

And when a target clustering point with the closest distance between the initial getting-on point and the initial getting-on point in the road section being larger than a third distance threshold exists in the at least one clustering point, determining the initial getting-on point and the target clustering point in the road section as target getting-on points.

5. The method of claim 1, wherein for each target pick-up point, determining a pick-up roll-up name for the target pick-up point comprises:

acquiring the point of interest name of the visible point of interest in the road data and second position information of the visible point of interest;

determining the nearest visible interest point corresponding to the target on-board point based on the first position information of the target on-board point and the second position information of the visible interest point;

and when the first distance between the nearest visible interest point and the target get-on point is smaller than a second distance threshold value, determining the interest point name of the nearest visible interest point as the get-on point name of the target get-on point.

6. The method of claim 5, wherein said determining a roll-in name for said target roll-in comprises:

when the first distance between the nearest visible interest point and the target get-on point is greater than or equal to the second distance threshold, acquiring a default get-on point name;

And determining the default get-on point name as the get-on point name of the target get-on point.

7. The method according to claim 5 or 6, characterized in that the method further comprises:

obtaining a trained visibility prediction model;

acquiring a crowdsourcing picture set and gate address information of each interest point included in the road data;

performing image recognition on each crowdsourcing picture in the crowdsourcing picture set to obtain the names of the interest points included in each crowdsourcing picture;

and predicting the gate address information and the interest point names of all the interest points by using the trained visibility prediction model to obtain a plurality of visible interest points in the road data.

8. The method according to any one of claims 1 to 6, further comprising:

receiving third position information of a riding starting point sent by a terminal;

determining a recommended get-on point from the plurality of target get-on points based on the third location information;

and sending the recommended get-on point to the terminal so as to present the recommended get-on point in a display interface of the terminal.

9. The method of claim 8, wherein determining a recommended get-on point from the plurality of target get-on points based on the third location information comprises:

Determining a target boarding point closest to the boarding start point from the plurality of target boarding points as a recommended boarding point when the boarding start point is determined to be located in a road area based on the third position information;

based on the third position information, when the riding starting point is determined to be positioned in the building area, acquiring a corresponding relation between a pre-generated area and the boarding point;

and acquiring the get-on point corresponding to the building area based on the corresponding relation, and determining the get-on point corresponding to the building area as the recommended get-on point.

10. The method as recited in claim 9, wherein the method further comprises:

acquiring riding order data sent by a plurality of terminals, and extracting the area and the boarding point of the terminal from the riding order data;

and clustering the area where the terminal is located and the corresponding get-on point to obtain the corresponding relation between the area and the get-on point.

11. A data processing apparatus, the apparatus comprising:

12. An electronic device, the electronic device comprising:

a memory for storing computer executable instructions;

a processor for implementing the method of any one of claims 1 to 10 when executing computer-executable instructions stored in the memory.

13. A computer-readable storage medium storing computer-executable instructions or a computer program, which when executed by a processor implement the method of any one of claims 1 to 10.